From mrauschkolb at ebs.com Tue Aug 2 20:01:26 2005 From: mrauschkolb at ebs.com (Rauschkolb, Mark) Date: Tue, 2 Aug 2005 16:01:26 -0400 Subject: An install that functions from both serial console and local keyboard? Message-ID: We want to have a minimally interactive install that will work over a serial console, or with a locally attached keyboard -- the only interaction at the moment is to reply to a "do you want to wipe the contents of the machine" I can get the config to work with either the attached keyboard (output to and get a reply from /dev/tty1) OR to work with the serial console (/dev/console) but I cannot get it to work with both. Is there a way to communicate with both devices simultaneously? If not, is there a way to determine if there is a keyboard/monitor attached? Thanks Mark The information contained in this e-mail is confidential. This e-mail is intended only for the stated addressee. If you are not an addressee, you must not disclose, copy, circulate or in any other way use or rely on the information contained in this e-mail. if you have received this e-mail in error, please inform us immediately and delete it and all copies from your system. EBS Dealing Resources International Limited. Registered address: 10 Paternoster Square, London EC4M 7DY, United Kingdom. Registered number 2669861. EBS Dealing Resources, Inc, registered in Delaware. Address: 535 Madison Avenue, 24th Floor, New York, NY 10022, USA, and One upper Pond road, Building F - Floor 3, Parsippany, NJ 07054, USA. EBS Dealing Resources Japan Limited, a Japanese Corporation. Address: Asteer Kayabacho Bldg, 6th Floor, 1-6-1, Shinkawa, Chuo-Ku, Tokyo 104-0033, Japan. From error27 at gmail.com Wed Aug 3 02:04:38 2005 From: error27 at gmail.com (Dan Carpenter) Date: Tue, 2 Aug 2005 19:04:38 -0700 Subject: An install that functions from both serial console and local keyboard? In-Reply-To: References: Message-ID: 2005/8/2, Rauschkolb, Mark : > > We want to have a minimally interactive install that will work over a > serial console, or with a locally attached keyboard -- the only > interaction at the moment is to reply to a "do you want to wipe the > contents of the machine" > > I can get the config to work with either the attached keyboard (output > to and get a reply from /dev/tty1) OR to work with the serial console > (/dev/console) but I cannot get it to work with both. > > Is there a way to communicate with both devices simultaneously? > Nope. There isn't an easy way to do that. > If not, is there a way to determine if there is a keyboard/monitor > attached? > If you used a USB keyboard you could type lsusb. regards, dan carpenter From jgao at optonline.net Wed Aug 3 02:12:10 2005 From: jgao at optonline.net (Jian Gao) Date: Tue, 02 Aug 2005 22:12:10 -0400 Subject: Expect Issue In-Reply-To: Message-ID: <0IKM00HBZJH9YE6X@mta9.srv.hcvlny.cv.net> I have some Expect script running fine in FC2 Kickstart, but they are not working any more in FC4 Kickstart. The weird thing is these scripts run just fine in regular FC4 environment. From pmatilai at laiskiainen.org Wed Aug 3 08:20:11 2005 From: pmatilai at laiskiainen.org (Panu Matilainen) Date: Wed, 3 Aug 2005 01:20:11 -0700 (PDT) Subject: Expect Issue In-Reply-To: <0IKM00HBZJH9YE6X@mta9.srv.hcvlny.cv.net> References: <0IKM00HBZJH9YE6X@mta9.srv.hcvlny.cv.net> Message-ID: On Tue, 2 Aug 2005, Jian Gao wrote: > I have some Expect script running fine in FC2 Kickstart, but they are not > working any more in FC4 Kickstart. The weird thing is these scripts run just > fine in regular FC4 environment. Expect needs pty's but /dev/pts is not mounted during kickstart %post. Adding 'mount /dev/pts' before the expect script should help. - Panu - From phr at doc.ic.ac.uk Wed Aug 3 12:35:14 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 3 Aug 2005 13:35:14 +0100 (BST) Subject: An install that functions from both serial console and local keyboard? In-Reply-To: References: Message-ID: On Tue, 2 Aug 2005, Rauschkolb, Mark wrote: >We want to have a minimally interactive install that will work over a >serial console, or with a locally attached keyboard -- the only >interaction at the moment is to reply to a "do you want to wipe the >contents of the machine" If you can assume kernel commandline args will put anaconda into serial mode, then the I/O will take care of itself. dialog(1) supports timeouts, as does bash's read builtin. Cheers, Phil From steve.mah at oracle.com Wed Aug 3 17:29:12 2005 From: steve.mah at oracle.com (Stephen Mah) Date: Wed, 03 Aug 2005 10:29:12 -0700 Subject: updating RPM directory Message-ID: <42F0FEE8.9090901@oracle.com> I think I've asked before, but can't seem to find that email. Does anyone have a script to update an existing directory with newer rpms from a given folder? For example, I have an RPM folder with customizations, and I want to update the RPMS from a directory that has update 1 plus newer erratas? Does anyone have such a script, before I try to write one? thanks steve From klaus.steden at thomson.net Wed Aug 3 17:43:42 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 3 Aug 2005 13:43:42 -0400 Subject: updating RPM directory In-Reply-To: <42F0FEE8.9090901@oracle.com> References: <42F0FEE8.9090901@oracle.com> Message-ID: <20050803174342.GP897@thomson.net> > I think I've asked before, but can't seem to find that email. > > Does anyone have a script to update an existing directory with newer > rpms from a given folder? > > For example, I have an RPM folder with customizations, and I want to > update the RPMS from a directory that has update 1 plus newer erratas? > Does anyone have such a script, before I try to write one? > Try this. I can't claim ownership or authorship of it - someone on this list wrote it, but I've been using it for the past two months. It's a life-saver. Klaus -------------- next part -------------- #!/usr/bin/python # # update_release.py 2.0 - merge rpm updates and third party rpms # into a RedHat distro and check for # dependencies/conflicts problems # Copyright (C) 2003 - Zouhir Hafidi (Zouhir.Hafidi at agat.univ-lille1.fr) # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # # Usage (some examples): # - get help: # update_release.py -h # - run in test mode and use current dir as a base dir # update_release.py -n # update_release.py -n -v # - run in test mode and use /export/install/redhat/8.0 as a base dir # update_release.py -n -d /export/install/redhat/8.0 # update_release.py -n -v -d /export/install/redhat/8.0 # - use current dir as a base dir. The installation dir and updates dir # are respectively by default updated_release/RedHat/RPMS/ and updates/ # update_release.py # update_release.py -v # - use /export/install/redhat/8.0 as a base dir. The installation dir and # updates dir are respectively by default updated_release/RedHat/RPMS/ and updates/ # update_release.py -d /export/install/redhat/8.0 # update_release.py -v -d /export/install/redhat/8.0 # - use /export/install/redhat/8.0 as a base dir and only check for # dependencies/conflicts under updated_release/RedHat/RPMS # update_release.py -c -d /export/install/redhat/8.0 # update_release.py -c -v -d /export/install/redhat/8.0 # - use /export/install/redhat/8.0 as a base dir and only check for # dependencies/conflicts under some/relative/dir # update_release.py -c -d /export/install/redhat/8.0 -i some/relative/dir # update_release.py -c -v -d /export/install/redhat/8.0 -i some/relative/dir import os import sys import getopt import commands import rpm import string import shutil def usage(): print "Usage: %s [-h] | [[-n] [-v] [-d dir] [-i dir] [-u dir] [-b branch] [-c] [-H]]" % sys.argv[0] print " -h|--help print this message" print " -n|--dryrun dry run mode (show a trace of what would be done)" print " -v|--verbose print more information when running" print " -d dir|--basedir=dir base dir (current dir by default)" print " -i dir|--installdir=dir installation dir (updated_release/RedHat/RPMS/ by default)" print " -u dir|--updatesdir=dir updates dir (updates/ by default)" print " -b branch|--branch=vers RedHat version (i.e. Fedora, RedHat)" print " -c|--checkonly check for dependencies/conflicts in the installation dir" print " -H|--nohdlist don't regenerate hdlists" # Description: # construct a dependency string (this function comes from anaconda) # Argument(s): # - rpm name # - rpm version # - dependency flags # Return value(s): # - dependency string def formatRequire (name, version, flags): string = name if flags: if flags & (rpm.RPMSENSE_LESS | rpm.RPMSENSE_GREATER | rpm.RPMSENSE_EQUAL): string = string + " " if flags & rpm.RPMSENSE_LESS: string = string + "<" if flags & rpm.RPMSENSE_GREATER: string = string + ">" if flags & rpm.RPMSENSE_EQUAL: string = string + "=" string = string + " %s" % version return string # Description: # construct a dictionary containing the headers of a set of rpms. # If a same rpm exists with different versions, then take into # account only the recent one. Source rpms are not processed. # Argument(s): # - a list of rpms filenames # Return value(s): # - a dictionary of the form: # {(rpm_name, rpm_arch):(rpm_header, rpm_filename), ...} # - a list containing source rpms and old rpms def getRpmsHeaders(rpms): headers = {} bad_rpms = [] ts = rpm.TransactionSet("/", ~(rpm._RPMVSF_NOSIGNATURES)) ts.closeDB() for rpm_file in rpms: fdno = os.open(rpm_file, os.O_RDONLY) header = ts.hdrFromFdno(fdno) os.close(fdno) if string.find(rpm_file, "debuginfo") >= 0: bad_rpms.append(rpm_file) if verbose_flag: print " skip %s which is a debug package" % os.path.basename(rpm_file) elif header[rpm.RPMTAG_SOURCEPACKAGE]: bad_rpms.append(rpm_file) if verbose_flag: print " skip %s which is a source package" % os.path.basename(rpm_file) else: key = (header[rpm.RPMTAG_NAME], header[rpm.RPMTAG_ARCH]) if headers.has_key(key): cmp = rpm.versionCompare(headers[key][0], header) if not cmp in [-1, 1]: print " rpm.versionCompare() returned error code %d while comparing %s and %s" % (cmp, os.path.basename(headers[key][1]), os.path.basename(rpm_file)) sys.exit(cmp) if cmp == -1: bad_rpms.append(headers[key][1]) if verbose_flag: print " use %s instead of %s" % (os.path.basename(rpm_file), os.path.basename(headers[key][1])) headers[key] = (header, rpm_file) else: bad_rpms.append(rpm_file) if verbose_flag: print " skip %s which is older than %s" % (os.path.basename(rpm_file), os.path.basename(headers[key][1])) else: headers[key] = (header, rpm_file) return (headers, bad_rpms) # Description: # delete unwanted entries from a list of rpms filenames # Argument(s): # - list of rpms filenames # - list of entries to delete # Return value(s): # None def skipRpms(rpms_list, skip_list): for item in skip_list: del rpms_list[rpms_list.index(item)] # Description: # check if a given set of rpms put together doesn't have any # dependencies nor conflicts problems # Argument(s): # - a dictionary of the form: # {(rpm_name, rpm_arch):(rpm_header, rpm_filename), ...} # Return value(s): # - a list of errors if any def checkDepsAndConflicts(headers): ts = rpm.TransactionSet("/", ~(rpm._RPMVSF_NOSIGNATURES)) ts.closeDB() for key in headers.keys(): ts.addInstall(headers[key][0], key, 'i') return ts.check() def main(): # Processing command line arguments try: opts, args = getopt.getopt(sys.argv[1:], 'hnvd:i:u:b:cHG', \ ["help", "dryrun", "verbose", "basedir=", "installdir=", \ "updatesdir=", "branch=", "checkonly", "nohdlist", "genhdlist"]) except getopt.GetoptError: usage() sys.exit(1) if args != []: usage() sys.exit(1) global verbose_flag verbose_flag = 0 dryrun_flag = 0 base_dir = os.getcwd() # updated_release_dir = "updated_release/RedHat/RPMS" updated_release_dir = "Fedora/RPMS" updates_dir = "updates" branch = "Fedora" checkonly_flag = 0 nohdlist_flag = 0 genhdlist_flag = 0 for o, a in opts: if o in ("-h", "--help"): usage() sys.exit() if o in ("-n", "--dryrun"): dryrun_flag = 1 if o in ("-v", "--verbose"): verbose_flag = 1 if o in ("-d", "--basedir"): base_dir = a if o in ("-i", "--installdir"): updated_release_dir = a if o in ("-u", "--updatesdir"): updates_dir = a if o in ("-b", "--branch"): branch = a if o in ("-c", "--checkonly"): checkonly_flag = 1 if o in ("-H", "--nohdlist"): nohdlist_flag = 1 if o in ("-G", "--genhdlist"): genhdlist_flag = 1 # Some initializations updated_release_dir = os.path.join(base_dir, updated_release_dir) updates_dir = os.path.join(base_dir, updates_dir) old_rpms_dir = os.path.join(base_dir, "old_rpms") regenerate_hdlist = 0 if dryrun_flag: verbose_flag = 1 # The updated release dir must exist if not os.path.isdir(updated_release_dir): print "can't access %s, nothing to do" % updated_release_dir sys.exit(2) if verbose_flag: print "processing rpms under %s ..." % updated_release_dir # Get all rpm names under updated_release_dir exit_status, output = commands.getstatusoutput("find %s -name '*\.rpm' | grep -v '.src.rpm'" % updated_release_dir) if exit_status != 0: sys.exit(3) updated_release_rpms = output.split() # Get headers of all rpms under updated_release_dir updated_release_headers, bad_rpms = getRpmsHeaders(updated_release_rpms) if bad_rpms != []: print " ERROR: the following rpms are either source rpms or older rpms. Please remove them:" for rpm_file in bad_rpms: print " %s" % rpm_file sys.exit(4) # Before merging, check if the current updated release dir contains # any dependencies or conflicts problems and exit if so errors = checkDepsAndConflicts(updated_release_headers) if errors: print "ERROR: the installation tree already contains conflicts and/or dependencies problems" for ((name, version, release), (reqname, reqversion), \ flags, suggest, sense) in errors: if sense==rpm.RPMDEP_SENSE_REQUIRES: print " depcheck: package %s needs %s" % ( name, formatRequire(reqname, reqversion, flags)) elif sense==rpm.RPMDEP_SENSE_CONFLICTS: print " depcheck: package %s conflicts with %s" % (name, reqname) sys.exit(4) if verbose_flag: print "done" if checkonly_flag: if errors: print "dependencies/conflicts problems exist in %s" % updated_release_dir else: print "no dependencies/conflicts problems in %s" % updated_release_dir sys.exit() # The updates dir must exist if not os.path.isdir(updates_dir): print "can't access %s, nothing to do" % updates_dir sys.exit(2) # If necessary, create the dir to put old rpms into if not os.path.isdir(old_rpms_dir): os.mkdir(old_rpms_dir) if verbose_flag: print "processing rpms under %s ..." % updates_dir # Get all rpm names under updates_dir exit_status, output = commands.getstatusoutput("find %s -follow -name '*\.rpm' | grep -v '.src.rpm'" % updates_dir) if exit_status != 0: sys.exit(3) updates_rpms = output.split() # Get headers of all rpms under updates_dir updates_headers, bad_rpms = getRpmsHeaders(updates_rpms) # Ignore unwanted rpms skipRpms(updates_rpms, bad_rpms) if bad_rpms != []: print " WARNING: the following rpms are either source rpms or older rpms. I will ignore them:" for rpm_file in bad_rpms: print " %s" % rpm_file if verbose_flag: print "done" # Take the entries in the updates_headers dictionary and merge them # with the entries in the updated_release_headers dictionary: # - for every entry in the updates_headers dictionary # - if a corresponding entry with an old version exists # in the updated_release_headers dictionary then drop it # in the old_rpms_headers dictionary and replace it with # the new one # - if there is no corresponding entry with an old version # then simply add the new one in the updated_release_headers # dictionary old_rpms_headers = {} for key in updates_headers.keys(): if updated_release_headers.has_key(key): cmp = rpm.versionCompare(updated_release_headers[key][0], updates_headers[key][0]) if cmp == 1: if verbose_flag: print "skip %s which is older than %s" % (os.path.basename(updates_headers[key][1]), os.path.basename(updated_release_headers[key][1])) elif cmp == 0: if verbose_flag: print "skip %s which is the same as %s" % (os.path.basename(updates_headers[key][1]), os.path.basename(updated_release_headers[key][1])) elif cmp == -1: if verbose_flag: print "exchange %s with %s" % (os.path.basename(updated_release_headers[key][1]), os.path.basename(updates_headers[key][1])) old_rpms_headers[key] = updated_release_headers[key] updated_release_headers[key] = updates_headers[key] else: print " rpm.versionCompare() returned error code %d while comparing %s and %s" % (cmp, os.path.basename(updated_release_headers[key][1]), os.path.basename(updates_headers[key][1])) sys.exit(cmp) else: updated_release_headers[key] = updates_headers[key] if verbose_flag: print "add %s" % os.path.basename(updates_headers[key][1]) # At this point, the merging process is done and the updated_release_headers # dictionary should be up to date. Before applying the changes to the updated # release dir, we must make sure that there's no dependencies nor conflicts # problems. If so, we do a "backtracking" on the headers which make problems # until a state without any dependencies/conflicts is reached. while 1: deps_and_conflicts = 0 errors = checkDepsAndConflicts(updated_release_headers) if errors: # Construct a list containing the names of the rpms that # make problems. If a same rpm exists for different # architectures, then all of them will not be used. # NOTE: we don't know how to discard only problematic rpms # since the architecture is not reported by the check() # method (is there a fix ?) if verbose_flag: print "merging will result in the following conflicts and/or dependencies problems:" name_list = [] for ((name, version, release), (reqname, reqversion), \ flags, suggest, sense) in errors: if sense==rpm.RPMDEP_SENSE_REQUIRES: if verbose_flag: print " depcheck: package %s needs %s" % ( name, formatRequire(reqname, reqversion, flags)) deps_and_conflicts = 1 if not name in name_list: name_list.append(name) elif sense==rpm.RPMDEP_SENSE_CONFLICTS: if verbose_flag: print " depcheck: package %s conflicts with %s" % (name, reqname) deps_and_conflicts = 1 if not name in name_list: name_list.append(name) # Discard entries with dependencies/conflicts problems in # the updated_release_headers dictionary if verbose_flag: print "backtracking:" for (name, arch) in updated_release_headers.keys(): if name in name_list : if os.path.dirname(updated_release_headers[(name, arch)][1]) == updated_release_dir: print "CRITICAL ERROR: attempt to discard %s" % updated_release_headers[(name, arch)][1] sys.exit(4) if verbose_flag: print " discard %s" % updated_release_headers[(name, arch)][1] del updated_release_headers[(name, arch)] # Put back those entries which have been replaced by recent # entries which resulted in dependencies/conflicts problems for (name, arch) in old_rpms_headers.keys(): if name in name_list : if verbose_flag: print " reuse %s" % old_rpms_headers[(name, arch)][1] updated_release_headers[(name, arch)] = old_rpms_headers[(name, arch)] del old_rpms_headers[(name, arch)] # The infinite loop *should* be broken after some iterations if deps_and_conflicts == 0: break # At this point, the updated_release_headers dictionary contains # entries with the rpms filenames that should be in the updated # release dir so that we have no dependencies/conflicts problems. # The old_rpms_headers dictionary contains entries with the rpms # filenames that should be moved from the updated release dir to # the old rpms dir. Just reflect these changes on the directories # themselves for key in updated_release_headers.keys(): if os.path.dirname(updated_release_headers[key][1]) == updated_release_dir: continue regenerate_hdlist = 1 if not dryrun_flag: # os.link(updated_release_headers[key][1], \ shutil.copy2(updated_release_headers[key][1], \ os.path.join(updated_release_dir, os.path.basename(updated_release_headers[key][1]))) else: print "should", if key in old_rpms_headers.keys(): print "update %s with %s" % (os.path.basename(old_rpms_headers[key][1]), \ os.path.basename(updated_release_headers[key][1])) if not dryrun_flag: os.rename(old_rpms_headers[key][1], \ os.path.join(old_rpms_dir, os.path.basename(old_rpms_headers[key][1]))) else: print "add %s" % os.path.basename(updated_release_headers[key][1]) # Regenerate hdlists if necessary if genhdlist_flag or ((not nohdlist_flag) and regenerate_hdlist): if not dryrun_flag: os.system("/usr/lib/anaconda-runtime/genhdlist --withnumbers --productpath %s %s/../.." % (branch, updated_release_dir) ) os.system("PYTHONPATH=/usr/lib/anaconda /usr/lib/anaconda-runtime/pkgorder %s/../.. i386 %s > pkgorder" % (updated_release_dir, branch) ) os.system("/usr/lib/anaconda-runtime/genhdlist --withnumbers --productpath %s --fileorder pkgorder %s/../.." % (branch, updated_release_dir) ) else: if verbose_flag: print "should", if verbose_flag: print "regenerate hdlist" if __name__ == "__main__": main() From cleve.cornelius at kroger.com Wed Aug 3 20:00:48 2005 From: cleve.cornelius at kroger.com (cleve.cornelius at kroger.com) Date: Wed, 3 Aug 2005 16:00:48 -0400 Subject: Cleve Cornelius/MIS/CORP/KrogerCo is out of the office. Message-ID: I will be out of the office starting 08/01/2005 and will not return until 08/08/2005. From sudoyang at gmail.com Wed Aug 3 23:28:15 2005 From: sudoyang at gmail.com (sudo Yang) Date: Wed, 3 Aug 2005 16:28:15 -0700 Subject: Determinig hostname and domain... Message-ID: <4f52331f05080316285d28fa56@mail.gmail.com> During kickstart, Anaconda displays this message "Determining hostname and domain...." If we do not use DNS during kickstart, is there an option in the KS file that can disable this? From klaus.steden at thomson.net Wed Aug 3 23:32:50 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 3 Aug 2005 19:32:50 -0400 Subject: Determinig hostname and domain... In-Reply-To: <4f52331f05080316285d28fa56@mail.gmail.com> References: <4f52331f05080316285d28fa56@mail.gmail.com> Message-ID: <20050803233250.GR897@thomson.net> > During kickstart, Anaconda displays this message "Determining hostname > and domain...." If we do not use DNS during kickstart, is there an > option in the KS file that can disable this? > You can specify a hostname and other network parameters by defining the 'network' statement in your ks.cfg. hth, Klaus From harry at mad-cat.co.uk Thu Aug 4 06:25:38 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 04 Aug 2005 07:25:38 +0100 (BST) Subject: updating RPM directory Message-ID: On Wed, 03 Aug, Stephen Mah wrote: > I think I've asked before, but can't seem to find that email. > > Does anyone have a script to update an existing directory with newer > rpms from a given folder? > > For example, I have an RPM folder with customizations, and I want to > update the RPMS from a directory that has update 1 plus newer erratas? > Does anyone have such a script, before I try to write one? Try novi: http://www.exmachinatech.net/01/novi/ I have been using it to create new yum repos and ks install trees. Not sure if it's what you want but it has saved me a huge amount of time messing around. Harry From David.Heley at bom.gov.au Thu Aug 4 06:30:18 2005 From: David.Heley at bom.gov.au (David Heley) Date: Thu, 4 Aug 2005 16:30:18 +1000 Subject: Kickstart-list Digest, Vol 18, Issue 2 Message-ID: <00E49660AB21DD44A0D4A3115DFCBF61269BD2@officeho2.bom.gov.au> I think it would be better if the subject matters were in the actual email instead of attachments. I tend not to read them as attachments. -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]On Behalf Of kickstart-list-request at redhat.com Sent: Thursday, 4 August 2005 16:26 To: kickstart-list at redhat.com Subject: Kickstart-list Digest, Vol 18, Issue 2 Send Kickstart-list mailing list submissions to kickstart-list at redhat.com To subscribe or unsubscribe via the World Wide Web, visit https://www.redhat.com/mailman/listinfo/kickstart-list or, via email, send a message with subject or body 'help' to kickstart-list-request at redhat.com You can reach the person managing the list at kickstart-list-owner at redhat.com When replying, please edit your Subject line so it is more specific than "Re: Contents of Kickstart-list digest..." From gobbledegeek at gmail.com Fri Aug 5 06:13:34 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Fri, 5 Aug 2005 11:43:34 +0530 Subject: ks.cfg crashes install Message-ID: <463aea5705080423134b8df31@mail.gmail.com> Hello Any format errors in the file below? Its making anaconda install crash... ---------------------- autostep autopart install cdrom lang en_US.UTF-8 langsupport --default=en_US.UTF-8 en_US.UTF-8 keyboard us network --bootproto=static --ip=192.168.1.254 --netmask=255.255.255.0 --gateway=192.168.1.1 --device=eth0 --nameserver=61.1.96.6 9 --hostname solvix.in rootpw solvix firewall --enabled --port=22:tcp --port=80:tcp --port=21:tcp --port=25:tcp --port=110:tcp firstboot -- selinux --disabled authconfig --enablemd5 --enableshadow timezone Asia/Calcutta bootloader --append="linux vga=791" --location=mbr %post /sbin/chkconfig --level 345 sshd on /sbin/chkconfig --level 345 httpd on /sbin/chkconfig --level 345 dhcpd on /bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" > /etc/issue ----------------------------- -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From klaus.steden at thomson.net Fri Aug 5 08:11:47 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 5 Aug 2005 04:11:47 -0400 Subject: ks.cfg crashes install In-Reply-To: <463aea5705080423134b8df31@mail.gmail.com> References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: <20050805081147.GD897@thomson.net> This: -- cut -- --gateway=192.168.1.1 --device=eth0 --nameserver=61.1.96.6 9 --hostname solvix.in -- cut -- looks wrong. I don't think line wraps are supported for command directives, and anyway .... what's the "9" mean? ;> hth, Klaus I could have sworn Gobbledegeek said this Fri, Aug 05, 2005 at 11:43:34AM +0530 ... > Hello > Any format errors in the file below? Its making anaconda > install crash... > ---------------------- > autostep > autopart > install > cdrom > lang en_US.UTF-8 > langsupport --default=en_US.UTF-8 en_US.UTF-8 > keyboard us > network --bootproto=static --ip=192.168.1.254 --netmask=255.255.255.0 > --gateway=192.168.1.1 --device=eth0 --nameserver=61.1.96.6 > 9 --hostname solvix.in > rootpw solvix > firewall --enabled --port=22:tcp --port=80:tcp --port=21:tcp > --port=25:tcp --port=110:tcp > firstboot -- > selinux --disabled > authconfig --enablemd5 --enableshadow > timezone Asia/Calcutta > bootloader --append="linux vga=791" --location=mbr > %post > /sbin/chkconfig --level 345 sshd on > /sbin/chkconfig --level 345 httpd on > /sbin/chkconfig --level 345 dhcpd on > /bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" > > /etc/issue > ----------------------------- > -- > Nonchalantly yours > GobbledeGeek > [Every thing but Gobbledegook.. !!] > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- Klaus Steden | Senior Systems Administrator | Technicolor Creative Services | TODO: Toronto | 1) Learn to use my new Unix account. klaus.steden at thomson.net | 2) Learn how to change this list. Phone: (416) 585-9995 | Fax: (416) 364-1585 | From herrold at owlriver.com Fri Aug 5 13:21:14 2005 From: herrold at owlriver.com (R P Herrold) Date: Fri, 5 Aug 2005 09:21:14 -0400 (EDT) Subject: ks] ks.cfg crashes install In-Reply-To: <463aea5705080423134b8df31@mail.gmail.com> References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: On Fri, 5 Aug 2005, Gobbledegeek wrote: > /bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" >> /etc/issue one assumes you nean to use 'echo' in teh preceeding line as to the crash cause, the python traceback would have the cause on the screen. - Russ Herrold From ben_m_f at yahoo.co.uk Fri Aug 5 14:09:44 2005 From: ben_m_f at yahoo.co.uk (Ben Fitzgerald) Date: Fri, 5 Aug 2005 15:09:44 +0100 Subject: ks] ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: <20050805140944.GB29755@bfitzgerald.co.uk> On Fri, Aug 05, 2005 at 09:21:14AM -0400, R P Herrold wrote: > On Fri, 5 Aug 2005, Gobbledegeek wrote: > > >/bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" > >>/etc/issue > > one assumes you nean to use 'echo' in teh preceeding line > > as to the crash cause, the python traceback would have the > cause on the screen. I've found this isn't always the case. Sometimes a lot of text will be dumped that makes the real reason disappear in output that is forced off-screen. Is there a way yet to log remotely? Besides outputting to a serial interface and capturing that? Assuming the interface is up, how about a netcat to a logging server? Guess I can hack initrd. On the same thread, can one easily enable telnetd during the install? I have done this during sun flash installs to telnet in and tail the install log so I can track build progress on a remote, headless server. ben. -- Registered Linux user number 339435 From klaus.steden at thomson.net Fri Aug 5 15:23:00 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 5 Aug 2005 11:23:00 -0400 Subject: ks] ks.cfg crashes install In-Reply-To: <20050805140944.GB29755@bfitzgerald.co.uk> References: <463aea5705080423134b8df31@mail.gmail.com> <20050805140944.GB29755@bfitzgerald.co.uk> Message-ID: <20050805152300.GE897@thomson.net> I could have sworn Ben Fitzgerald said this Fri, Aug 05, 2005 at 03:09:44PM +0100 ... > > I've found this isn't always the case. Sometimes a lot of text will be > dumped that makes the real reason disappear in output that is forced > off-screen. Is there a way yet to log remotely? Besides outputting to a > serial interface and capturing that? Assuming the interface is up, how > about a netcat to a logging server? Guess I can hack initrd. > I've been thinking about implementing a remote syslog facility in Anaconda but so far haven't had the time. I think it's high time one got added. Klaus From dky at utcc.utoronto.ca Fri Aug 5 16:45:53 2005 From: dky at utcc.utoronto.ca (Derek Yeung) Date: Fri, 5 Aug 2005 12:45:53 -0400 (EDT) Subject: ks] ks.cfg crashes install In-Reply-To: <20050805140944.GB29755@bfitzgerald.co.uk> References: <463aea5705080423134b8df31@mail.gmail.com> <20050805140944.GB29755@bfitzgerald.co.uk> Message-ID: Presumably that is possible if you started the network either in %pre or in %post -- each with it's own set of perks (and problems?) ;-) /dky On Fri, 5 Aug 2005, Ben Fitzgerald wrote: > On Fri, Aug 05, 2005 at 09:21:14AM -0400, R P Herrold wrote: >> On Fri, 5 Aug 2005, Gobbledegeek wrote: >> >>> /bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" >>>> /etc/issue >> >> one assumes you nean to use 'echo' in teh preceeding line >> >> as to the crash cause, the python traceback would have the >> cause on the screen. > > I've found this isn't always the case. Sometimes a lot of text will be > dumped that makes the real reason disappear in output that is forced > off-screen. Is there a way yet to log remotely? Besides outputting to a > serial interface and capturing that? Assuming the interface is up, how > about a netcat to a logging server? Guess I can hack initrd. > > On the same thread, can one easily enable telnetd during the install? > I have done this during sun flash installs to telnet in and tail the > install log so I can track build progress on a remote, headless server. > > ben. > > From klaus.steden at thomson.net Fri Aug 5 17:05:50 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Fri, 5 Aug 2005 13:05:50 -0400 Subject: ks] ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> <20050805140944.GB29755@bfitzgerald.co.uk> Message-ID: <20050805170549.GL897@thomson.net> > > Presumably that is possible if you started the network either in %pre or > in %post -- each with it's own set of perks (and problems?) ;-) > >From a cursory examination of the appropriate Anaconda Python module, it shouldn't be too difficult, assuming that any remote host provided must be provided by IP, since DNS is not necessarily working. At any rate, as soon as the IP stack is up and running and the system has an address and a gateway (if necessary), then sending messages via UDP to an external syslogd is elementary. Still, I like the mechanism available on *BSD via the 'logger' command. $0.02, Klaus From gobbledegeek at gmail.com Sat Aug 6 06:05:54 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Fri, 5 Aug 2005 23:05:54 -0700 Subject: ks] ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: <463aea5705080523057255d77f@mail.gmail.com> it doesn't!! it crashes before the 1st anaconda screen. install works ok without kickstart... I'll change the cat to echo thanks ... Rgrds On 8/5/05, R P Herrold wrote: > On Fri, 5 Aug 2005, Gobbledegeek wrote: > > > /bin/cat " (Based on Fedora Core 3)\n\r Kernel \r on an \m \n\r" > >> /etc/issue > > one assumes you nean to use 'echo' in teh preceeding line > > as to the crash cause, the python traceback would have the > cause on the screen. > > - Russ Herrold > -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From gobbledegeek at gmail.com Mon Aug 8 08:16:00 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Mon, 8 Aug 2005 13:46:00 +0530 Subject: ks.cfg crashes install In-Reply-To: <20050806160032.BF062734C7@hormel.redhat.com> References: <20050806160032.BF062734C7@hormel.redhat.com> Message-ID: <463aea57050808011625dc826e@mail.gmail.com> Well in my case, anaconda crashes before first screen - screen blanks immediately on crash. Works ok without ks. Is the sequence of commands important in ks.cfg? -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From phr at doc.ic.ac.uk Mon Aug 8 09:41:17 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Mon, 8 Aug 2005 10:41:17 +0100 (BST) Subject: ks.cfg crashes install In-Reply-To: <463aea57050808011625dc826e@mail.gmail.com> References: <20050806160032.BF062734C7@hormel.redhat.com> <463aea57050808011625dc826e@mail.gmail.com> Message-ID: On Mon, 8 Aug 2005, Gobbledegeek wrote: >Well in my case, anaconda crashes before first screen - screen blanks >immediately on crash. Works ok without ks. Is the sequence of commands >important in ks.cfg? %packages should follow anaconda "commands"; %post should follow %packages. Within these sections, order is irrelevent. What error message do you see when anaconda crashes. What's the very last thing to happen/print before the crash? Cheers, Phil From clumens at redhat.com Mon Aug 8 13:39:13 2005 From: clumens at redhat.com (Chris Lumens) Date: Mon, 8 Aug 2005 09:39:13 -0400 (EDT) Subject: ks.cfg crashes install In-Reply-To: <463aea5705080423134b8df31@mail.gmail.com> References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: > langsupport --default=en_US.UTF-8 en_US.UTF-8 Is this an FC4 or a Rawhide install? The langsupport keyword has been removed from kickstart in Rawhide and will cause an error to be printed on the screen explaining how it's deprecated. > firstboot -- This might be it. Is there supposed to be anything after the -- that got mangled in your post? - Chris From tdiehl at rogueind.com Mon Aug 8 14:54:56 2005 From: tdiehl at rogueind.com (Tom Diehl) Date: Mon, 8 Aug 2005 10:54:56 -0400 (EDT) Subject: ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: On Mon, 8 Aug 2005, Chris Lumens wrote: > > langsupport --default=en_US.UTF-8 en_US.UTF-8 > > Is this an FC4 or a Rawhide install? The langsupport keyword has been > removed from kickstart in Rawhide and will cause an error to be printed > on the screen explaining how it's deprecated. In addition fc4 kickstart is B0rked!! See the archives of this list or bugzilla for more info. I do not have the bug handy but there is a simple workaround. > > > firstboot -- > > This might be it. Is there supposed to be anything after the -- that > got mangled in your post? > > - Chris > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > Tom Diehl tdiehl at rogueind.com Spamtrap address mtd123 at rogueind.com From clumens at redhat.com Mon Aug 8 15:05:34 2005 From: clumens at redhat.com (Chris Lumens) Date: Mon, 8 Aug 2005 11:05:34 -0400 (EDT) Subject: ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: > In addition fc4 kickstart is B0rked!! See the archives of this list or > bugzilla for more info. I do not have the bug handy but there is a simple > workaround. Yes, but I believe you're referring to specifying packages that should not be installed. I don't see that he's doing any of that though. - Chris From tdiehl at rogueind.com Mon Aug 8 15:51:25 2005 From: tdiehl at rogueind.com (Tom Diehl) Date: Mon, 8 Aug 2005 11:51:25 -0400 (EDT) Subject: ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: On Mon, 7 Aug 2005, Chris Lumens wrote: > > In addition fc4 kickstart is B0rked!! See the archives of this list or > > bugzilla for more info. I do not have the bug handy but there is a simple > > workaround. > > Yes, but I believe you're referring to specifying packages that should > not be installed. I don't see that he's doing any of that though. Ok, I did not see the original post. All I saw was FC4 and broken kickstart. :-) I found the one I was referring to, for the archives it is: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160209 Regards, Tom Diehl tdiehl at rogueind.com Spamtrap address mtd123 at rogueind.com From bill at wiliweld.com Mon Aug 8 16:23:24 2005 From: bill at wiliweld.com (Bill Schoolcraft) Date: Mon, 8 Aug 2005 09:23:24 -0700 (PDT) Subject: ks.cfg crashes install In-Reply-To: References: <20050806160032.BF062734C7@hormel.redhat.com> <463aea57050808011625dc826e@mail.gmail.com> Message-ID: <20050808092038.F28159@liam.billschoolcraft.com> At Mon, 8 Aug 2005 it looks like Philip Rowlands composed: > On Mon, 8 Aug 2005, Gobbledegeek wrote: > >> Well in my case, anaconda crashes before first screen - screen blanks >> immediately on crash. Works ok without ks. Is the sequence of commands >> important in ks.cfg? > > %packages should follow anaconda "commands"; %post should follow > %packages. Within these sections, order is irrelevent. > > What error message do you see when anaconda crashes. What's the very > last thing to happen/print before the crash? > Hello Family, I believe, from my experience, that if you go and look at the console on screen + it will show you the stdout of what command was being issued when failure occurred and the previous successful commands. It's like a running "tail -f" on /var/log/messages Hope that helps. -- Bill Schoolcraft PO Box 210076 San Francisco, CA 94121 http://billschoolcraft.com From chandan-dutta.chowdhury at hp.com Tue Aug 9 09:24:51 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Tue, 9 Aug 2005 14:54:51 +0530 Subject: Getting install log on a remote machine Message-ID: Hello All, I want to get the install time log on a remote machine, so that if I am away from the console of the machine being installed I can still have access to the anaconda logs of the machine being installed. Is there a way to do this in kickstart pre install script (asking anaconda to send logs to a network log server) ? Thanks in advance Chandan From klaus.steden at thomson.net Tue Aug 9 16:01:53 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 9 Aug 2005 12:01:53 -0400 Subject: Getting install log on a remote machine In-Reply-To: References: Message-ID: <20050809160153.GT897@thomson.net> > Hello All, > > I want to get the install time log on a remote machine, so that if I am > away from the console of the machine being installed I can still have > access to the anaconda logs of the machine being installed. > > Is there a way to do this in kickstart pre install script (asking > anaconda to send logs to a network log server) ? > Nope. Seems this question gets asked almost weekly. Klaus From miller_duane at bah.com Tue Aug 9 17:37:10 2005 From: miller_duane at bah.com (Duane Miller) Date: Tue, 09 Aug 2005 10:37:10 -0700 Subject: Accessing the CD from %post In-Reply-To: References: Message-ID: <42F8E9C6.3010103@bah.com> I built a DVD to install from with kickstart file built in. One more thing I need to do: Can I access the files on the DVD from my %post or %pre script? If so how? D From phr at doc.ic.ac.uk Tue Aug 9 17:53:32 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Tue, 9 Aug 2005 18:53:32 +0100 (BST) Subject: Accessing the CD from %post In-Reply-To: <42F8E9C6.3010103@bah.com> References: <42F8E9C6.3010103@bah.com> Message-ID: On Tue, 9 Aug 2005, Duane Miller wrote: >I built a DVD to install from with kickstart file built in. > >One more thing I need to do: > >Can I access the files on the DVD from my %post or %pre script? If so >how? Erm, mount the DVD/CD, access files. Did you have something more complex in mind? Cheers, Phil From miller_duane at bah.com Tue Aug 9 17:56:42 2005 From: miller_duane at bah.com (Duane Miller) Date: Tue, 09 Aug 2005 10:56:42 -0700 Subject: Accessing the CD from %post In-Reply-To: References: <42F8E9C6.3010103@bah.com> Message-ID: <42F8EE5A.1040506@bah.com> Sounds easy enough. But isn't it already mounted? You're installing from it. D Philip Rowlands wrote: >On Tue, 9 Aug 2005, Duane Miller wrote: > > > >>I built a DVD to install from with kickstart file built in. >> >>One more thing I need to do: >> >>Can I access the files on the DVD from my %post or %pre script? If so >>how? >> >> > >Erm, mount the DVD/CD, access files. Did you have something more complex >in mind? > > >Cheers, >Phil > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > > From clumens at redhat.com Tue Aug 9 19:53:34 2005 From: clumens at redhat.com (Chris Lumens) Date: Tue, 9 Aug 2005 15:53:34 -0400 (EDT) Subject: Getting install log on a remote machine In-Reply-To: References: Message-ID: > I want to get the install time log on a remote machine, so that if I am > away from the console of the machine being installed I can still have > access to the anaconda logs of the machine being installed. > > Is there a way to do this in kickstart pre install script (asking > anaconda to send logs to a network log server) ? Not yet. I just started working on all the logging support in anaconda today. The goal is to convert it to using Python's logging class, which will allow us to log to files, syslog running on a remote machine, mail, web servers, sockets, etc. I don't know if all those mechanisms will make the cut for what we'll support, but at least one network mechanism will be supported. I guess now's as good a time as any to ask if people have any particular features, annoyances, or comments on logging. - Chris From klaus.steden at thomson.net Tue Aug 9 20:11:17 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 9 Aug 2005 16:11:17 -0400 Subject: Getting install log on a remote machine In-Reply-To: References: Message-ID: <20050809201117.GX897@thomson.net> > >I want to get the install time log on a remote machine, so that if I am > >away from the console of the machine being installed I can still have > >access to the anaconda logs of the machine being installed. > > > >Is there a way to do this in kickstart pre install script (asking > >anaconda to send logs to a network log server) ? > > Not yet. I just started working on all the logging support in anaconda > today. The goal is to convert it to using Python's logging class, which > will allow us to log to files, syslog running on a remote machine, mail, > web servers, sockets, etc. I don't know if all those mechanisms will > make the cut for what we'll support, but at least one network mechanism > will be supported. > > I guess now's as good a time as any to ask if people have any particular > features, annoyances, or comments on logging. > I'd be happy to help out, if you're looking; I have been meaning to implement something like this on my own, but a little hesitant to jump right into something I'm still feeling my way around. Syslogd support I would think would be essential ... I don't know about any of the others, for my part. Klaus From cmalek at caltech.edu Tue Aug 9 20:22:28 2005 From: cmalek at caltech.edu (Chris Malek) Date: Tue, 09 Aug 2005 13:22:28 -0700 Subject: Getting install log on a remote machine In-Reply-To: <20050809201117.GX897@thomson.net> References: <20050809201117.GX897@thomson.net> Message-ID: <42F91084.7090001@caltech.edu> Klaus Steden wrote: > Syslogd support I would think would be essential ... I don't know about any of > the others, for my part. I second that. This is something I've been meaning to look into, when time and tasks permit. Chris -- cmalek at caltech.edu Office: (626) 395-4073 Nosce Teipsum Fax: (626) 405-9851 From herrold at owlriver.com Tue Aug 9 20:57:16 2005 From: herrold at owlriver.com (R P Herrold) Date: Tue, 9 Aug 2005 16:57:16 -0400 (EDT) Subject: ks] Re: Getting install log on a remote machine In-Reply-To: References: Message-ID: On Tue, 9 Aug 2005, Chris Lumens wrote: > Not yet. I just started working on all the logging support in anaconda > today. The goal is to convert it to using Python's logging class, which > will allow us to log to files, syslog running on a remote machine, mail, > web servers, sockets, etc. I don't know if all those mechanisms will > make the cut for what we'll support, but at least one network mechanism > will be supported. > > I guess now's as good a time as any to ask if people have any particular > features, annoyances, or comments on logging. Assumedly, you will need to pass out a logserver IP or name, and append it to a working /etc/syslog.conf, an (re)start the syslogd. I have seen parameters to pass out timeservers in DHCP; I assume we can do logservers in like fashion (DHCP options 004 or 042 for timeservers) -- It seems to be: Option code 7 (Log Server Option). per this documentation: http://www.cisco.com/warp/public/109/dhcp_and_docsis_config_cm.html but I have not seen DHCP provided logservers documented in Linux space, nor tested this. (I will look into it) In testing, the option line: option log-servers 10.16.1.253 ; seems to work as a matter of syntax for my local /etc/dhcpd.conf , and I assume the Option can be extracted in a Python application to build the syslog destination IP. - Russ Herrold From jim at rossberry.com Tue Aug 9 20:58:51 2005 From: jim at rossberry.com (Jim Wildman) Date: Tue, 9 Aug 2005 16:58:51 -0400 (EDT) Subject: Getting install log on a remote machine In-Reply-To: References: Message-ID: > > Not yet. I just started working on all the logging support in anaconda > today. The goal is to convert it to using Python's logging class, which > will allow us to log to files, syslog running on a remote machine, mail, > web servers, sockets, etc. I don't know if all those mechanisms will > make the cut for what we'll support, but at least one network mechanism > will be supported. > > I guess now's as good a time as any to ask if people have any particular > features, annoyances, or comments on logging. > > - Chris When enabled, it should throw everything possible down the pipe. Let syslog on the other end sort it out. I would love to 'tail -f /var/log/kickstart' on my syslog server and watch the install progress. This would be really great for headless pxe installs. ------------------------------------------------------------------------ Jim Wildman, CISSP, RHCE jim at rossberry.com http://www.rossberry.com "Society in every state is a blessing, but Government, even in its best state, is a necessary evil; in its worst state, an intolerable one." Thomas Paine From clumens at redhat.com Tue Aug 9 21:15:34 2005 From: clumens at redhat.com (Chris Lumens) Date: Tue, 9 Aug 2005 17:15:34 -0400 (EDT) Subject: Getting install log on a remote machine In-Reply-To: References: Message-ID: > When enabled, it should throw everything possible down the pipe. Let > syslog on the other end sort it out. I would love to 'tail -f > /var/log/kickstart' on my syslog server and watch the install progress. > This would be really great for headless pxe installs. You'll get an option to specify your logging level, with the default set to something in the middle. Half the point of reworking the logging support is to make it less verbose so the average user doesn't freak out seeing tons of debugging output fly by and think they have a problem. I guess for you guys who will be enabling network logging support, it'd make sense to assume you won't be confused and crank the default output up a bit. - Chris From phr at doc.ic.ac.uk Tue Aug 9 23:32:45 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 10 Aug 2005 00:32:45 +0100 (BST) Subject: Accessing the CD from %post In-Reply-To: <42F8EE5A.1040506@bah.com> References: <42F8E9C6.3010103@bah.com> <42F8EE5A.1040506@bah.com> Message-ID: On Tue, 9 Aug 2005, Duane Miller wrote: >Sounds easy enough. But isn't it already mounted? You're installing >from it. Depends how we define "installing from". The anaconda first stage, anaconda second stage and Fedora/RPMS directory can all live on difference media. Because the latter comes from CD doesn't necessarily imply the rest do, nor that the CD is left mounted in %post. For added safety, why not put a check in the script? Cheers, Phil From klaus.steden at thomson.net Wed Aug 10 01:35:01 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Tue, 9 Aug 2005 21:35:01 -0400 Subject: ks] Re: Getting install log on a remote machine In-Reply-To: References: Message-ID: <20050810013501.GA897@thomson.net> > On Tue, 9 Aug 2005, Chris Lumens wrote: > > >Not yet. I just started working on all the logging support in anaconda > >today. The goal is to convert it to using Python's logging class, which > >will allow us to log to files, syslog running on a remote machine, mail, > >web servers, sockets, etc. I don't know if all those mechanisms will > >make the cut for what we'll support, but at least one network mechanism > >will be supported. > > > >I guess now's as good a time as any to ask if people have any particular > >features, annoyances, or comments on logging. > > Assumedly, you will need to pass out a logserver IP or name, > and append it to a working /etc/syslog.conf, an > (re)start the syslogd. > > I have seen parameters to pass out timeservers in DHCP; I > assume we can do logservers in like fashion (DHCP options 004 > or 042 for timeservers) -- It seems to be: > Option code 7 (Log Server Option). > per this documentation: > http://www.cisco.com/warp/public/109/dhcp_and_docsis_config_cm.html > but I have not seen DHCP provided logservers documented in > Linux space, nor tested this. (I will look into it) > > In testing, the option line: > option log-servers 10.16.1.253 ; > seems to work as a matter of syntax for my local > /etc/dhcpd.conf , and I assume the Option can be extracted in > a Python application to build the syslog destination IP. > I've seen it provided by both ISC and official (in this case, SGI) vendors, so I think it's a good bet that a syslog server name can be provided during lease acquisition. However, it is entirely possible and indeed workable, I think, to simply send the raw syslog message using sockets and avoid a syslog API altogether, leaving the remote log server to make sense of it (which has already been mentioned by someone else, and is I think the default behaviour of syslog - but I'm wiling to be corrected). But smarter heads than mine will decide. Klaus From gobbledegeek at gmail.com Wed Aug 10 05:23:13 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Wed, 10 Aug 2005 10:53:13 +0530 Subject: ks.cfg crashes install In-Reply-To: References: <463aea5705080423134b8df31@mail.gmail.com> Message-ID: <463aea570508092223210f17f3@mail.gmail.com> Its a custom single cd fc3 distro. i think langsupport is ok with fc3 kickstart? Rgrds On 8/8/05, Tom Diehl wrote: > On Mon, 8 Aug 2005, Chris Lumens wrote: > > > > langsupport --default=en_US.UTF-8 en_US.UTF-8 > > > > Is this an FC4 or a Rawhide install? The langsupport keyword has been > > removed from kickstart in Rawhide and will cause an error to be printed > > on the screen explaining how it's deprecated. > > In addition fc4 kickstart is B0rked!! See the archives of this list or > bugzilla for more info. I do not have the bug handy but there is a simple > workaround. > > > > > > firstboot -- > > > > This might be it. Is there supposed to be anything after the -- that > > got mangled in your post? > > > > - Chris > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > Tom Diehl tdiehl at rogueind.com Spamtrap address mtd123 at rogueind.com > -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From gobbledegeek at gmail.com Wed Aug 10 05:26:23 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Wed, 10 Aug 2005 10:56:23 +0530 Subject: ks.cfg crashes install In-Reply-To: <20050809160025.D8FA6739AE@hormel.redhat.com> References: <20050809160025.D8FA6739AE@hormel.redhat.com> Message-ID: <463aea5705080922261d6120c1@mail.gmail.com> Well The crash doesn't give me time to switch screens with ALT-F3. It promptly reboots. Screen Blanking is the last thing. Rgrds > > ---------- Forwarded message ---------- > From: Bill Schoolcraft > To: Discussion list about Kickstart > Date: Mon, 8 Aug 2005 09:23:24 -0700 (PDT) > Subject: Re: ks.cfg crashes install > At Mon, 8 Aug 2005 it looks like Philip Rowlands composed: > > > On Mon, 8 Aug 2005, Gobbledegeek wrote: > > > >> Well in my case, anaconda crashes before first screen - screen blanks > >> immediately on crash. Works ok without ks. Is the sequence of commands > >> important in ks.cfg? > > > > %packages should follow anaconda "commands"; %post should follow > > %packages. Within these sections, order is irrelevent. > > > > What error message do you see when anaconda crashes. What's the very > > last thing to happen/print before the crash? > > > > Hello Family, > > I believe, from my experience, that if you go and look at the > console on screen + it will show you the stdout of what > command was being issued when failure occurred and the previous > successful commands. It's like a running "tail -f" on > /var/log/messages > > Hope that helps. > > -- > Bill Schoolcraft > PO Box 210076 > San Francisco, CA 94121 > http://billschoolcraft.com > > > > > > ---------- Forwarded message ---------- > From: "Chowdhury, Chandan Dutta" > To: "Discussion list about Kickstart" > Date: Tue, 9 Aug 2005 14:54:51 +0530 > Subject: Getting install log on a remote machine > Hello All, > > I want to get the install time log on a remote machine, so that if I am > away from the console of the machine being installed I can still have > access to the anaconda logs of the machine being installed. > > Is there a way to do this in kickstart pre install script (asking > anaconda to send logs to a network log server) ? > > > > Thanks in advance > > Chandan > > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From gobbledegeek at gmail.com Wed Aug 10 05:33:35 2005 From: gobbledegeek at gmail.com (Gobbledegeek) Date: Wed, 10 Aug 2005 11:03:35 +0530 Subject: Kickstart-list Digest, Vol 18, Issue 6 In-Reply-To: <20050808160025.8CD8173959@hormel.redhat.com> References: <20050808160025.8CD8173959@hormel.redhat.com> Message-ID: <463aea5705080922334a21c418@mail.gmail.com> I'll check on the firstboot -- . I think I copied it from kickstart help document in anaconda sources though. Rgrds > > > firstboot -- > > This might be it. Is there supposed to be anything after the -- that > got mangled in your post? > > - Chris > > > -- Nonchalantly yours GobbledeGeek [Every thing but Gobbledegook.. !!] From chandan-dutta.chowdhury at hp.com Wed Aug 10 05:47:59 2005 From: chandan-dutta.chowdhury at hp.com (Chowdhury, Chandan Dutta) Date: Wed, 10 Aug 2005 11:17:59 +0530 Subject: Getting install log on a remote machine Message-ID: -----Original Message----- From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of Klaus Steden Sent: Wednesday, August 10, 2005 1:41 AM To: Discussion list about Kickstart Subject: Re: Getting install log on a remote machine > >I want to get the install time log on a remote machine, so that if I > >am away from the console of the machine being installed I can still > >have access to the anaconda logs of the machine being installed. > > > >Is there a way to do this in kickstart pre install script (asking > >anaconda to send logs to a network log server) ? > > Not yet. I just started working on all the logging support in > anaconda today. The goal is to convert it to using Python's logging > class, which will allow us to log to files, syslog running on a remote > machine, mail, web servers, sockets, etc. I don't know if all those > mechanisms will make the cut for what we'll support, but at least one > network mechanism will be supported. > > I guess now's as good a time as any to ask if people have any > particular features, annoyances, or comments on logging. > I'd be happy to help out, if you're looking; I have been meaning to implement something like this on my own, but a little hesitant to jump right into something I'm still feeling my way around. Syslogd support I would think would be essential ... I don't know about any of the others, for my part. Klaus Hello All, Thank you all for all your responses. My concern is if it is done at the anaconda level (ofcource that is the right place to do it) how will I be able to get the same feature on the previous versions(rhel2.1 or rhel3 etc). I was looking for may be a hack, some script which I can put in the pre install part of the kickstart so that it works for all the distributions, if you have any suggestion on this , plz help me out. I tried running a webserver using python on /tmp in the install machine. Unfortunately I dint find the BaseHTTPServer.py in the runtime env of the machine being installed. As for the modifications to anaconda, it will be great to have the remote logging feature in anaconda. Thanks Chandan From ajaymulwani at gmail.com Wed Aug 10 10:10:53 2005 From: ajaymulwani at gmail.com (Ajay Mulwani) Date: Wed, 10 Aug 2005 15:40:53 +0530 Subject: Accessing the CD from %post In-Reply-To: <42F8EE5A.1040506@bah.com> References: <42F8E9C6.3010103@bah.com> <42F8EE5A.1040506@bah.com> Message-ID: <5520b2790508100310571adf50@mail.gmail.com> > Sounds easy enough. But isn't it already mounted? You're installing > from it. But once the install is over, I don't think it is mounted anymore. Try to add these in ks.cfg ### %post /usr/sbin/updfstab /bin/mount /mnt/cdrom ### The DVD/CD should be accessible now at /mnt/cdrom. Ajay -------------- next part -------------- An HTML attachment was scrubbed... URL: From susanta_p at vfortress.com Fri Aug 12 10:33:38 2005 From: susanta_p at vfortress.com (Susant Padhi) Date: Fri, 12 Aug 2005 16:03:38 +0530 Subject: kiskstart installation does not asks for network configuration Message-ID: <1123842818.17303.6.camel@localhost.localdomain> Hi all, Following is my kickstart file where I am not written for network configuration parameter. When I am using this kickstart file, the anaconda skips the network configuration screen. What I am doing wrong here? ------------------ KICKSTART FILE ------------------- xconfig --depth=16 --resolution=800x600 --defaultdesktop=GNOME --startxonboot #Package install information %packages --resolvedeps @ gnome-desktop @ dialup @ base-x @ sql-server @ editors @ printing @ web-server @ graphical-internet grub kernel e2fsprogs %post if [ -z "` grep ServerFlags /etc/X11/xorg.conf`" >/dev/null 2>&1 ]; then cat /usr/bin/x.conf >> /etc/X11/xorg.conf fi rm /usr/bin/x.conf chkconfig --level 345 httpd on chkconfig --level 345 mysqld on chkconfig --level 345 xinetd on Please help. regards Susant From gotosathya at gmail.com Wed Aug 10 11:38:16 2005 From: gotosathya at gmail.com (Sathya M) Date: Wed, 10 Aug 2005 17:08:16 +0530 Subject: configuration to mailbox In-Reply-To: <1123842818.17303.6.camel@localhost.localdomain> References: <1123842818.17303.6.camel@localhost.localdomain> Message-ID: Hi all, I have doubt in /redhat linux 9.0, here is my requriements 1) Need to configure or create GLOBAL ADDRESS LIST for my users. if is it possible then how?? 2) copy of all mails should forward to a common mailbox Thanks in advance. Regards, Sathya From clumens at redhat.com Wed Aug 10 15:33:54 2005 From: clumens at redhat.com (Chris Lumens) Date: Wed, 10 Aug 2005 11:33:54 -0400 (EDT) Subject: ks] Re: Getting install log on a remote machine In-Reply-To: <20050810013501.GA897@thomson.net> References: <20050810013501.GA897@thomson.net> Message-ID: > However, it is entirely possible and indeed workable, I think, to simply send > the raw syslog message using sockets and avoid a syslog API altogether, > leaving the remote log server to make sense of it (which has already been > mentioned by someone else, and is I think the default behaviour of syslog - > but I'm wiling to be corrected). But smarter heads than mine will decide. I think a good case can be made for both. The syslog interface gives you a quick way to get remote logging up and running because it doesn't require any special programming. On the other hand, having a socket interface gives you the ability to write whatever crazy program you want to process the log information. On the other hand, I don't want to make the boot parameters too complicated. There's already too many of them and some are pretty goofy. - Chris From herrold at owlriver.com Wed Aug 10 15:56:29 2005 From: herrold at owlriver.com (R P Herrold) Date: Wed, 10 Aug 2005 11:56:29 -0400 (EDT) Subject: Getting install log on a remote machine In-Reply-To: References: <20050810013501.GA897@thomson.net> Message-ID: On Wed, 10 Aug 2005, Chris Lumens wrote: > I think a good case can be made for both. The syslog interface gives > you a quick way to get remote logging up and running because it doesn't > require any special programming. On the other hand, having a socket > interface gives you the ability to write whatever crazy program you want > to process the log information. The fact that one is quick and based on a longtime *nix standard approach, and the other is for use by a (new) 'crazy program' says something to me. -- Russ Herrold Those who do not understand Unix are condemned to reinvent it, poorly. -- Henry Spencer From daniel.spratlen at cox.net Wed Aug 10 16:06:03 2005 From: daniel.spratlen at cox.net (Daniel Spratlen) Date: Wed, 10 Aug 2005 12:06:03 -0400 Subject: ks] Re: Getting install log on a remote machine In-Reply-To: References: <20050810013501.GA897@thomson.net> Message-ID: <1123689963.4296.10.camel@boo> On Wed, 2005-08-10 at 11:33 -0400, Chris Lumens wrote: > > However, it is entirely possible and indeed workable, I think, to simply send > > the raw syslog message using sockets and avoid a syslog API altogether, > > leaving the remote log server to make sense of it (which has already been > > mentioned by someone else, and is I think the default behaviour of syslog - > > but I'm wiling to be corrected). But smarter heads than mine will decide. > > I think a good case can be made for both. The syslog interface gives > you a quick way to get remote logging up and running because it doesn't > require any special programming. On the other hand, having a socket > interface gives you the ability to write whatever crazy program you want > to process the log information. > > On the other hand, I don't want to make the boot parameters too > complicated. There's already too many of them and some are pretty > goofy. > > - Chris I agree, using syslog to accomplish remote logging would be extremely nice. Having other methods may be useful, but in environments with large installs of *nix boxes, dedicated syslog servers are common already. As for getting the syslog option via dhcp, that would also be very nice, however it should be an available option in the kickstart configuration file as well. In addition to allowing logging of an install, it'd be nice if the same kickstart configuration variable could be used to specify the ip of the syslog server to be used after the install is done. Sure you can do this fairly easily with a post script, but it'd be one less thing to worry about. thanks, Daniel From klaus.steden at thomson.net Wed Aug 10 16:36:56 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 10 Aug 2005 12:36:56 -0400 Subject: ks] Re: Getting install log on a remote machine In-Reply-To: References: <20050810013501.GA897@thomson.net> Message-ID: <20050810163655.GC897@thomson.net> > >However, it is entirely possible and indeed workable, I think, to simply > >send the raw syslog message using sockets and avoid a syslog API > >altogether, leaving the remote log server to make sense of it (which has > >already been mentioned by someone else, and is I think the default > >behaviour of syslog - but I'm wiling to be corrected). But smarter heads > >than mine will decide. > > I think a good case can be made for both. The syslog interface gives you a > quick way to get remote logging up and running because it doesn't require > any special programming. On the other hand, having a socket interface gives > you the ability to write whatever crazy program you want to process the log > information. > > On the other hand, I don't want to make the boot parameters too > complicated. There's already too many of them and some are pretty goofy. > What about a dual-pronged approach? The first being updating Anaconda to support the syslog API, the second being to provide a 'logger' command in bootstrap that can be called by kickstart itself, like the one used on FBSD, like this: logger -t kickstart -p local7.warn @loghost "Could not find /install/foobar.rpm to install!" or something like that. Klaus From herrold at owlriver.com Wed Aug 10 17:03:51 2005 From: herrold at owlriver.com (R P Herrold) Date: Wed, 10 Aug 2005 13:03:51 -0400 (EDT) Subject: Getting install log on a remote machine In-Reply-To: <20050810163655.GC897@thomson.net> References: <20050810013501.GA897@thomson.net> <20050810163655.GC897@thomson.net> Message-ID: On Wed, 10 Aug 2005, Klaus Steden wrote: > What about a dual-pronged approach? The first being updating Anaconda to > support the syslog API, the second being to provide a 'logger' command in > bootstrap that can be called by kickstart itself, like the one used on FBSD, > like this: > > logger -t kickstart -p local7.warn @loghost \ > "Could not find /install/foobar.rpm to install!" > > or something like that. I do not see the '@loghost' option in the man page of the util-linux-2.12a-16 derived 'logger', nor does it work locally. Interesting concept to be able to specify the syslog server directly in an end user accessible command -- remote DoS attack vectors spring unbidden to mind. - Russ Herrold From phr at doc.ic.ac.uk Wed Aug 10 17:36:47 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 10 Aug 2005 18:36:47 +0100 (BST) Subject: Getting install log on a remote machine In-Reply-To: References: <20050810013501.GA897@thomson.net> <20050810163655.GC897@thomson.net> Message-ID: Lots of interesting ideas on this thread, but in case people aren't aware of it can I also suggest the anaconda-devel list: https://www.redhat.com/mailman/listinfo/anaconda-devel-list "anaconda-devel-list is a list for the discussion of the Red Hat Linux installer." This might be appropriate in the wider context of anaconda usage, not just with kickstart files. Cheers, Phil From klaus.steden at thomson.net Wed Aug 10 17:49:53 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 10 Aug 2005 13:49:53 -0400 Subject: Getting install log on a remote machine In-Reply-To: References: <20050810013501.GA897@thomson.net> <20050810163655.GC897@thomson.net> Message-ID: <20050810174953.GD897@thomson.net> > >What about a dual-pronged approach? The first being updating Anaconda to > >support the syslog API, the second being to provide a 'logger' command in > >bootstrap that can be called by kickstart itself, like the one used on > >FBSD, > >like this: > > > > logger -t kickstart -p local7.warn @loghost \ > > "Could not find /install/foobar.rpm to install!" > > > >or something like that. > > I do not see the '@loghost' option in the man page of the > util-linux-2.12a-16 derived 'logger', nor does it work > locally. > > Interesting concept to be able to specify the syslog server > directly in an end user accessible command -- remote DoS > attack vectors spring unbidden to mind. > This is a vulnerability with any syslog server, or in fact, any server connected to a network that syslogs network requests. I could just as easily DoS the system by attacking another service. Syslogd should be defensive in anticipation of this, and indeed, on the three systems I justed checked (FreeBSD, IRIX, and Linux), all have a switch that can be used to disable logging of remotely submitted syslog messages. In the case of systems not directly connected to hostile networks, this scenario is more of a minor nuisance than a major security vulnerability. Klaus From info at hostinthebox.net Wed Aug 10 19:17:41 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 10 Aug 2005 13:17:41 -0600 Subject: kiskstart installation does not asks for network configuration In-Reply-To: <1123842818.17303.6.camel@localhost.localdomain> References: <1123842818.17303.6.camel@localhost.localdomain> Message-ID: <42FA52D5.3000700@hostinthebox.net> Susant Padhi wrote: > Hi all, > > Following is my kickstart file where I am not written for network > configuration parameter. When I am using this kickstart file, the > anaconda skips the network configuration screen. What I am doing wrong > here? > > > ------------------ > KICKSTART FILE > ------------------- > > xconfig --depth=16 --resolution=800x600 --defaultdesktop=GNOME > --startxonboot > #Package install information > %packages --resolvedeps > @ gnome-desktop > @ dialup > @ base-x > @ sql-server > @ editors > @ printing > @ web-server > @ graphical-internet > grub > kernel > e2fsprogs > %post > if [ -z "` grep ServerFlags /etc/X11/xorg.conf`" >/dev/null 2>&1 ]; > then > cat /usr/bin/x.conf >> /etc/X11/xorg.conf > fi > rm /usr/bin/x.conf > chkconfig --level 345 httpd on > chkconfig --level 345 mysqld on > chkconfig --level 345 xinetd on > > Please help. > > regards > Susant > Susant - As far as I understand Kicksart (and I may be wrong here), it is meant to be completely automated. The whole idea surrounding a Kickstart install is that there is no outside interaction needed in order to get the install done, after initial configuration of said kickstart file and friends. That being the case, I don't think that you can break in the middle of a kickstart for configuration of something such as this. However, you can automate the process by taking a look at: http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-options.html ...with emphasis on the 'network' directive. Hope that helps -dant From miller_duane at bah.com Wed Aug 10 19:52:15 2005 From: miller_duane at bah.com (Duane Miller) Date: Wed, 10 Aug 2005 12:52:15 -0700 Subject: kiskstart installation does not asks for network configuration In-Reply-To: <42FA52D5.3000700@hostinthebox.net> References: <1123842818.17303.6.camel@localhost.localdomain> <42FA52D5.3000700@hostinthebox.net> Message-ID: <42FA5AEF.70304@bah.com> Why only one line in the top section? And it's not complete. Maybe it doesn't like the unknown video card. What does anaconda-ks.cfg look like when you do a manual install? If it only skips the network configuration screen, it's ususally because the network card is not recognized. Does a manual install recognize the card? D Dan Trainor wrote: >Susant Padhi wrote: > > >>Hi all, >> >>Following is my kickstart file where I am not written for network >>configuration parameter. When I am using this kickstart file, the >>anaconda skips the network configuration screen. What I am doing wrong >>here? >> >> >>------------------ >>KICKSTART FILE >>------------------- >> >>xconfig --depth=16 --resolution=800x600 --defaultdesktop=GNOME >>--startxonboot >>#Package install information >>%packages --resolvedeps >>@ gnome-desktop >>@ dialup >>@ base-x >>@ sql-server >>@ editors >>@ printing >>@ web-server >>@ graphical-internet >>grub >>kernel >>e2fsprogs >>%post >>if [ -z "` grep ServerFlags /etc/X11/xorg.conf`" >/dev/null 2>&1 ]; >>then >> cat /usr/bin/x.conf >> /etc/X11/xorg.conf >>fi >>rm /usr/bin/x.conf >>chkconfig --level 345 httpd on >>chkconfig --level 345 mysqld on >>chkconfig --level 345 xinetd on >> >>Please help. >> >>regards >>Susant >> >> >> > >Susant - > >As far as I understand Kicksart (and I may be wrong here), it is meant >to be completely automated. The whole idea surrounding a Kickstart >install is that there is no outside interaction needed in order to get >the install done, after initial configuration of said kickstart file and >friends. > >That being the case, I don't think that you can break in the middle of a >kickstart for configuration of something such as this. However, you can >automate the process by taking a look at: > >http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/sysadmin-guide/s1-kickstart2-options.html > >...with emphasis on the 'network' directive. > >Hope that helps >-dant > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > > From Christian.Rohrmeier at SCHERING.DE Thu Aug 11 15:26:40 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 11 Aug 2005 17:26:40 +0200 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: Message-ID: Hello Dear Kickstarters, Currently I support RHEL 2.1, 3, and 4, WS and AS, and X86 and X86_64 version (except for RHEL 2.1). For each of these I have a kickstart script and an install tree, and for each of these I have a boot CD to install from. We install by boot from a RHEL 3 WS X86 Update 5 CD, for example, and entering "linux ks=http://install.server.name/install/rhel3wsx86". Recently I was asked not just to support the latest version of each of those RHEL versions (RHEL 2.1 Update 6, RHEL 3 Update 5, RHEL 4 Update 1), but to also support older updates. This because some software, such as IBM Websphere, is only supported on specific update releases (for example RHEL 3 AS Update 1 and Update 3). If I were to implement this support via my current methods, I calculate: RHEL 2.1 X 2 Version (WS, AS) X 7 Updates = 14 CDs, 14 install trees, and 14 kickstart scripts (each one has to have URL= pointing to the right install tree) RHEL 3 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 6 Updates = 24 CDs, 24 install trees, and 24 kickstart scripts (each one has to have URL= pointing to the right install tree) RHEL 4 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 2 Updates = 8 CDs, 8 install trees, and 8 kickstart scripts (each one has to have URL= pointing to the right install tree) I would have to support 46 kickstart scripts, install CDs, and install trees. Perhaps getting around all those install trees is not possible, but disk space is cheap, so thats not the problem. However, having to make 46 boot CDs available for installation, and making 46 kickstart scripts, that is just horrible. Issues: 1) I have found that installing from an RHEL 3 AS CD, for example, doesn't allow you to install from a RHEL 3 WS tree (it complains the CD doesn't match the files in the tree.) The same goes for trying to install a RHEL 3 WS Update 5 tree by booting a RHEL 3 WS Update 4 CD, for example. It seems therefor that you can't make one Installer CD that allows you to install different releases of RHEL. 2) Unifying kickstart scripts would be just grand. There is no difference in the disk partitioning, package install, or other confguration between RHEL 3 AS or WS (well, not for me anyway.) But I have to specify the installation media in the kickstart script via "URL=". Perhaps in RHEL 3 and 4 (not in 2.1) this could be taken care of by specifying a release number at the boot: prompt and reading the boot: prompt options in %pre and then using %include to set the proper URL= line. Problem 1 seems to be the worst of the two. If anyone has dealt with the same issues, I would be very interested to hear about your solutions. Thanks for any ideas, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 From brilong at cisco.com Thu Aug 11 16:06:28 2005 From: brilong at cisco.com (Brian Long) Date: Thu, 11 Aug 2005 12:06:28 -0400 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: References: Message-ID: <1123776389.3824.63.camel@brilong-lnx.cisco.com> On Thu, 2005-08-11 at 17:26 +0200, Christian.Rohrmeier at SCHERING.DE wrote: > Hello Dear Kickstarters, > I would have to support 46 kickstart scripts, install CDs, and install > trees. We've written and implemented a kickstart-tools web CGI which allows the end user to customize their boot disk by selecting a hardware and software profile and optionally some extra items like disk partitions, etc. It then creates a boot disk / driver disk combo, boot.iso or kickme shar file (which can be run on an existing Linux install to modify grub to kickstart the machine when rebooted). This is all available at http://kickstart-tools.sf.net. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From jason at jw2.org Thu Aug 11 21:55:31 2005 From: jason at jw2.org (Jason White) Date: Thu, 11 Aug 2005 17:55:31 -0400 Subject: Multiple LUNs (scsi_mod) and RHEL4 Message-ID: <20050811215531.GA27344@jw2.org> I'm trying to kickstart a machine, and partition disks on 3 fibre attached LUNs (lpfc/emulex), but the installer only sees the first one (/dev/sda), and not the other two (/dev/sdb, /dev/sdc) Out of the box, the RHEL4 initrd is configured to only discover a single scsi LUN. In order to fix this, you have to use a custom initrd: # echo "options scsi_mod max_luns=xx" >> /etc/modprobe.conf # mkinitrd -f # reboot (where xx > 1) This works fine *after* the box is installed, but is there a way to make this work during the kickstart install? I tried using some %pre scripts, but it didn't work: ---ks.cfg--- ... %pre rmmod scsi_mod modprobe scsi_mod max_luns=128 ... %post echo "options scsi_mod max_luns=255" >> /etc/modprobe.conf mkinitrd -f -v ---ks.cfg--- ---kickstart errors--- *Running kickstart %pre script(s) rmmod: scsi_mod: Resource temporarily unavailable modprobe: could no parse modules.dep * WARNING - Error code 256 encountered running a kickstart %pre/%post * script * All kickstart %pre script(s) have been run * Exception parsing ks.cfg: specified disk sdb in partition command does not exist ---kickstart errors--- Thanks, Jason From jkeating at j2solutions.net Thu Aug 11 22:01:02 2005 From: jkeating at j2solutions.net (Jesse Keating) Date: Thu, 11 Aug 2005 15:01:02 -0700 Subject: Multiple LUNs (scsi_mod) and RHEL4 In-Reply-To: <20050811215531.GA27344@jw2.org> References: <20050811215531.GA27344@jw2.org> Message-ID: <1123797663.3379.8.camel@localhost.localdomain> On Thu, 2005-08-11 at 17:55 -0400, Jason White wrote: > This works fine *after* the box is installed, but is there a way to > make this work during the kickstart install? I tried using some %pre > scripts, but it didn't work: Add the line to your boot line when booting the installer, or in the kernel line when starting kickstart. -- Jesse Keating RHCE (http://geek.j2solutions.net) Fedora Legacy Team (http://www.fedoralegacy.org) GPG Public Key (http://geek.j2solutions.net/jkeating.j2solutions.pub) Was I helpful? Let others know: http://svcs.affero.net/rm.php?r=jkeating -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: From john.j.poole at usa-spaceops.com Thu Aug 11 22:40:44 2005 From: john.j.poole at usa-spaceops.com (Poole, John J) Date: Thu, 11 Aug 2005 18:40:44 -0400 Subject: Incorporating Anaconda fixes onto a KS boot CD Message-ID: <2CE66A1ABBED8C4B85DFF5E0860987A3051D6B41@usaflcms03.usa-spaceops.ksc.nasa.gov> Hi, I reload our machines via kickstart using both NFS and HTTP. I initially boot up on the RHEL Update 5 CDROM #1 to start the KS process. A recent Anaconda-9.1.5.8-1.RHEL.i386 has been released with a fix for e1000 Ethernet cards timing out during start of network kick start installations. What is the easiest way to incorporate the Anaconda release into my kickstart? I assume that I must put the Anaconda-9.1.5.8-1.RHEL.i386 onto the boot cd. I have been using the stock Update 5 install cd and have never had to build/rebuild a boot cd. Some simple instructions or pointers on incorporating the updated Anaconda/pump files into the Update 5 install cdrom would be appreciated. I am an Admin and not a programmer/developer. Thanks, John J. Poole United Space Alliance From klaus.steden at thomson.net Thu Aug 11 23:12:29 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 11 Aug 2005 19:12:29 -0400 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: References: Message-ID: <20050811231229.GR897@thomson.net> > > Hello Dear Kickstarters, > > Currently I support RHEL 2.1, 3, and 4, WS and AS, and X86 and X86_64 > version (except for RHEL 2.1). For each of these I have a kickstart script > and an install tree, and for each of these I have a boot CD to install > from. We install by boot from a RHEL 3 WS X86 Update 5 CD, for example, and > entering "linux ks=http://install.server.name/install/rhel3wsx86". > > Recently I was asked not just to support the latest version of each of > those RHEL versions (RHEL 2.1 Update 6, RHEL 3 Update 5, RHEL 4 Update 1), > but to also support older updates. This because some software, such as IBM > Websphere, is only supported on specific update releases (for example RHEL > 3 AS Update 1 and Update 3). > > If I were to implement this support via my current methods, I calculate: > > RHEL 2.1 X 2 Version (WS, AS) X 7 Updates = 14 CDs, 14 install trees, and > 14 kickstart scripts (each one has to have URL= pointing to the right > install tree) > RHEL 3 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 6 Updates = > 24 CDs, 24 install trees, and 24 kickstart scripts (each one has to have > URL= pointing to the right install tree) > RHEL 4 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 2 Updates = > 8 CDs, 8 install trees, and 8 kickstart scripts (each one has to have URL= > pointing to the right install tree) > > I would have to support 46 kickstart scripts, install CDs, and install > trees. > > Perhaps getting around all those install trees is not possible, but disk > space is cheap, so thats not the problem. However, having to make 46 boot > CDs available for installation, and making 46 kickstart scripts, that is > just horrible. > > Issues: > 1) I have found that installing from an RHEL 3 AS CD, for example, doesn't > allow you to install from a RHEL 3 WS tree (it complains the CD doesn't > match the files in the tree.) The same goes for trying to install a RHEL 3 > WS Update 5 tree by booting a RHEL 3 WS Update 4 CD, for example. It seems > therefor that you can't make one Installer CD that allows you to install > different releases of RHEL. > > 2) Unifying kickstart scripts would be just grand. There is no difference > in the disk partitioning, package install, or other confguration between > RHEL 3 AS or WS (well, not for me anyway.) But I have to specify the > installation media in the kickstart script via "URL=". Perhaps in RHEL 3 > and 4 (not in 2.1) this could be taken care of by specifying a release > number at the boot: prompt and reading the boot: prompt options in %pre and > then using %include to set the proper URL= line. > > Problem 1 seems to be the worst of the two. If anyone has dealt with the > same issues, I would be very interested to hear about your solutions. > > Thanks for any ideas, > Yikes. I thought I was looking at complexity with four machine types and three architectures ... but you my friend have quite an impressive task on your hands. I made some adjustments to the Anaconda/Kickstart process to get around this problem; I am in the process of putting together docs for the Anaconda Wiki explaining the how, what, and hopefully why of what I did ... but essentially, I end up with a bootstrap CD that allows me to do something like this: wkstn name= wkstn64 name= dev name= dev64 name= etc. Each of these stanzas has different information written into the isolinux.cfg instructing the bootstrap process to pick a different target and flavour. It would be theoretically possible to implement this relatively cleanly using a series of menus, each for a different OS flavour, which support the variety of arches you're using, but that's more something you'll have to customize to suit your needs. I will hopefully have a full explanation on the Wiki within the fortnight, and that may be of some help. hth, Klaus From ebrown at lanl.gov Fri Aug 12 01:54:05 2005 From: ebrown at lanl.gov (Edward F. Brown) Date: Thu, 11 Aug 2005 19:54:05 -0600 (MDT) Subject: Incorporating Anaconda fixes onto a KS boot CD In-Reply-To: <2CE66A1ABBED8C4B85DFF5E0860987A3051D6B41@usaflcms03.usa-spaceops.ksc. nasa.gov> References: <2CE66A1ABBED8C4B85DFF5E0860987A3051D6B41@usaflcms03.usa-spaceops.ksc.nasa.gov> Message-ID: <32944.128.165.7.86.1123811645.squirrel@webmail.lanl.gov> anaconda-9.1.5.8-1 IS the version included with update 5 (and it did not appear to fix the problem). Update 6, available in beta now, includes anaconda-9.1.6.7-1.RHEL. The changelog for this version includes: * Wed Jul 13 2005 Paul Nasrat - 9.1.6.3-1.RHEL - multiple NICs fixes for kickstart (clumens, #158556) so we can only hope, again, that this finally resolves the problem. In the meantime, filing a support ticket with RedHat is the best way of keeping this issue on their radar, though be prepared to spend more time beta-testing solutions. -Ed > Hi, > I reload our machines via kickstart using both NFS and HTTP. I > initially boot up on the RHEL Update 5 CDROM #1 to start the KS > process. A recent > Anaconda-9.1.5.8-1.RHEL.i386 has been released with a fix for e1000 > Ethernet cards timing out during start of network kick start > installations. > What is the easiest way to incorporate the Anaconda release into my > kickstart? I assume that I must put the Anaconda-9.1.5.8-1.RHEL.i386 > onto the boot > cd. I have been using the stock Update 5 install cd and have never had to > build/rebuild a boot cd. Some simple instructions or pointers on > incorporating the > updated Anaconda/pump files into the Update 5 install cdrom would be > appreciated. I am an Admin and not a programmer/developer. > Thanks, > > John J. Poole > United Space Alliance > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From jason at jw2.org Fri Aug 12 02:14:37 2005 From: jason at jw2.org (Jason White) Date: Thu, 11 Aug 2005 22:14:37 -0400 Subject: Multiple LUNs (scsi_mod) and RHEL4 In-Reply-To: <1123797663.3379.8.camel@localhost.localdomain> References: <20050811215531.GA27344@jw2.org> <1123797663.3379.8.camel@localhost.localdomain> Message-ID: <20050812021437.GA3195@jw2.org> * Jesse Keating (jkeating at j2solutions.net) [050811 18:01]: > On Thu, 2005-08-11 at 17:55 -0400, Jason White wrote: > > This works fine *after* the box is installed, but is there a way to > > make this work during the kickstart install? I tried using some %pre > > scripts, but it didn't work: > Add the line to your boot line when booting the installer, or in the > kernel line when starting kickstart. I tried both, but neither seem to work: Boot line: "boot: ks max_luns=256" Kernel line: ---isolinux.cfg--- label ks kernel vmlinuz append ks=http://myserver/ks.cfg ksdevice=eth0 initrd=initrd.img max_luns=256 ---isolinux.cfg--- I also tried putting the following line at the very beginning of my ks.cfg (before the %pre section): device scsi scsi_mod --opts="max_luns=256" Jason From erik.sjolund at home.se Fri Aug 12 06:40:22 2005 From: erik.sjolund at home.se (Erik =?ISO-8859-1?Q?Sj=F6lund?=) Date: Fri, 12 Aug 2005 08:40:22 +0200 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: References: Message-ID: <1123828822.11057.29.camel@otto> > If anyone has dealt with the > same issues, I would be very interested to hear about your solutions. > > Thanks for any ideas, You may want to take a look at http://xml2hostconf.sourceforge.net/ Among other things it generates kickstart scripts out of an XML file that describes your network. It is primarily meant to be used in situations where clients are installed over the network with PXE. Of course you could instead copy the generated kickstart scripts to another medium and bring it to the client computers by other means. A new release ( 2.0.0 ) was released yesterday. A new feature in this release is among others a cgi script in Python that makes the kickstart file only downloadable one single time from the correct ip address. If you want to avoid to always have your kickstart files published this might be a good thing. cheers, Erik Sj?lund From error27 at gmail.com Fri Aug 12 17:13:04 2005 From: error27 at gmail.com (Dan Carpenter) Date: Fri, 12 Aug 2005 10:13:04 -0700 Subject: Incorporating Anaconda fixes onto a KS boot CD In-Reply-To: <32944.128.165.7.86.1123811645.squirrel@webmail.lanl.gov> References: <2CE66A1ABBED8C4B85DFF5E0860987A3051D6B41@usaflcms03.usa-spaceops.ksc.nasa.gov> <32944.128.165.7.86.1123811645.squirrel@webmail.lanl.gov> Message-ID: I've been working on creating a "disk zero" that seperates out the kernel and installer from the packages so that you can make quick updates. Basically the plan is to just rm the RPM directory, change the diskinfo file to say it's cd 0. Copy in a new install kernel. Create a patchup/ directory with a script that will be run right before the %post. The patchup script will install an updated kernel, and udev etc. To install you just boot to disk zero, then it tells you to insert cd 1 for the packages install part. Anaconda already handles that bit. My plan is to get disk zero working for CD installs then afterwards work on network installs etc. I started last weekend and I'll try to work on it some more tomorrow. regards, dan From Christian.Rohrmeier at SCHERING.DE Mon Aug 15 08:23:09 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Mon, 15 Aug 2005 10:23:09 +0200 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: <20050811231229.GR897@thomson.net> Message-ID: Hi Klaus and the others that gave me your suggestions, please let me (us) know when you have the info on the Wiki. I think that would be a good start for how to be flexible when handling this type of mass kickstart scripting. Thanks, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Klaus Steden To Sent by: Discussion list about Kickstart kickstart-list-bo unces at redhat.com cc Subject 12.08.2005 01:12 Re: 46 kick start scripts, 46 install trees, 46 boot CDs Please respond to Discussion list about Kickstart > > Hello Dear Kickstarters, > > Currently I support RHEL 2.1, 3, and 4, WS and AS, and X86 and X86_64 > version (except for RHEL 2.1). For each of these I have a kickstart script > and an install tree, and for each of these I have a boot CD to install > from. We install by boot from a RHEL 3 WS X86 Update 5 CD, for example, and > entering "linux ks=http://install.server.name/install/rhel3wsx86". > > Recently I was asked not just to support the latest version of each of > those RHEL versions (RHEL 2.1 Update 6, RHEL 3 Update 5, RHEL 4 Update 1), > but to also support older updates. This because some software, such as IBM > Websphere, is only supported on specific update releases (for example RHEL > 3 AS Update 1 and Update 3). > > If I were to implement this support via my current methods, I calculate: > > RHEL 2.1 X 2 Version (WS, AS) X 7 Updates = 14 CDs, 14 install trees, and > 14 kickstart scripts (each one has to have URL= pointing to the right > install tree) > RHEL 3 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 6 Updates = > 24 CDs, 24 install trees, and 24 kickstart scripts (each one has to have > URL= pointing to the right install tree) > RHEL 4 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 2 Updates = > 8 CDs, 8 install trees, and 8 kickstart scripts (each one has to have URL= > pointing to the right install tree) > > I would have to support 46 kickstart scripts, install CDs, and install > trees. > > Perhaps getting around all those install trees is not possible, but disk > space is cheap, so thats not the problem. However, having to make 46 boot > CDs available for installation, and making 46 kickstart scripts, that is > just horrible. > > Issues: > 1) I have found that installing from an RHEL 3 AS CD, for example, doesn't > allow you to install from a RHEL 3 WS tree (it complains the CD doesn't > match the files in the tree.) The same goes for trying to install a RHEL 3 > WS Update 5 tree by booting a RHEL 3 WS Update 4 CD, for example. It seems > therefor that you can't make one Installer CD that allows you to install > different releases of RHEL. > > 2) Unifying kickstart scripts would be just grand. There is no difference > in the disk partitioning, package install, or other confguration between > RHEL 3 AS or WS (well, not for me anyway.) But I have to specify the > installation media in the kickstart script via "URL=". Perhaps in RHEL 3 > and 4 (not in 2.1) this could be taken care of by specifying a release > number at the boot: prompt and reading the boot: prompt options in %pre and > then using %include to set the proper URL= line. > > Problem 1 seems to be the worst of the two. If anyone has dealt with the > same issues, I would be very interested to hear about your solutions. > > Thanks for any ideas, > Yikes. I thought I was looking at complexity with four machine types and three architectures ... but you my friend have quite an impressive task on your hands. I made some adjustments to the Anaconda/Kickstart process to get around this problem; I am in the process of putting together docs for the Anaconda Wiki explaining the how, what, and hopefully why of what I did ... but essentially, I end up with a bootstrap CD that allows me to do something like this: wkstn name= wkstn64 name= dev name= dev64 name= etc. Each of these stanzas has different information written into the isolinux.cfg instructing the bootstrap process to pick a different target and flavour. It would be theoretically possible to implement this relatively cleanly using a series of menus, each for a different OS flavour, which support the variety of arches you're using, but that's more something you'll have to customize to suit your needs. I will hopefully have a full explanation on the Wiki within the fortnight, and that may be of some help. hth, Klaus _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Christian.Rohrmeier at SCHERING.DE Mon Aug 15 13:22:24 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Mon, 15 Aug 2005 15:22:24 +0200 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: Message-ID: Hi All, Just a note on some things I've been trying, and how they fail: - you can't use %include to dynamically specify the install media (url=). So I guess it doesn't run the %pre script before it checks for install media. Which is sad. - you can interchangeably use WS and AS install CDs (and I assume ES as well). Which is one boot CD less to distribute, so thats good. - you can't switch between Update releases on the install CDs and the install trees. Booting from RHEL 3 WS Update 4 won't let you install from anything else than an Update 4 tree. As Klaus suggested, bootstrapping the install CD and modifying iso.linux is the only way. Mit freundlichen Gr??en, Christian Rohrmeier _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Christian.Rohrmei er at schering.de Sent by: To kickstart-list-bo Discussion list about Kickstart unces at redhat.com cc Discussion list about Kickstart 15.08.2005 10:23 , kickstart-list-bounces at redhat.com Subject Please respond to Re: 46 kick start scripts, 46 Discussion list install trees, 46 boot CDs about Kickstart Hi Klaus and the others that gave me your suggestions, please let me (us) know when you have the info on the Wiki. I think that would be a good start for how to be flexible when handling this type of mass kickstart scripting. Thanks, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Klaus Steden To Sent by: Discussion list about Kickstart kickstart-list-bo unces at redhat.com cc Subject 12.08.2005 01:12 Re: 46 kick start scripts, 46 install trees, 46 boot CDs Please respond to Discussion list about Kickstart > > Hello Dear Kickstarters, > > Currently I support RHEL 2.1, 3, and 4, WS and AS, and X86 and X86_64 > version (except for RHEL 2.1). For each of these I have a kickstart script > and an install tree, and for each of these I have a boot CD to install > from. We install by boot from a RHEL 3 WS X86 Update 5 CD, for example, and > entering "linux ks=http://install.server.name/install/rhel3wsx86". > > Recently I was asked not just to support the latest version of each of > those RHEL versions (RHEL 2.1 Update 6, RHEL 3 Update 5, RHEL 4 Update 1), > but to also support older updates. This because some software, such as IBM > Websphere, is only supported on specific update releases (for example RHEL > 3 AS Update 1 and Update 3). > > If I were to implement this support via my current methods, I calculate: > > RHEL 2.1 X 2 Version (WS, AS) X 7 Updates = 14 CDs, 14 install trees, and > 14 kickstart scripts (each one has to have URL= pointing to the right > install tree) > RHEL 3 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 6 Updates = > 24 CDs, 24 install trees, and 24 kickstart scripts (each one has to have > URL= pointing to the right install tree) > RHEL 4 X 2 Versions (WS, AS) X 2 architectures (X86, X86_64) X 2 Updates = > 8 CDs, 8 install trees, and 8 kickstart scripts (each one has to have URL= > pointing to the right install tree) > > I would have to support 46 kickstart scripts, install CDs, and install > trees. > > Perhaps getting around all those install trees is not possible, but disk > space is cheap, so thats not the problem. However, having to make 46 boot > CDs available for installation, and making 46 kickstart scripts, that is > just horrible. > > Issues: > 1) I have found that installing from an RHEL 3 AS CD, for example, doesn't > allow you to install from a RHEL 3 WS tree (it complains the CD doesn't > match the files in the tree.) The same goes for trying to install a RHEL 3 > WS Update 5 tree by booting a RHEL 3 WS Update 4 CD, for example. It seems > therefor that you can't make one Installer CD that allows you to install > different releases of RHEL. > > 2) Unifying kickstart scripts would be just grand. There is no difference > in the disk partitioning, package install, or other confguration between > RHEL 3 AS or WS (well, not for me anyway.) But I have to specify the > installation media in the kickstart script via "URL=". Perhaps in RHEL 3 > and 4 (not in 2.1) this could be taken care of by specifying a release > number at the boot: prompt and reading the boot: prompt options in %pre and > then using %include to set the proper URL= line. > > Problem 1 seems to be the worst of the two. If anyone has dealt with the > same issues, I would be very interested to hear about your solutions. > > Thanks for any ideas, > Yikes. I thought I was looking at complexity with four machine types and three architectures ... but you my friend have quite an impressive task on your hands. I made some adjustments to the Anaconda/Kickstart process to get around this problem; I am in the process of putting together docs for the Anaconda Wiki explaining the how, what, and hopefully why of what I did ... but essentially, I end up with a bootstrap CD that allows me to do something like this: wkstn name= wkstn64 name= dev name= dev64 name= etc. Each of these stanzas has different information written into the isolinux.cfg instructing the bootstrap process to pick a different target and flavour. It would be theoretically possible to implement this relatively cleanly using a series of menus, each for a different OS flavour, which support the variety of arches you're using, but that's more something you'll have to customize to suit your needs. I will hopefully have a full explanation on the Wiki within the fortnight, and that may be of some help. hth, Klaus _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From Paul.Miles at quadriga.com Mon Aug 15 14:14:20 2005 From: Paul.Miles at quadriga.com (Miles, Paul) Date: Mon, 15 Aug 2005 15:14:20 +0100 Subject: Custom dialog messages Message-ID: <5D339AE687B0C8438185AE62ED06697D7A8105@ukchmail01.mail.quadriga.com> Hello, I want to display custom messages in anaconda's PRE and POST section. Currently I'm using a statically compiled version of dialog to display these. Annoyingly Dialog is being rendered using '+' and '-' signs rather than drawing a proper box. Any ideas how to solve this ? Thanks Paul This e-mail is the property of Quadriga Worldwide Ltd The message (and any associated files) is intended only for the use of the individual or entity to which it is addressed and may contain information that is confidential, subject to copyright or constitutes a trade secret. If you are not the intended recipient you are hereby notified that any dissemination, copying or distribution of this message, or files associated with this message, is strictly prohibited. If you have received this message in error, please notify us immediately by replying to the message and deleting it from your computer. Messages sent to and from us may be monitored. Internet communications cannot be guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain viruses. Therefore, we do not accept responsibility for any errors or omissions that are present in this message, or any attachment, that have arisen as a result of e-mail transmission. If verification is required, please request a hard-copy version. Any views or opinions presented are solely those of the author and do not necessarily represent those of the company. From phr at doc.ic.ac.uk Mon Aug 15 14:32:44 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Mon, 15 Aug 2005 15:32:44 +0100 (BST) Subject: Custom dialog messages In-Reply-To: <5D339AE687B0C8438185AE62ED06697D7A8105@ukchmail01.mail.quadriga.com> References: <5D339AE687B0C8438185AE62ED06697D7A8105@ukchmail01.mail.quadriga.com> Message-ID: On Mon, 15 Aug 2005, Miles, Paul wrote: >I want to display custom messages in anaconda's PRE and POST section. > >Currently I'm using a statically compiled version of dialog to display >these. Annoyingly Dialog is being rendered using '+' and '-' signs >rather than drawing a proper box. Looks like the TERM type is hardcoded/non-existent/misdetected. As a guess, try setting TERM=linux (and exporting) before calling dialog. Any particular reason to use a statically compiled version (or are you referring to the standard kickstart-bundled one)? Cheers, Phil From Paul.Miles at quadriga.com Tue Aug 16 08:56:06 2005 From: Paul.Miles at quadriga.com (Miles, Paul) Date: Tue, 16 Aug 2005 09:56:06 +0100 Subject: Custom dialog messages Message-ID: <5D339AE687B0C8438185AE62ED06697D7A8106@ukchmail01.mail.quadriga.com> Hi Phil, There isn't a standard kickstart dialog is there ? I certainly wouldn't have gone to the effort of statically compiling it if I'd seen it there. I tried : export TERM=linux but it made no change. Any other thoughts? Thanks, Paul P.S Sorry for the stupidly long signature, I have no control over it I'm afraid. --- This e-mail is the property of Quadriga Worldwide Ltd The message (and any associated files) is intended only for the use of the individual or entity to which it is addressed and may contain information that is confidential, subject to copyright or constitutes a trade secret. If you are not the intended recipient you are hereby notified that any dissemination, copying or distribution of this message, or files associated with this message, is strictly prohibited. If you have received this message in error, please notify us immediately by replying to the message and deleting it from your computer. Messages sent to and from us may be monitored. Internet communications cannot be guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain viruses. Therefore, we do not accept responsibility for any errors or omissions that are present in this message, or any attachment, that have arisen as a result of e-mail transmission. If verification is required, please request a hard-copy version. Any views or opinions presented are solely those of the author and do not necessarily represent those of the company. From Christian.Rohrmeier at SCHERING.DE Tue Aug 16 10:28:32 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Tue, 16 Aug 2005 12:28:32 +0200 Subject: Mass RHEL Management (was re: Custom dialog messages) In-Reply-To: <5D339AE687B0C8438185AE62ED06697D7A8106@ukchmail01.mail.quadriga.com> Message-ID: Hi all, Yeah, I am wondering about this too. RHEL 2.1 certainly doesn't have Dialog in it Kickstart environment. Because of this I have implemented a very lengthy (but rather functional) interactive shell-based installation script. The whole thing runs in %PRE, such that all user-interaction is done with right off the bat and the rest of the install runs without any user input. I would ofcourse love to make this work with Dialog, which is much nicer than just having echo/read statements. I for example provide defaults for, but allow changes to hostname system description (for MOTD/SSH banner) IP Gateway NM network line settings console keyboard layout timezone DNS servers (1 -3) domain name domain search order NTP server SMTP gateway etc. All this gets asked interactively in the PRE% script. The idea is to have one group (that has no admin experience) setup the servers and hand them over ready to go, to have applications, etc., installed by another group (application administrators), and finally a third group is responsible for running the system (real unix admins). Anyhow, Dialog is unfortunately definately not part of the standard Kickstart environment. You may now ask: Christian, why don't you just make your own boot media? And indeed it seems that is the way I am forced to go. Forced, because I am not willingly going in that direction because I have RHEL 2.1, 3, and 4, I have WS and AS, I have various Update releases, and I have 32bit and 64bit servers to support. If I have to customize an install CD for all of those, I'd go insane! (I'm already going insane just dealing with the kickstart scripts and install trees for all of the above mentioned versions of RHEL.) So, my idea was: do EVERYTHING in the KS script, and use only standard boot/install CDs. To save myself work. But the more I do in KS, the more I notice how much is missing, especially in RHEL 2.1 (the bane of my existence: RHEL 2.1 KS doesn't do %INCLUDE, which makes it VERY hard to work with.) -Christian PS. I also have to support IDE, SCSI, and Compaq RAID controller-based systems. So far I have avoided trying to automate the selection of the installation drive by having the user manually select what drive controller to be used for the installation (i.e. "cciss/c0d0", "sda", or "hda"). I can look in "/proc/ide/hdX/media" to see if there is an IDE hard drive, but finding out which SCSI or (worse still) which CPQ RAID controller has which disks is eluding me at the moment. If someone has a schell script that I can stick into %PRE that automates install hard drive selection, that would make my day. _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 "Miles, Paul" To Sent by: kickstart-list-bo cc unces at redhat.com Subject Re: Custom dialog messages 16.08.2005 10:56 Please respond to Discussion list about Kickstart Hi Phil, There isn't a standard kickstart dialog is there ? I certainly wouldn't have gone to the effort of statically compiling it if I'd seen it there. I tried : export TERM=linux but it made no change. Any other thoughts? Thanks, Paul P.S Sorry for the stupidly long signature, I have no control over it I'm afraid. --- This e-mail is the property of Quadriga Worldwide Ltd The message (and any associated files) is intended only for the use of the individual or entity to which it is addressed and may contain information that is confidential, subject to copyright or constitutes a trade secret. If you are not the intended recipient you are hereby notified that any dissemination, copying or distribution of this message, or files associated with this message, is strictly prohibited. If you have received this message in error, please notify us immediately by replying to the message and deleting it from your computer. Messages sent to and from us may be monitored. Internet communications cannot be guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or contain viruses. Therefore, we do not accept responsibility for any errors or omissions that are present in this message, or any attachment, that have arisen as a result of e-mail transmission. If verification is required, please request a hard-copy version. Any views or opinions presented are solely those of the author and do not necessarily represent those of the company. _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From steve.adams at ixora.com.au Wed Aug 17 04:56:03 2005 From: steve.adams at ixora.com.au (Steve Adams) Date: Wed, 17 Aug 2005 14:56:03 +1000 Subject: badblocks Message-ID: <003201c5a2e7$f83ba250$ae8735cb@ixora.com.au> Hi All, I am trying to scan for badblocks during the partitioning phase of a kickstart install. However, despite that "--badblocks" is specified on the "part" lines in ks.cfg, the format commands shown on virtual console 3 during the install are just. * Format command: ['/usr/sbin/mke2fs', '/tmp/hda1', '-i', '4096', '-j'] I do not see a -c being passed to mke2fs, or any other evidence that the badblocks check is actually happening. Is there anything else that I should be doing? Sorry if this is a FAQ, but I searched and did not find it. @ Regards, @ Steve Adams @ http://www.ixora.com.au/ - For DBAs @ http://www.christianity.net.au/ - For all From clumens at redhat.com Wed Aug 17 14:02:06 2005 From: clumens at redhat.com (Chris Lumens) Date: Wed, 17 Aug 2005 10:02:06 -0400 (EDT) Subject: badblocks In-Reply-To: <003201c5a2e7$f83ba250$ae8735cb@ixora.com.au> References: <003201c5a2e7$f83ba250$ae8735cb@ixora.com.au> Message-ID: > I am trying to scan for badblocks during the partitioning phase of a kickstart install. > However, despite that "--badblocks" is specified on the "part" lines in ks.cfg, > the format commands shown on virtual console 3 during the install are just. > > * Format command: ['/usr/sbin/mke2fs', '/tmp/hda1', '-i', '4096', '-j'] > > I do not see a -c being passed to mke2fs, or any other evidence that the badblocks > check is actually happening. Is there anything else that I should be doing? > > Sorry if this is a FAQ, but I searched and did not find it. According to kickstart.py, badblocks checking is no longer supported. You should be seeing a message about it in /tmp/anaconda.log. I'll verify that the documentation reflects this and also perhaps modify kickstart.py to complain much more loudly. Here's the relevant section: elif str == "--badblocks": # no longer support badblocks checking log.warning("--badblocks specified but is no longer supported") - Chris From klaus.steden at thomson.net Thu Aug 18 03:15:35 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 17 Aug 2005 23:15:35 -0400 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: References: Message-ID: <20050818031535.GT897@thomson.net> For anyone who's interested, I've published some documentation on the Anaconda Wiki discussing the techniques I used to extend Kickstart to cover a range of system configurations and distros using the same boot media. http://fedoraproject.org/wiki/AnacondaExtendingKickstart Hope you find it useful, and happy Kickstarting! cheers, Klaus From menscher at uiuc.edu Thu Aug 18 05:17:46 2005 From: menscher at uiuc.edu (Damian Menscher) Date: Thu, 18 Aug 2005 00:17:46 -0500 (CDT) Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: <20050818031535.GT897@thomson.net> References: <20050818031535.GT897@thomson.net> Message-ID: On Wed, 17 Aug 2005, Klaus Steden wrote: > http://fedoraproject.org/wiki/AnacondaExtendingKickstart I'm curious: why not just do it as: linux ks=http://server/path/to/ks-cfg.cgi?hostname=host&class=wkstn This is something I was just about to do, since I need to install a bunch of machines, and want to hardcode a few things into each machine (hostname, ip, ssh keys) based on who it is. Is there some reason this won't work? Damian Menscher -- -=#| Physics Grad Student & SysAdmin @ U Illinois Urbana-Champaign |#=- -=#| 488 LLP, 1110 W. Green St, Urbana, IL 61801 Ofc:(217)333-0038 |#=- -=#| 4602 Beckman, VMIL/MS, Imaging Technology Group:(217)244-3074 |#=- -=#| www.uiuc.edu/~menscher/ Fax:(217)333-9819 |#=- -=#| The above opinions are not necessarily those of my employers. |#=- From klaus.steden at thomson.net Thu Aug 18 05:43:28 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 18 Aug 2005 01:43:28 -0400 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: References: <20050818031535.GT897@thomson.net> Message-ID: <20050818054328.GU897@thomson.net> > >http://fedoraproject.org/wiki/AnacondaExtendingKickstart > > I'm curious: why not just do it as: > > linux ks=http://server/path/to/ks-cfg.cgi?hostname=host&class=wkstn > > This is something I was just about to do, since I need to install a > bunch of machines, and want to hardcode a few things into each machine > (hostname, ip, ssh keys) based on who it is. Is there some reason this > won't work? > No, that would definitely work ... I wanted more than that, though. - I wanted to make sure that installation would abort if a component (disk, NIC, host name, etc.) wasn't found - not fall back to asking me questions. I've got a fairly homogeneous hardware environment - if a hard drive doesn't work, I've got a bigger problem. - I wanted more compact command-line syntax, so that non-admin people could make use of it, too. - I wanted to be able to use a single medium to bootstrap and install multiple systems without having to cut a CD for every new version. Klaus From digit-x at gmx.net Thu Aug 18 09:31:41 2005 From: digit-x at gmx.net (digit-x at gmx.net) Date: Thu, 18 Aug 2005 11:31:41 +0200 (MEST) Subject: Kickstart first timer Message-ID: <8994.1124357501@www30.gmx.net> Hi at list not familiar to fedora core linux/kick start i`ve got several questions and hope they`ll be answered within this list. In a currrent project i`ve to generate a bootable Fedora Core 2 recovery DVD with specific pre-configured applications (apache/tomcat/....) enclosing a software developed by our company. After investigating i found kickstart, but am not quite sure, if this tool is able to solve my problem. Pre-requisites: 1.) A dedicated hardware platform will be used (using an image can be an option) 2.) Dedicated pre-installed and pre-configured packages have to be available 3.) our proprietary Software shall be installed within this recovery/installation process. 4.) Further options should be integrated into the setup process (e.g. mounting a USB filesystem; Data Backup from USB filesystem) Will that be possible with Kickstart? If not, do you have any suggestions how to automate these tasks? Thank you for replying Ingo -- Lust, ein paar Euro nebenbei zu verdienen? Ohne Kosten, ohne Risiko! Satte Provisionen f?r GMX Partner: http://www.gmx.net/de/go/partner From Christian.Rohrmeier at SCHERING.DE Thu Aug 18 10:31:22 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 18 Aug 2005 12:31:22 +0200 Subject: 46 kick start scripts, 46 install trees, 46 boot CDs In-Reply-To: <20050818054328.GU897@thomson.net> Message-ID: Hi Klaus, et. al, The info on the wiki looks excellent. I will give it a shot and report back here how it goes making this work in our environment. Thanks again, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 Klaus Steden To Sent by: Discussion list about Kickstart kickstart-list-bo unces at redhat.com cc Subject 18.08.2005 07:43 Re: 46 kick start scripts, 46 install trees, 46 boot CDs Please respond to Discussion list about Kickstart > >http://fedoraproject.org/wiki/AnacondaExtendingKickstart > > I'm curious: why not just do it as: > > linux ks=http://server/path/to/ks-cfg.cgi?hostname=host&class=wkstn > > This is something I was just about to do, since I need to install a > bunch of machines, and want to hardcode a few things into each machine > (hostname, ip, ssh keys) based on who it is. Is there some reason this > won't work? > No, that would definitely work ... I wanted more than that, though. - I wanted to make sure that installation would abort if a component (disk, NIC, host name, etc.) wasn't found - not fall back to asking me questions. I've got a fairly homogeneous hardware environment - if a hard drive doesn't work, I've got a bigger problem. - I wanted more compact command-line syntax, so that non-admin people could make use of it, too. - I wanted to be able to use a single medium to bootstrap and install multiple systems without having to cut a CD for every new version. Klaus _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From digit-x at gmx.net Thu Aug 18 14:30:13 2005 From: digit-x at gmx.net (digit-x at gmx.net) Date: Thu, 18 Aug 2005 16:30:13 +0200 (MEST) Subject: 1 simple question Message-ID: <18357.1124375413@www21.gmx.net> due to the fact that the linux distribution used in our company is fedora core 2 ( without kickstart configurator installed on it ) i`d like to ask where the installer (rpm/src) can be found -- GMX DSL = Maximale Leistung zum minimalen Preis! 2000 MB nur 2,99, Flatrate ab 4,99 Euro/Monat: http://www.gmx.net/de/go/dsl From phr at doc.ic.ac.uk Thu Aug 18 15:29:48 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 18 Aug 2005 16:29:48 +0100 (BST) Subject: 1 simple question In-Reply-To: <18357.1124375413@www21.gmx.net> References: <18357.1124375413@www21.gmx.net> Message-ID: On Thu, 18 Aug 2005 digit-x at gmx.net wrote: >due to the fact that the linux distribution used in our company is fedora >core 2 ( without kickstart configurator installed on it ) i`d like to ask >where the installer (rpm/src) can be found ftp://download.fedora.redhat.com/pub/fedora/linux/core/2/SRPMS/anaconda-10.0-5.src.rpm Depending on your requirements, you might also need: anaconda-help-9.92-1.src.rpm anaconda-images-10-3.src.rpm Cheers, Phil From omer at faruk.net Fri Aug 19 05:37:23 2005 From: omer at faruk.net (Omer Faruk Sen) Date: Fri, 19 Aug 2005 08:37:23 +0300 (EEST) Subject: ks.cfg located on a usb drive Message-ID: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2> Hi, I am trying to install my server using ks.cfg but the server I want to install has no floppy drive. I have just usb drive which recognized as /dev/sda1. I wanted to use it but since I don't know the right kernel parameters I couldn't use it. I have also searched the list archives but I couldn't find an answer. I have tried linux ks=/dev/sda1/ks.cfg linux ks=usb but none of them worked. How can I use ks.cfg file located on a usb drive. USB drive is recognized and required modules installed by anaconda. The system I am trying to install is RH ES4. Is it possible for you in the list to give me the right kernel parameters ? -- Omer Faruk Sen http://www.faruk.net From omer at faruk.net Fri Aug 19 06:03:22 2005 From: omer at faruk.net (Omer Faruk Sen) Date: Fri, 19 Aug 2005 09:03:22 +0300 (EEST) Subject: ks.cfg located on a usb drive In-Reply-To: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2> References: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2> Message-ID: <50638.193.140.74.2.1124431402.squirrel@193.140.74.2> I have another question. Is it possible to install servers with CD-ROM using kickstart? As far as I understand kickstart installation uses ftp or http locations to install server. Am I right? > > Hi, > > I am trying to install my server using ks.cfg but the server I want to > install has no floppy drive. I have just usb drive which recognized as > /dev/sda1. I wanted to use it but since I don't know the right kernel > parameters I couldn't use it. I have also searched the list archives but I > couldn't find an answer. > > I have tried > > linux ks=/dev/sda1/ks.cfg > linux ks=usb > > but none of them worked. How can I use ks.cfg file located on a usb drive. > USB drive is recognized and required modules installed by anaconda. The > system I am trying to install is RH ES4. > > Is it possible for you in the list to give me the right kernel parameters > ? > > > > -- > Omer Faruk Sen > http://www.faruk.net > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Omer Faruk Sen http://www.faruk.net From pmatilai at laiskiainen.org Fri Aug 19 19:28:04 2005 From: pmatilai at laiskiainen.org (Panu Matilainen) Date: Fri, 19 Aug 2005 12:28:04 -0700 (PDT) Subject: Custom dialog messages In-Reply-To: <5D339AE687B0C8438185AE62ED06697D7A8106@ukchmail01.mail.quadriga.com> References: <5D339AE687B0C8438185AE62ED06697D7A8106@ukchmail01.mail.quadriga.com> Message-ID: On Tue, 16 Aug 2005, Miles, Paul wrote: > Hi Phil, > > There isn't a standard kickstart dialog is there ? I certainly wouldn't > have gone to the effort of statically compiling it if I'd seen it there. > > I tried : export TERM=linux but it made no change. > > Any other thoughts? I started using python for both %pre and %post, from there you can use the snack module (available even in %pre stage) to build whatever dialogs you want and as an added bonus it looks exactly like anaconda itself :) Snack was showing the box borders with wrong characters (at least in %post) as well but setting LC_ALL=C cured that, might help with dialog as well. - Panu - From uwelists at u-b.de Sat Aug 20 09:36:53 2005 From: uwelists at u-b.de (Uwe E. Bilger) Date: Sat, 20 Aug 2005 11:36:53 +0200 Subject: Installing image via PXE Message-ID: Hello, I'm not sure whether I'm mailing to the correct list. Please, after taring and feathering, point me to a correct list, in case I'm wrong. I would like to install, or maybe better transfer?, an image, which I created with dd, from a server to the client via pxe. I know that there is a bunch of OSS out there doing a great job administering a bunch of images for server farms, etc. That's too complex. I only have three machines with three different snapshot images, which I would like to transfer back for recovery purposes in an automated manner. I /can/ set up a kickstart installation over pxe. I'm wondering if the pxe kernel has options or something like that to simply transfer an image to the target machine, instead of kicking off a kickstart installation. I'm familiar with basic pxe installation procedures, coming from redhat's (fedora's) kickstart routine, so I don't need to know those basics which can be found in documentation as well. Just the step of: don't want to use kickstart, just want to use an image. Thanks for the help in advance. Uwe From error27 at gmail.com Mon Aug 22 03:57:38 2005 From: error27 at gmail.com (Dan Carpenter) Date: Mon, 22 Aug 2005 05:57:38 +0200 Subject: ks.cfg located on a usb drive In-Reply-To: <50638.193.140.74.2.1124431402.squirrel@193.140.74.2> References: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2> <50638.193.140.74.2.1124431402.squirrel@193.140.74.2> Message-ID: 2005/8/19, Omer Faruk Sen : > I have another question. Is it possible to install servers with CD-ROM > using kickstart? Yes. I don't know how to do this with a USB floppy but with a regular floppy then you just boot from the CDROM and say ks=floppy regards, dan carpenter From miller_duane at bah.com Mon Aug 22 07:08:10 2005 From: miller_duane at bah.com (Miller Duane) Date: Mon, 22 Aug 2005 03:08:10 -0400 Subject: ks.cfg located on a usb drive References: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2><50638.193.140.74.2.1124431402.squirrel@193.140.74.2> Message-ID: <5C063D1B790A3449B0B58BE75AEDE8300EAD58@MCLNEXVS02.resource.ds.bah.com> Or if the ks.cfg is on the cdrom then ks=cdrom:/ks.cfg D -----Original Message----- From: kickstart-list-bounces at redhat.com on behalf of Dan Carpenter Sent: Sun 8/21/2005 11:57 PM To: omer at faruk.net; Discussion list about Kickstart Subject: Re: ks.cfg located on a usb drive 2005/8/19, Omer Faruk Sen : > I have another question. Is it possible to install servers with CD-ROM > using kickstart? Yes. I don't know how to do this with a USB floppy but with a regular floppy then you just boot from the CDROM and say ks=floppy regards, dan carpenter _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list -------------- next part -------------- A non-text attachment was scrubbed... Name: winmail.dat Type: application/ms-tnef Size: 2934 bytes Desc: not available URL: From omer at faruk.net Mon Aug 22 08:13:29 2005 From: omer at faruk.net (Omer Faruk Sen) Date: Mon, 22 Aug 2005 11:13:29 +0300 (EEST) Subject: ks.cfg located on a usb drive In-Reply-To: <5C063D1B790A3449B0B58BE75AEDE8300EAD58@MCLNEXVS02.resource.ds.bah.com > References: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2><50638.193.140.74.2.1124431402.squirrel@193.140.74.2> <5C063D1B790A3449B0B58BE75AEDE8300EAD58@MCLNEXVS02.resource.ds.bah.com> Message-ID: <52701.193.140.74.2.1124698409.squirrel@193.140.74.2> I have found the answer. For the sake of record: linux ks=hd:sda1/ks.cfg > Or if the ks.cfg is on the cdrom then ks=cdrom:/ks.cfg > D > -----Original Message----- > From: kickstart-list-bounces at redhat.com on behalf of Dan Carpenter > Sent: Sun 8/21/2005 11:57 PM > To: omer at faruk.net; Discussion list about Kickstart > Subject: Re: ks.cfg located on a usb drive > > 2005/8/19, Omer Faruk Sen : >> I have another question. Is it possible to install servers with CD-ROM >> using kickstart? > > Yes. I don't know how to do this with a USB floppy but with a regular > floppy then you just boot from the CDROM and say ks=floppy > > regards, > dan carpenter > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list -- Omer Faruk Sen http://www.faruk.net From drkludge at cox.net Tue Aug 23 06:28:09 2005 From: drkludge at cox.net (Greg Morgan) Date: Mon, 22 Aug 2005 23:28:09 -0700 Subject: ks.cfg located on a usb drive In-Reply-To: <50638.193.140.74.2.1124431402.squirrel@193.140.74.2> References: <61787.193.140.74.2.1124429843.squirrel@193.140.74.2> <50638.193.140.74.2.1124431402.squirrel@193.140.74.2> Message-ID: <430AC1F9.8090803@cox.net> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Omer Faruk Sen wrote: > I have another question. Is it possible to install servers with CD-ROM > using kickstart? As far as I understand kickstart installation uses ftp or > http locations to install server. Am I right? There a many ways to use a kickstart file. The question may be more of where can anaconda find the kickstart file. However, you are not limited to using just http, or ftp installations when you use a kickstart file. You can use NFS and CD-ROM too. A floppy disk, another harddrive, a USB drive, etc can hold the ks.cfg file for any of these methods. Greg -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFDCsH5xyxe5L6mr7IRAqZ4AKCFL6AS3QVr9fCfzhWFgwzxAE6oqgCdHTzv 9lYD1XkQaKYxE/8LP20hab4= =Pk6A -----END PGP SIGNATURE----- From mail2vinodj at yahoo.co.in Tue Aug 23 15:07:10 2005 From: mail2vinodj at yahoo.co.in (vinod kumar) Date: Tue, 23 Aug 2005 16:07:10 +0100 (BST) Subject: hello Message-ID: <20050823150710.74164.qmail@web8502.mail.in.yahoo.com> my name is vinod kumar j i need the details how to recompile linux kernel 2.6 to view windows ntfs files regards vinod --------------------------------- Check out Yahoo! India Rakhi Special for Rakhi shopping, contests and lots more. http://in.promos.yahoo.com/rakhi/index.html -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew.w.robinson at mms.gov Tue Aug 23 17:07:20 2005 From: andrew.w.robinson at mms.gov (Robinson, Andrew W.) Date: Tue, 23 Aug 2005 12:07:20 -0500 Subject: hello Message-ID: <82ABF19AC0F67D40A0E2B2F5949BE8E20413D8@IMSNEXPRI02.service.agency.mms.pri> Check the Linux-NTFS project: http://linux-ntfs.sourceforge.net/ Andrew Robinson ________________________________ From: kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com] On Behalf Of vinod kumar Sent: Tuesday, August 23, 2005 10:07 AM To: kickstart-list at redhat.com Subject: hello my name is vinod kumar j i need the details how to recompile linux kernel 2.6 to view windows ntfs files regards vinod ________________________________ Check out Yahoo! India Rakhi Special for Rakhi shopping, contests and lots more. http://in.promos.yahoo.com/rakhi/index.html -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew.w.robinson at mms.gov Wed Aug 24 15:46:13 2005 From: andrew.w.robinson at mms.gov (Robinson, Andrew W.) Date: Wed, 24 Aug 2005 10:46:13 -0500 Subject: Cleaning up the VMWare disk Message-ID: <82ABF19AC0F67D40A0E2B2F5949BE8E20413DB@IMSNEXPRI02.service.agency.mms.pri> One of the key learnings Craig and I learned at the VMWare class was that there is a limit to the number of files that can exist on a vmfs file system. We've been using the big vamps file system on imsnolesxOMM to park files that someone might want in the future. Bzzzzt! We plan to migrate off every file that is not owned by a currently registered virtual machine. We would love to completely delete any files that will not be needed ever again. The attached spreadsheet contains a list of the "un-owned" files that we plan to migrate or delete. Please check the list and let us know if any of those files should not be migrated. Please let us know which ones can be deleted. Thanks! Andrew -------------- next part -------------- A non-text attachment was scrubbed... Name: vmfsreport.xls Type: application/vnd.ms-excel Size: 26112 bytes Desc: vmfsreport.xls URL: From andrew.w.robinson at mms.gov Wed Aug 24 15:56:05 2005 From: andrew.w.robinson at mms.gov (Robinson, Andrew W.) Date: Wed, 24 Aug 2005 10:56:05 -0500 Subject: Cleaning up the VMWare disk Message-ID: <82ABF19AC0F67D40A0E2B2F5949BE8E20413DC@IMSNEXPRI02.service.agency.mms.pri> My apologies. This was NOT intended for the kickstart list obviously. Sorry to clutter your mailboxes. Andrew Robinson > -----Original Message----- > From: Robinson, Andrew W. > Sent: Wednesday, August 24, 2005 10:46 AM > To: Henshaw Young, Beverly; Fouquet, Errol; Sharp, Dennis; > Thomason, William B > Cc: Discussion list about Kickstart > Subject: Cleaning up the VMWare disk > > One of the key learnings Craig and I learned at the VMWare class was > that there is a limit to the number of files that can exist on a vmfs > file system. We've been using the big vamps file system on > imsnolesxOMM > to park files that someone might want in the future. Bzzzzt! > We plan to > migrate off every file that is not owned by a currently registered > virtual machine. We would love to completely delete any files > that will > not be needed ever again. The attached spreadsheet contains a list of > the "un-owned" files that we plan to migrate or delete. > Please check the > list and let us know if any of those files should not be migrated. > Please let us know which ones can be deleted. > > Thanks! > > Andrew > From tj_yang at hotmail.com Wed Aug 24 21:10:38 2005 From: tj_yang at hotmail.com (T.J. Yang) Date: Wed, 24 Aug 2005 16:10:38 -0500 Subject: How to eject the bootable kickstart CD after build ? Message-ID: Hi, there I tried "/usr/bin/eject /dev/cdrom" or "reboot" as last line in ks.cfg in the hope to eject the CD after finish the ks.cfg pre/post commands. But these two command doesn't work. Anyone know how to auto eject the bootable CD in ks.cfg ? Thanks for your pointer T.J. Yang From error27 at gmail.com Wed Aug 24 21:13:56 2005 From: error27 at gmail.com (Dan Carpenter) Date: Wed, 24 Aug 2005 14:13:56 -0700 Subject: How to eject the bootable kickstart CD after build ? In-Reply-To: References: Message-ID: The installer is still using the cdrom so you can't eject it. regards, dan carpenter From tj_yang at hotmail.com Wed Aug 24 21:21:09 2005 From: tj_yang at hotmail.com (T.J. Yang) Date: Wed, 24 Aug 2005 16:21:09 -0500 Subject: How to eject the bootable kickstart CD after build ? In-Reply-To: Message-ID: >From: Dan Carpenter >Reply-To: Discussion list about Kickstart >To: Discussion list about Kickstart >Subject: Re: How to eject the bootable kickstart CD after build ? >Date: Wed, 24 Aug 2005 14:13:56 -0700 > >The installer is still using the cdrom so you can't eject it. hmm, If this is the case, why I can manuallly eject the CD by pushing the ejcect button after the harddrive starting to get formatted ? Also from the ctrl-alt-F3 screen, I saw /usr/bin/eject /dev/cdrom failed "no such device" message. looks I just need to find the right device name ? tj > >regards, >dan carpenter > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list From info at hostinthebox.net Wed Aug 24 22:54:26 2005 From: info at hostinthebox.net (Dan Trainor) Date: Wed, 24 Aug 2005 16:54:26 -0600 Subject: reading /proc/cmdline Message-ID: <430CFAA2.1020707@hostinthebox.net> Hello, all - I use the following example of code in my kickstart file, to get values from /proc/cmdline to use later on during the install: if grep -i -q "mystring[a-zA-Z0-9]" /proc/cmdline then MYSTRING=`cat /proc/cmdline | sed 's/.*mystring=\([^ ]*\).*/\1/'` fi I think Klaus in fact gave this to me. However, I've ran into a few problems. It seems that this does not like strings with periods in them, such as an IP address. Since I know a bit less than nothing about regular expressions, I was hoping that someone could give me a hand with this. I need to modify it to allow me to assign a dotted quad string to the MYSTRING variable. Again, any help would be greatly appreciated. Thanks! -dant From klaus.steden at thomson.net Wed Aug 24 23:06:06 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Wed, 24 Aug 2005 19:06:06 -0400 Subject: reading /proc/cmdline In-Reply-To: <430CFAA2.1020707@hostinthebox.net> References: <430CFAA2.1020707@hostinthebox.net> Message-ID: <20050824230606.GR897@thomson.net> > Hello, all - > > I use the following example of code in my kickstart file, to get values > from /proc/cmdline to use later on during the install: > > if grep -i -q "mystring[a-zA-Z0-9]" /proc/cmdline > then > MYSTRING=`cat /proc/cmdline | sed 's/.*mystring=\([^ ]*\).*/\1/'` > fi > > I think Klaus in fact gave this to me. However, I've ran into a few > problems. It seems that this does not like strings with periods in > them, such as an IP address. > > Since I know a bit less than nothing about regular expressions, I was > hoping that someone could give me a hand with this. I need to modify it > to allow me to assign a dotted quad string to the MYSTRING variable. > > Again, any help would be greatly appreciated. > Hello again, Dan, How about ... if grep -iq "mystring=[a-zA-Z0-9]" /proc/cmdline then MYSTRING=`cat /proc/cmdline | sed 's/.*mystring=\([^ ]*\).*/\1/'` fi Quick testing with sh on Linux tells me that it should preserve the dots. It should just be deleting everything up to, but not including 'mystring=' and whatever follows that, and then deleting everything after that string, and returning the match. REs can be tricky, though, so holler if you get stuck. Klaus From phr at doc.ic.ac.uk Wed Aug 24 23:32:54 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Thu, 25 Aug 2005 00:32:54 +0100 (BST) Subject: How to eject the bootable kickstart CD after build ? In-Reply-To: References: Message-ID: On Wed, 24 Aug 2005, T.J. Yang wrote: >hmm, If this is the case, why I can manuallly eject the CD by pushing >the ejcect button after the harddrive starting to get formatted ? > >Also from the ctrl-alt-F3 screen, I saw /usr/bin/eject /dev/cdrom >failed "no such device" message. looks I just need to find the right >device name ? Is this is a %post --nochroot section? If not, the /dev/cdrom won't have been created, hence the "no such device" message. Cheers, Phil From Christian.Rohrmeier at SCHERING.DE Thu Aug 25 11:03:05 2005 From: Christian.Rohrmeier at SCHERING.DE (Christian.Rohrmeier at SCHERING.DE) Date: Thu, 25 Aug 2005 13:03:05 +0200 Subject: More networking params on the boot command line In-Reply-To: <20050824230606.GR897@thomson.net> Message-ID: Hi All, On the way to making my super CD that supports dozens of RHEL versions and updates, I wanted to test support for non-DCHP installs (just incase). And I can feed the following boot options to the loader: linux ks=http://some.server/some.script ip=1.2.3.4 netmask=255.255.255.128 gateway=1.2.3.1 dns=1.2.3.2 And that works, it doesn't try to use DHCP for any of that information. But it DOES sit there and try to figure out its hostname and domain name! And because I haven't supplied those, it times out trying to determine those values. Does anyone happen to know how you can tell anaconda what the systems hostname and domain are supposed to be via the boot commandline? Maybe its just host= domain=? hostname=? node=? Well, before I guess more, and before I try to google even more, I figured I'd ask you guys because I know you know. =) Thanks, Christian _________________ Christian Rohrmeier Schering AG Corporate IT - Infrastructure and Services Computer Systems and Operations System Administration - Research and Development Tel +49 30 468 15794 Fax +49 30 468 95794 From somewhere_or_other at byu.edu Thu Aug 25 14:23:10 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 08:23:10 -0600 Subject: disabling dhcp for eth1 in kernel Message-ID: <430DD44E.2040308@byu.edu> Hey all. Admittedly, I'm new to this whole kickstart thing, but I can't seem to find an answer to my question anywhere else. For reasons that are out of my control, I can't provide a DHCP address to the eth1 interface on the node I'm trying to install. I don't need the eth1 interface to perform the installation, so I'd like disable it using a kernel parameter so that the step of configuring the eth1 interface during kickstart (before it retrieves the ks.cfg over the network) is non-interactive. This is the last thing I need to do to make my configuration fully non-interactive and fully remote. Essentially I need something like this, although I'm making up the parameters: linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=eth0 disable_eth1 Alternatively if someone has a way to specify a static address as a kernel parameter, that'd be okay too. Anyone have any ideas? Thanks, Lloyd Brown From harry at mad-cat.co.uk Thu Aug 25 15:01:14 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 25 Aug 2005 16:01:14 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: Hi, If you only have one interface physically connected to the network (eth0) then try this: linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=link This will get ks to use the interface which has a link. Not sure if that is what you are looking for but I have to set up boxes with 3 nics - only one of which is connected to a network - works well for me. Regards Harry On Thu, 25 Aug, Lloyd T Brown wrote: > Hey all. > > Admittedly, I'm new to this whole kickstart thing, but I can't seem to > find an answer to my question anywhere else. For reasons that are out > of my control, I can't provide a DHCP address to the eth1 interface on > the node I'm trying to install. I don't need the eth1 interface to > perform the installation, so I'd like disable it using a kernel > parameter so that the step of configuring the eth1 interface during > kickstart (before it retrieves the ks.cfg over the network) is > non-interactive. This is the last thing I need to do to make my > configuration fully non-interactive and fully remote. Essentially I > need something like this, although I'm making up the parameters: > > linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=eth0 disable_eth1 > > Alternatively if someone has a way to specify a static address as a > kernel parameter, that'd be okay too. > > Anyone have any ideas? > > Thanks, > Lloyd Brown > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From somewhere_or_other at byu.edu Thu Aug 25 15:16:43 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 09:16:43 -0600 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <430DE0DB.2030705@byu.edu> Well, sort of. The idea is to have the kickstart be fully remote and non-interactive, so going up there to physically disconnect one of the interfaces doesn't really solve the problem, since I could just as easily go up there to put a static IP into the dialog box. I just want to avoid going up there at all. Unfortunately, the network architecture here is a little strange, and it precludes me from supplying a DHCP address for eth1 like I do for eth0. It has to do with the DHCP server's eth1 (same physical segment as the node's eth1) having two VLANs attached to it with separate static IP addresses and custom MTUs. So, without being able to attach the eth1 address to at least one of those VLAN numbers, and set the MTU, all before the ks.cfg file is retrieved, there doesn't seem to be a way of using DHCP for my solution. Unfortunately, as I said before, this network architecture is out of my control, so I just have to figure it out. Some of our nodes have Platform Rocks on them, though, and they somehow do the very thing I'm trying. I can't figure out how they do it, though. If anyone can point me the right direction for that investigation, that'd be great too. Thanks, Lloyd Harry Mills wrote: >Hi, > >If you only have one interface physically connected to the network (eth0) then try this: > >linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=link > >This will get ks to use the interface which has a link. > >Not sure if that is what you are looking for but I have to set up boxes with 3 nics - only one of which is connected to a network - works well for me. > >Regards > >Harry > > > >On Thu, 25 Aug, Lloyd T Brown wrote: > > >>Hey all. >> >>Admittedly, I'm new to this whole kickstart thing, but I can't seem to >>find an answer to my question anywhere else. For reasons that are out >>of my control, I can't provide a DHCP address to the eth1 interface on >>the node I'm trying to install. I don't need the eth1 interface to >>perform the installation, so I'd like disable it using a kernel >>parameter so that the step of configuring the eth1 interface during >>kickstart (before it retrieves the ks.cfg over the network) is >>non-interactive. This is the last thing I need to do to make my >>configuration fully non-interactive and fully remote. Essentially I >>need something like this, although I'm making up the parameters: >> >>linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=eth0 disable_eth1 >> >>Alternatively if someone has a way to specify a static address as a >>kernel parameter, that'd be okay too. >> >>Anyone have any ideas? >> >>Thanks, >>Lloyd Brown >> >>_______________________________________________ >>Kickstart-list mailing list >>Kickstart-list at redhat.com >>https://www.redhat.com/mailman/listinfo/kickstart-list >> >> >> > > > > > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > > > From ext.helice.rivera at sncf.fr Thu Aug 25 16:12:06 2005 From: ext.helice.rivera at sncf.fr (EXT / HELICE RIVERA Cedric) Date: Thu, 25 Aug 2005 18:12:06 +0200 Subject: disabling dhcp for eth1 in kernel Message-ID: Hi, Maybe you can try this : linux ks=nfs:nfs_server:/path/to/ ksdevice=eth0 ip=x.x.x.x If your kickstart config file is complete and it's name is : "/path/to/ks/x.x.x.x-kickstart" then the installation will be completely non-interactive, unless the disk partition table is clean ... I hope this helps. ++ -----Message d'origine----- De : kickstart-list-bounces at redhat.com [mailto:kickstart-list-bounces at redhat.com]De la part de Lloyd T Brown Envoy? : jeudi 25 ao?t 2005 16:23 ? : kickstart-list at redhat.com Objet : disabling dhcp for eth1 in kernel Hey all. Admittedly, I'm new to this whole kickstart thing, but I can't seem to find an answer to my question anywhere else. For reasons that are out of my control, I can't provide a DHCP address to the eth1 interface on the node I'm trying to install. I don't need the eth1 interface to perform the installation, so I'd like disable it using a kernel parameter so that the step of configuring the eth1 interface during kickstart (before it retrieves the ks.cfg over the network) is non-interactive. This is the last thing I need to do to make my configuration fully non-interactive and fully remote. Essentially I need something like this, although I'm making up the parameters: linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=eth0 disable_eth1 Alternatively if someone has a way to specify a static address as a kernel parameter, that'd be okay too. Anyone have any ideas? Thanks, Lloyd Brown _______________________________________________ Kickstart-list mailing list Kickstart-list at redhat.com https://www.redhat.com/mailman/listinfo/kickstart-list From somewhere_or_other at byu.edu Thu Aug 25 16:54:31 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 10:54:31 -0600 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <430DF7C7.5070807@byu.edu> I'll try that later, but, just for reference, which interface does the ip=x.x.x.x paramter apply to? Can I specify which one? I can DHCP eth0 just fine. It's eth1 that I either need to set statically or disable entirely. Lloyd EXT / HELICE RIVERA Cedric wrote: > Hi, > > Maybe you can try this : > >linux ks=nfs:nfs_server:/path/to/ ksdevice=eth0 ip=x.x.x.x > > If your kickstart config file is complete and it's name is : > >"/path/to/ks/x.x.x.x-kickstart" > >then the installation will be completely non-interactive, unless the disk >partition table is clean ... > >I hope this helps. > >++ > > > >-----Message d'origine----- >De : kickstart-list-bounces at redhat.com >[mailto:kickstart-list-bounces at redhat.com]De la part de Lloyd T Brown >Envoy? : jeudi 25 ao?t 2005 16:23 >? : kickstart-list at redhat.com >Objet : disabling dhcp for eth1 in kernel > > >Hey all. > >Admittedly, I'm new to this whole kickstart thing, but I can't seem to >find an answer to my question anywhere else. For reasons that are out >of my control, I can't provide a DHCP address to the eth1 interface on >the node I'm trying to install. I don't need the eth1 interface to >perform the installation, so I'd like disable it using a kernel >parameter so that the step of configuring the eth1 interface during >kickstart (before it retrieves the ks.cfg over the network) is >non-interactive. This is the last thing I need to do to make my >configuration fully non-interactive and fully remote. Essentially I >need something like this, although I'm making up the parameters: > >linux ks=nfs:nfs_server:/path/to/ks.cfg ksdevice=eth0 disable_eth1 > >Alternatively if someone has a way to specify a static address as a >kernel parameter, that'd be okay too. > >Anyone have any ideas? > >Thanks, >Lloyd Brown > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > > > From forltran at yahoo.com Thu Aug 25 17:03:42 2005 From: forltran at yahoo.com (Lambert Tran) Date: Thu, 25 Aug 2005 10:03:42 -0700 (PDT) Subject: Mass Installation Message-ID: <20050825170342.81816.qmail@web80313.mail.yahoo.com> Greetings, Thanks in advance for all info. Our task is to install many workstations from a handcrafted system. This is basically a automated cloning job. We prefer not to use DHCP. This task was done fairly easy when we used Solaris, and HPUX (I had JumpStart, and Ignite setup). While I think KickStart could do the same thing, I am not sure how easy it could be setup. Any thoughts on unforeseen problems? BTW, we will use RedHat4 (both AS and ES). Thanks, -ltran From harry at mad-cat.co.uk Thu Aug 25 17:06:39 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 25 Aug 2005 18:06:39 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: I may well be totally wrong but if you specify the ksdevice then that is the device that will be used to fetch the ks.cfg and no user intervention is required. This is how it seems to work on my set up here. The only time I am prompted for manual network information is when kickstart can't find the ks.cfg file using the location specification in the 'ks=blah' boot parameter. KS then falls back to manually asking for network details. Is it possible that the machine cannot locate/mount the NFS server or resolve the hostname? Harry On Thu, 25 Aug, Lloyd T Brown wrote: > I'll try that later, but, just for reference, which interface does the > ip=x.x.x.x parameter apply to? Can I specify which one? I can DHCP eth0 > just fine. It's eth1 that I either need to set statically or disable > entirely. > > Lloyd > > EXT / HELICE RIVERA Cedric wrote: > > > Hi, > > > > Maybe you can try this : > > > >linux ks=nfs:nfs_server:/path/to/ ksdevice=eth0 ip=x.x.x.x > > > > If your kickstart config file is complete and it's name is : > > > >"/path/to/ks/x.x.x.x-kickstart" > > > >then the installation will be completely non-interactive, unless the disk > >partition table is clean ... > > > >I hope this helps. > > > >++ > > > > > > > >-----Message d'origine----- > >De : kickstart-list-bounces at redhat.com > >[mailto:kickstart-list-bounces at redhat.com]De la part de Lloyd T Brown > >Envoy??? : jeudi 25 ao???t 2005 16:23 > >??? : kickstart-list at redhat.com > >Objet : disabling dhcp for eth1 in kernel > > > > > >_______________________________________________ > >Kickstart-list mailing list > >Kickstart-list at redhat.com > >https://www.redhat.com/mailman/listinfo/kickstart-list From klaus.steden at thomson.net Thu Aug 25 17:25:45 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 25 Aug 2005 13:25:45 -0400 Subject: Mass Installation In-Reply-To: <20050825170342.81816.qmail@web80313.mail.yahoo.com> References: <20050825170342.81816.qmail@web80313.mail.yahoo.com> Message-ID: <20050825172544.GT897@thomson.net> > Greetings, > > Thanks in advance for all info. > Our task is to install many workstations from a handcrafted system. This is > basically a automated cloning job. We prefer not to use DHCP. This task was > done fairly easy when we used Solaris, and HPUX (I had JumpStart, and Ignite > setup). > While I think KickStart could do the same thing, I am not sure how easy it > could be setup. Any thoughts on unforeseen problems? BTW, we will use RedHat4 > (both AS and ES). > Should be fairly straightforward. A number of people on this list have large install bases across more than one distro (myself included). The Anaconda Wiki covers a lot of this stuff in detail (http://fedoraproject.org/wiki/Anaconda). KickStart will definitely do the trick, although you may have to do some tuning to your kickstart config file, etc. good luck, Klaus From somewhere_or_other at byu.edu Thu Aug 25 18:06:12 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 12:06:12 -0600 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <430E0894.9070306@byu.edu> That was what I thought too. Now, like I said, this is my first time working with Kickstart, so I may be missing something. Basically, I use the following parameters: ks=nfs:192.168.201.28:/opt/rhel_image_kickstart/m4b-1-17.ks/ks.cfg ksdevice=eth0 The installation begins, loads a few modules, states that it is trying to get a dhcp address for eth0, then tries with eth1. Since the eth1 dhcp fails, as expected, a dialog pops up to ask for user input, either to try dhcp again, or to put a static IP address in. I put a static address in, and the installation proceeds exactly as planned. I guess what I'm saying is that I'm sure you're right that when I specify the ksdevice that the device I specify will be the only one used to fetch the ks.cfg, etc. All the evidence seems to support this. The problem is that, despite this, Kickstart still seems to want to configure eth1, even though it's totally unnecessary. As far as your question, I'm pretty sure the machine can locate the nfs server. I only export the directory on the NFS server to 192.168.201.x subnet, and I'm putting in a 192.168.2.x address into eth1 by hand. I'm pretty sure that, since the installation proceeds alright from there on, that it's working correctly. That, and once the system is up, I can mount it by hand without any trouble. Since I'm using an IP address for the nfs server, I don't think resolving the hostname is the issue either. Thanks, Lloyd Harry Mills wrote: >I may well be totally wrong but if you specify the ksdevice then that is the device that will be used to fetch the ks.cfg and no user intervention is required. This is how it seems to work on my set up here. > >The only time I am prompted for manual network information is when kickstart can't find the ks.cfg file using the location specification in the 'ks=blah' boot parameter. KS then falls back to manually asking for network details. > >Is it possible that the machine cannot locate/mount the NFS server or resolve the hostname? > > >Harry > >On Thu, 25 Aug, Lloyd T Brown wrote: > > >>I'll try that later, but, just for reference, which interface does the >>ip=x.x.x.x parameter apply to? Can I specify which one? I can DHCP eth0 >>just fine. It's eth1 that I either need to set statically or disable >>entirely. >> >>Lloyd >> >>EXT / HELICE RIVERA Cedric wrote: >> >> >> >>> Hi, >>> >>> Maybe you can try this : >>> >>>linux ks=nfs:nfs_server:/path/to/ ksdevice=eth0 ip=x.x.x.x >>> >>> If your kickstart config file is complete and it's name is : >>> >>>"/path/to/ks/x.x.x.x-kickstart" >>> >>>then the installation will be completely non-interactive, unless the disk >>>partition table is clean ... >>> >>>I hope this helps. >>> >>>++ >>> >>> >>> >>>-----Message d'origine----- >>>De : kickstart-list-bounces at redhat.com >>>[mailto:kickstart-list-bounces at redhat.com]De la part de Lloyd T Brown >>>Envoy??? : jeudi 25 ao???t 2005 16:23 >>>??? : kickstart-list at redhat.com >>>Objet : disabling dhcp for eth1 in kernel >>> >>> >>> > > > >>>_______________________________________________ >>>Kickstart-list mailing list >>>Kickstart-list at redhat.com >>>https://www.redhat.com/mailman/listinfo/kickstart-list >>> >>> > > > > > > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > > > From harry at mad-cat.co.uk Thu Aug 25 19:07:23 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 25 Aug 2005 20:07:23 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: Yep - does seem a little confusing. What does the KS log say (+)? Can you see it get the correct IP from the DHCP server on eth0? On my install (Which is booted from a USB key) I see it get the network info for eth2 (my onboard card) and then fetch the ks.cfg. It never attempts to initialise any of the other cards. The only other oddness I have come across is where later kernel's probe order is different - my onboard always used to be eth0 until one OS version later... it was eth2 :) Cheers Harry On Thu, 25 Aug, Lloyd T Brown wrote: > That was what I thought too. Now, like I said, this is my first time > working with Kickstart, so I may be missing something. Basically, I use > the following parameters: > > ks=nfs:192.168.201.28:/opt/rhel_image_kickstart/m4b-1-17.ks/ks.cfg > ksdevice=eth0 > > The installation begins, loads a few modules, states that it is trying > to get a dhcp address for eth0, then tries with eth1. Since the eth1 > dhcp fails, as expected, a dialog pops up to ask for user input, either > to try dhcp again, or to put a static IP address in. I put a static > address in, and the installation proceeds exactly as planned. > > I guess what I'm saying is that I'm sure you're right that when I > specify the ksdevice that the device I specify will be the only one used > to fetch the ks.cfg, etc. All the evidence seems to support this. The > problem is that, despite this, Kickstart still seems to want to > configure eth1, even though it's totally unnecessary. > > As far as your question, I'm pretty sure the machine can locate the nfs > server. I only export the directory on the NFS server to 192.168.201.x > subnet, and I'm putting in a 192.168.2.x address into eth1 by hand. I'm > pretty sure that, since the installation proceeds alright from there on, > that it's working correctly. That, and once the system is up, I can > mount it by hand without any trouble. Since I'm using an IP address for > the nfs server, I don't think resolving the hostname is the issue either. > > Thanks, > Lloyd From somewhere_or_other at byu.edu Thu Aug 25 19:26:51 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 13:26:51 -0600 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <430E1B7B.2000604@byu.edu> Yes, the ks log shows the ip address being acquired correctly for eth0. There's this weird message about it waiting 15 seconds, though. I'm not sure what that means. The message about the eth0 getting a dhcp address is followed very shortly by a couple messages stating what the IP address for the NFS server is and what the path is to the ks.cfg file, so I'm pretty sure it's working correctly. After that, messages about it trying to acquire a dhcp address for eth1 come too, including a failure message. I'll see if I can get something more exact and post it, if that'd be helpful. Thanks, Lloyd Harry Mills wrote: >Yep - does seem a little confusing. > >What does the KS log say (+)? Can you see it get the correct IP from the DHCP server on eth0? On my install (Which is booted from a USB key) I see it get the network info for eth2 (my onboard card) and then fetch the ks.cfg. It never attempts to initialise any of the other cards. > >The only other oddness I have come across is where later kernel's probe order is different - my onboard always used to be eth0 until one OS version later... it was eth2 :) > >Cheers > >Harry > > > > From brilong at cisco.com Thu Aug 25 19:36:38 2005 From: brilong at cisco.com (Brian Long) Date: Thu, 25 Aug 2005 15:36:38 -0400 Subject: disabling dhcp for eth1 in kernel In-Reply-To: <430E1B7B.2000604@byu.edu> References: <430E1B7B.2000604@byu.edu> Message-ID: <1124998598.4152.24.camel@brilong-lnx.cisco.com> On Thu, 2005-08-25 at 13:26 -0600, Lloyd T Brown wrote: > Yes, the ks log shows the ip address being acquired correctly for > eth0. There's this weird message about it waiting 15 seconds, though. > I'm not sure what that means. The message about the eth0 getting a dhcp > address is followed very shortly by a couple messages stating what the > IP address for the NFS server is and what the path is to the ks.cfg > file, so I'm pretty sure it's working correctly. After that, messages > about it trying to acquire a dhcp address for eth1 come too, including a > failure message. I'll see if I can get something more exact and post > it, if that'd be helpful. Do you have "network --bootproto=dhcp --device=eth0" in ks.cfg? If you have zero "network" lines, the docs say it's only supposed to use eth0, but I wonder if this might be causing your problems. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From harry at mad-cat.co.uk Thu Aug 25 19:51:11 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 25 Aug 2005 20:51:11 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: I would agree with this. It seems that you are mounting the nfs directory successfully and kickstart has downloaded the ks.cfg. Check your ks.cfg for network lines. I have just added this to my ks.cfg for testing: network --device=eth0 network --device=eth1 on top of the entry network --device=eth2 --bootproto dhcp Kickstart is now sending DHCP requests for other network cards and failing - then prompting me for manual input. Try removing all network lines from your ks config except: network --device=eth0 --bootproto dhcp and give it another go. Harry On Thu, 25 Aug, Brian Long wrote: > On Thu, 2005-08-25 at 13:26 -0600, Lloyd T Brown wrote: > > Yes, the ks log shows the ip address being acquired correctly for > > eth0. There's this weird message about it waiting 15 seconds, though. > > I'm not sure what that means. The message about the eth0 getting a dhcp > > address is followed very shortly by a couple messages stating what the > > IP address for the NFS server is and what the path is to the ks.cfg > > file, so I'm pretty sure it's working correctly. After that, messages > > about it trying to acquire a dhcp address for eth1 come too, including a > > failure message. I'll see if I can get something more exact and post > > it, if that'd be helpful. > > Do you have "network --bootproto=dhcp --device=eth0" in ks.cfg? If you > have zero "network" lines, the docs say it's only supposed to use eth0, > but I wonder if this might be causing your problems. > > /Brian/ > -- Harry Mills Chief Technical Manager Opendium Ltd. http://www.opendium.com/ From klaus.steden at thomson.net Thu Aug 25 20:17:26 2005 From: klaus.steden at thomson.net (Klaus Steden) Date: Thu, 25 Aug 2005 16:17:26 -0400 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <20050825201726.GV897@thomson.net> I could have sworn Harry Mills said this Thu, Aug 25, 2005 at 08:07:23PM +0100 ... > > What does the KS log say (+)? Can you see it get the correct IP > from the DHCP server on eth0? On my install (Which is booted from a USB key) > I see it get the network info for eth2 (my onboard card) and then fetch the > ks.cfg. It never attempts to initialise any of the other cards. > Harry, I am interested in switching to USB boot media instead of optical ... got any pointers for someone familiar with the process, but not so familiar with the particular technique? Klaus From harry at mad-cat.co.uk Thu Aug 25 20:52:45 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Thu, 25 Aug 2005 21:52:45 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: Klaus, I played around with doing eveything the 'hard way' - downloading syslinux and creating my own initrd but eventually I simply took diskboot.img from the Fedora installation CD (There is one specifically for HD/USB pen drive booting). This is also syslinux based which makes it very very easy to customize. The one thing I have not done is put together a complete installation on a USB pendrive. The RPMS's are served over http. This is partly because my usb pendrives are only 64Mb but mainly because I wanted to integrate the kickstart and yum repository as described here: http://www.linuxdevcenter.com/pub/a/linux/2005/02/17/kickstart_updates.html If you simply want to boot off the pen drive and fetch the files from elsewhere then it is pretty straight forward - although some BIOSes simply don't want to play - you can get info about this problem if you download the syslinux distribution:- README.usbkey http://syslinux.zytor.com/ I used a 64Mb Easydisk USB 2.0 pen drive. I am not sure there is much more to it but if there is anything else you want to know give me a shout. Regards Harry On Thu, 25 Aug, Klaus Steden wrote: > I could have sworn Harry Mills said this Thu, Aug 25, 2005 at 08:07:23PM +0100 ... > > > > What does the KS log say (+)? Can you see it get the correct IP > > from the DHCP server on eth0? On my install (Which is booted from a USB key) > > I see it get the network info for eth2 (my onboard card) and then fetch the > > ks.cfg. It never attempts to initialise any of the other cards. > > > Harry, > > I am interested in switching to USB boot media instead of optical ... got any > pointers for someone familiar with the process, but not so familiar with the > particular technique? > > Klaus > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -- Harry Mills Chief Technical Manager Opendium Ltd. http://www.opendium.com/ From info at hostinthebox.net Thu Aug 25 21:42:23 2005 From: info at hostinthebox.net (Dan Trainor) Date: Thu, 25 Aug 2005 15:42:23 -0600 Subject: reading /proc/cmdline In-Reply-To: <20050824230606.GR897@thomson.net> References: <430CFAA2.1020707@hostinthebox.net> <20050824230606.GR897@thomson.net> Message-ID: <430E3B3F.6070305@hostinthebox.net> Klaus Steden wrote: >>Hello, all - >> >>I use the following example of code in my kickstart file, to get values >>from /proc/cmdline to use later on during the install: >> >>if grep -i -q "mystring[a-zA-Z0-9]" /proc/cmdline >>then >> MYSTRING=`cat /proc/cmdline | sed 's/.*mystring=\([^ ]*\).*/\1/'` >>fi >> >>I think Klaus in fact gave this to me. However, I've ran into a few >>problems. It seems that this does not like strings with periods in >>them, such as an IP address. >> >>Since I know a bit less than nothing about regular expressions, I was >>hoping that someone could give me a hand with this. I need to modify it >>to allow me to assign a dotted quad string to the MYSTRING variable. >> >>Again, any help would be greatly appreciated. >> > > Hello again, Dan, > > How about ... > > if grep -iq "mystring=[a-zA-Z0-9]" /proc/cmdline > then > MYSTRING=`cat /proc/cmdline | sed 's/.*mystring=\([^ ]*\).*/\1/'` > fi > > Quick testing with sh on Linux tells me that it should preserve the dots. It > should just be deleting everything up to, but not including 'mystring=' and > whatever follows that, and then deleting everything after that string, and > returning the match. > > REs can be tricky, though, so holler if you get stuck. > > Klaus > Klaus - This whole thing was due to sloppyness on my part. I forgot to use the "=" sign. However, it works now. Thanks again! -dant From somewhere_or_other at byu.edu Thu Aug 25 22:38:17 2005 From: somewhere_or_other at byu.edu (Lloyd T Brown) Date: Thu, 25 Aug 2005 16:38:17 -0600 Subject: disabling dhcp for eth1 in kernel In-Reply-To: References: Message-ID: <430E4859.7020300@byu.edu> Actually, I had a line for both eth1 and eth0 in the ks.cfg. By some random coincidence, eth1 was first, although that may not mean much. I commented out the line for eth1, and the install went straight through. Normally, this would probably not set up the eth1 interface at all, but the appropriate files for our really weird network setup are being copied over in the %post script, so everything should be fine. It's installing right now. I'll let you all know if I have any other problems. So, I guess that means that the installer was pulling the ks.cfg over eth0, noticed the eth1 line (static, but without any address) and tried to configure it. Do you suppose that the fact that the eth1 line was above the eth0 line in the file had anything to do with it? I may have to experiment some more tomorrow. Thanks again. You guys have been a great help Lloyd Brown BYU Supercomputing (http://marylou.byu.edu) Harry Mills wrote: >I would agree with this. It seems that you are mounting the nfs directory successfully and kickstart has downloaded the ks.cfg. > >Check your ks.cfg for network lines. I have just added this to my ks.cfg for testing: > >network --device=eth0 >network --device=eth1 > >on top of the entry >network --device=eth2 --bootproto dhcp > >Kickstart is now sending DHCP requests for other network cards and failing - then prompting me for manual input. > >Try removing all network lines from your ks config except: > >network --device=eth0 --bootproto dhcp > >and give it another go. > >Harry > > >On Thu, 25 Aug, Brian Long wrote: > > >>On Thu, 2005-08-25 at 13:26 -0600, Lloyd T Brown wrote: >> >> >>>Yes, the ks log shows the ip address being acquired correctly for >>>eth0. There's this weird message about it waiting 15 seconds, though. >>>I'm not sure what that means. The message about the eth0 getting a dhcp >>>address is followed very shortly by a couple messages stating what the >>>IP address for the NFS server is and what the path is to the ks.cfg >>>file, so I'm pretty sure it's working correctly. After that, messages >>>about it trying to acquire a dhcp address for eth1 come too, including a >>>failure message. I'll see if I can get something more exact and post >>>it, if that'd be helpful. >>> >>> >>Do you have "network --bootproto=dhcp --device=eth0" in ks.cfg? If you >>have zero "network" lines, the docs say it's only supposed to use eth0, >>but I wonder if this might be causing your problems. >> >>/Brian/ >> >> >> > > > From mike at ftl.com Thu Aug 25 22:55:13 2005 From: mike at ftl.com (Mike Smith) Date: Thu, 25 Aug 2005 15:55:13 -0700 Subject: Mass Installation In-Reply-To: <20050825170342.81816.qmail@web80313.mail.yahoo.com> References: <20050825170342.81816.qmail@web80313.mail.yahoo.com> Message-ID: <430E4C51.3020802@ftl.com> Lambert Tran wrote: >Greetings, > >Thanks in advance for all info. >Our task is to install many workstations from a handcrafted system. This is >basically a automated cloning job. We prefer not to use DHCP. This task was >done fairly easy when we used Solaris, and HPUX (I had JumpStart, and Ignite >setup). >While I think KickStart could do the same thing, I am not sure how easy it >could be setup. Any thoughts on unforeseen problems? BTW, we will use RedHat4 >(both AS and ES). > >Thanks, >-ltran > >_______________________________________________ >Kickstart-list mailing list >Kickstart-list at redhat.com >https://www.redhat.com/mailman/listinfo/kickstart-list > > Just curious why you wouldn't want to use DHCP. We are a mixed shop with RedHat, Solaris and HPUX, and we do all of ours via DHCP. Our only obsticle was with some old E420's that didn't have PROM 3.25 or greater. This was needed to allow jumpstarts to happen across subnets. After we did that and set up the 'helpers' on the cisco gear, we have it nailed down to a single jumpstart server and a single dhcp server, thus doing away for a BOOTP server on each subnet. The ip's we are serving are static ip's. DHCP is ONLY used for re-jumping/igniting/kickstarting. If you need some more info, let me know. - Mike Basically we From harry at mad-cat.co.uk Thu Aug 25 23:20:50 2005 From: harry at mad-cat.co.uk (Harry Mills) Date: Fri, 26 Aug 2005 00:20:50 +0100 (BST) Subject: disabling dhcp for eth1 in kernel Message-ID: Glad it's working. On Thu, 25 Aug, Lloyd T Brown wrote: > Actually, I had a line for both eth1 and eth0 in the ks.cfg. By some > random coincidence, eth1 was first, although that may not mean much. I > commented out the line for eth1, and the install went straight through. > Normally, this would probably not set up the eth1 interface at all, but > the appropriate files for our really weird network setup are being > copied over in the %post script, so everything should be fine. It's > installing right now. I'll let you all know if I have any other problems. > > So, I guess that means that the installer was pulling the ks.cfg over > eth0, noticed the eth1 line (static, but without any address) and tried > to configure it. Do you suppose that the fact that the eth1 line was > above the eth0 line in the file had anything to do with it? I may have > to experiment some more tomorrow. I think this is exactly what happened, but I am sure the order makes no difference. KS tried to configure eth1 because it was included in the ks.cfg but was lacking a complete confgiuration. The proof of this would be to put the line back in but with all the config info for it included. > Thanks again. You guys have been a great help > Lloyd Brown > BYU Supercomputing (http://marylou.byu.edu) > > :) Harry From michael.boman at gmail.com Sun Aug 28 04:02:11 2005 From: michael.boman at gmail.com (Michael Boman) Date: Sun, 28 Aug 2005 12:02:11 +0800 Subject: Kickstart fails when trying to install extra packages Message-ID: <48989dd705082721027dbd5f11@mail.gmail.com> I have a ISO based on a RHEL 4 clone (CentOS 4) which has been updated with the latest release of the available packages and had some extra packages added to it. The problem is that when I try to install the extra packages in the %packages section the kickstart just fails, while the ISO as a whole works just fine without the kickstart (can do manual installation), minus the extra packages. Do I need to modify comps.xml now days to get it to work, or have I missed something? I never needed to modify comps.xml when I used Fedora Core 1 as a base. I have attached the kickstart configuration to this email. Best regards Michael Boman -- IT Security Researcher & Developer http://proxy.11a.nu -------------- next part -------------- A non-text attachment was scrubbed... Name: ks.cfg Type: application/octet-stream Size: 1752 bytes Desc: not available URL: From xagonzalezm at gmail.com Sun Aug 28 17:42:08 2005 From: xagonzalezm at gmail.com (Alberto) Date: Sun, 28 Aug 2005 19:42:08 +0200 Subject: Boot in rescue mode and launch a command?? Message-ID: <7fa4b410508281042eb86639@mail.gmail.com> Hi, I am trying to create a ks.cfg file to boot into rescue mode. I get everything ok until it ask me last question about if I want to look for hd and mount it. I want to Skip it automatically and then go to shell prompt and launch a command. How can I add that to the ks.cfg file?? This is my ks file: $ cat wbel4-rescue-ks.cfg # Kickstart configuration file RHEL/WBEL 4 Rescue Mode # #System language # lang en_US.UTF-8 # #Language modules to install # langsupport --default=en_GB.UTF-8 en_GB.UTF-8 # #System keyboard # keyboard uk # #System mouse # mouse none # #Retrieve rescue system from NFS # nfs --server=XX.XX.XX.XX --dir=/whitebox/4/en/os/i386 # #Network information # network --bootproto=dhcp And /tftpboot/linux-install/pxelinux.cfg/default label 3 kernel wbel4/vmlinuz append initrd=wbel4/initrd.img ramdisk_size=10000 text rescue ks=nfs:XX.XX.XX.XX:/kickstart/wbel4-rescue-ks.cfg From teng at dataway.com Mon Aug 29 18:14:04 2005 From: teng at dataway.com (Tedman Eng) Date: Mon, 29 Aug 2005 11:14:04 -0700 Subject: Boot in rescue mode and launch a command?? Message-ID: <37ED92F9890FAF4BB947613C66FF8B1A08BB2B76@dw-mail.dataway.com> Add "nomount" to your kernel append line. nomount Don't automatically mount any installed Linux partitions in rescue mode. Found docs mirrored online here: http://docs.rage.net/system/anaconda-9.0/command-line.txt > -----Original Message----- > From: Alberto [mailto:xagonzalezm at gmail.com] > Sent: Sunday, August 28, 2005 10:42 AM > To: kickstart-list at redhat.com > Subject: Boot in rescue mode and launch a command?? > > > Hi, > > I am trying to create a ks.cfg file to boot into rescue mode. I get > everything ok until it ask me last question about if I want to look > for hd and mount it. I want to Skip it automatically and then go to > shell prompt and launch a command. How can I add that to the ks.cfg > file?? This is my ks file: > > > > $ cat wbel4-rescue-ks.cfg > # Kickstart configuration file RHEL/WBEL 4 Rescue Mode > # > #System language > # > lang en_US.UTF-8 > # > #Language modules to install > # > langsupport --default=en_GB.UTF-8 en_GB.UTF-8 > # > #System keyboard > # > keyboard uk > # > #System mouse > # > mouse none > # > #Retrieve rescue system from NFS > # > nfs --server=XX.XX.XX.XX --dir=/whitebox/4/en/os/i386 > # > #Network information > # > network --bootproto=dhcp > > > > And /tftpboot/linux-install/pxelinux.cfg/default > > label 3 > kernel wbel4/vmlinuz > append initrd=wbel4/initrd.img ramdisk_size=10000 text rescue > ks=nfs:XX.XX.XX.XX:/kickstart/wbel4-rescue-ks.cfg > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > From info at hostinthebox.net Mon Aug 29 23:22:58 2005 From: info at hostinthebox.net (Dan Trainor) Date: Mon, 29 Aug 2005 17:22:58 -0600 Subject: Echo'ing data to vt3 when running %post Message-ID: <431398D2.7070605@hostinthebox.net> Hello, all - I have a few applications taht are run during %post, which I've been having trouble with - mainly with data input corruption. I was trying to see if it was possible to echo all output frmo a %post section to vt3, or even better, a file. Anyone know how I can do this? Thanks! -adnt From skvidal at phy.duke.edu Tue Aug 30 00:05:34 2005 From: skvidal at phy.duke.edu (seth vidal) Date: Mon, 29 Aug 2005 20:05:34 -0400 Subject: Echo'ing data to vt3 when running %post In-Reply-To: <431398D2.7070605@hostinthebox.net> References: <431398D2.7070605@hostinthebox.net> Message-ID: <1125360334.2001.3.camel@cutter> On Mon, 2005-08-29 at 17:22 -0600, Dan Trainor wrote: > Hello, all - > > I have a few applications taht are run during %post, which I've been > having trouble with - mainly with data input corruption. > > I was trying to see if it was possible to echo all output frmo a %post > section to vt3, or even better, a file. > > Anyone know how I can do this? > Something cool I picked up from the admin of the math department here at duke: %post --nochroot ln -s /mnt/sysimage/etc/termcap /etc/termcap # download the post.sh into the chroot chroot /mnt/sysimage /usr/bin/wget -O /tmp/post.sh \ http://someplace/post.sh >>/mnt/sysimage/root/foo.output 2>&1 chmod +x /mnt/sysimage/tmp/post.sh # run it in an xterm DISPLAY=:1.0 xterm -geometry 80x25+150+160 -e chroot /mnt/sysimage \ /tmp/post.sh >>/mnt/sysimage/root/foo.output 2>&1 /bin/rm -f /mnt/sysimage/tmp/post.sh so in a graphical install it will pop up the xterm to run your %post :) -sv From brilong at cisco.com Tue Aug 30 12:00:18 2005 From: brilong at cisco.com (Brian Long) Date: Tue, 30 Aug 2005 08:00:18 -0400 Subject: Echo'ing data to vt3 when running %post In-Reply-To: <1125360334.2001.3.camel@cutter> References: <431398D2.7070605@hostinthebox.net> <1125360334.2001.3.camel@cutter> Message-ID: <1125403218.3964.13.camel@brilong-lnx.cisco.com> On Mon, 2005-08-29 at 20:05 -0400, seth vidal wrote: > On Mon, 2005-08-29 at 17:22 -0600, Dan Trainor wrote: > > Hello, all - > > > > I have a few applications taht are run during %post, which I've been > > having trouble with - mainly with data input corruption. > > > > I was trying to see if it was possible to echo all output frmo a %post > > section to vt3, or even better, a file. > > > > Anyone know how I can do this? > > > > Something cool I picked up from the admin of the math department here at > duke: > > > %post --nochroot > ln -s /mnt/sysimage/etc/termcap /etc/termcap > > # download the post.sh into the chroot > chroot /mnt/sysimage /usr/bin/wget -O /tmp/post.sh \ > http://someplace/post.sh >>/mnt/sysimage/root/foo.output 2>&1 > > > chmod +x /mnt/sysimage/tmp/post.sh > # run it in an xterm > DISPLAY=:1.0 xterm -geometry 80x25+150+160 -e chroot /mnt/sysimage \ > /tmp/post.sh >>/mnt/sysimage/root/foo.output 2>&1 > > /bin/rm -f /mnt/sysimage/tmp/post.sh > > > > so in a graphical install it will pop up the xterm to run your %post :) This is very cool. I guess I could look at /proc/cmdline to verify the installer wasn't running with "text", "serial" or "console=ttyS*" before bringing up the xterm. Right now, I just "chvt 3", run my %post commands and then "chvt 1" which is broken in graphical installs but works well on HP Proliants doing text-based installs. /Brian/ -- Brian Long | | | IT Data Center Systems | .|||. .|||. Cisco Linux Developer | ..:|||||||:...:|||||||:.. Phone: (919) 392-7363 | C i s c o S y s t e m s From xagonzalezm at gmail.com Tue Aug 30 22:03:37 2005 From: xagonzalezm at gmail.com (Alberto) Date: Wed, 31 Aug 2005 00:03:37 +0200 Subject: Boot in rescue mode and launch a command?? In-Reply-To: <37ED92F9890FAF4BB947613C66FF8B1A08BB2B76@dw-mail.dataway.com> References: <37ED92F9890FAF4BB947613C66FF8B1A08BB2B76@dw-mail.dataway.com> Message-ID: <7fa4b41050830150351f561ff@mail.gmail.com> If I add nomount I cant see my scsi disk device. If I boot without nomount and choose skip I see the scsi device disk ( fdisk -l ). Any other option that asign a device to my disk booting with nomout?? And what about launching a command... %post doesnt seems to work. On 8/29/05, Tedman Eng wrote: > > Add "nomount" to your kernel append line. > > nomount Don't automatically mount any installed Linux partitions > in rescue mode. > > > Found docs mirrored online here: > http://docs.rage.net/system/anaconda-9.0/command-line.txt > > > > > > -----Original Message----- > > From: Alberto [mailto:xagonzalezm at gmail.com] > > Sent: Sunday, August 28, 2005 10:42 AM > > To: kickstart-list at redhat.com > > Subject: Boot in rescue mode and launch a command?? > > > > > > Hi, > > > > I am trying to create a ks.cfg file to boot into rescue mode. I get > > everything ok until it ask me last question about if I want to look > > for hd and mount it. I want to Skip it automatically and then go to > > shell prompt and launch a command. How can I add that to the ks.cfg > > file?? This is my ks file: > > > > > > > > $ cat wbel4-rescue-ks.cfg > > # Kickstart configuration file RHEL/WBEL 4 Rescue Mode > > # > > #System language > > # > > lang en_US.UTF-8 > > # > > #Language modules to install > > # > > langsupport --default=en_GB.UTF-8 en_GB.UTF-8 > > # > > #System keyboard > > # > > keyboard uk > > # > > #System mouse > > # > > mouse none > > # > > #Retrieve rescue system from NFS > > # > > nfs --server=XX.XX.XX.XX --dir=/whitebox/4/en/os/i386 > > # > > #Network information > > # > > network --bootproto=dhcp > > > > > > > > And /tftpboot/linux-install/pxelinux.cfg/default > > > > label 3 > > kernel wbel4/vmlinuz > > append initrd=wbel4/initrd.img ramdisk_size=10000 text rescue > > ks=nfs:XX.XX.XX.XX:/kickstart/wbel4-rescue-ks.cfg > > > > _______________________________________________ > > Kickstart-list mailing list > > Kickstart-list at redhat.com > > https://www.redhat.com/mailman/listinfo/kickstart-list > > > > _______________________________________________ > Kickstart-list mailing list > Kickstart-list at redhat.com > https://www.redhat.com/mailman/listinfo/kickstart-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From uwelists at u-b.de Wed Aug 31 20:54:32 2005 From: uwelists at u-b.de (Uwe E. Bilger) Date: Wed, 31 Aug 2005 22:54:32 +0200 Subject: Installing a (compressed) image over nfs Message-ID: Hello all, I'm trying to create a recovery scenario where, upon recovery request, an image is loaded from a server to the system via nfs. My attempt so far was to do a minimum kickstart installation, basically installing nothing, but rather "waiting" for the %post section. But from the beginng: >From the system to be recovered, I took an image with dd and transferred it over to the server, compressing it there: dd if=/dev/hda | gzip -c /nfsmount/host2-hda.gz On recovery, the post section is supposed to overwrite the whole HD with this particular image again: [...] #Firewall configuration firewall --disabled #Do not configure XWindows skipx #Package install information %packages --resolvedeps %post mkdir /mnt/server mount -t nfs 192.168.2.254:/backup/images /mnt/vitels gunzip -c /mnt/vitels/host2-hda.gz | dd of=/dev/hda And this aint working at all. Any ideas why? Below is the rest of the ks.cfg I'm also open to suggestions on how else I can resolve this. I didn't find any documentation about the rescue option. Could that be a possibility? Can you point me to some docs? Thank you so much! Uwe #System language lang en_US #Language modules to install langsupport en_US #System keyboard keyboard de-latin1 #System mouse mouse genericwheelps/2 #Sytem timezone timezone America/New_York #Root password rootpw --iscrypted $1$LBYw4gs2$kV4hSBc22FQDabrINN5sn/ #Reboot after installation reboot #Install OS instead of upgrade install #Use Web installation url --url http://192.168.2.254/fedora/4 #System bootloader configuration bootloader --location=mbr #Clear the Master Boot Record zerombr yes #Partition clearing information clearpart --all --initlabel #Disk partitioning information part /boot --fstype ext3 --size 100 part / --fstype ext3 --size 1 --grow part swap --size 1020 #System authorization infomation auth --useshadow --enablemd5 #Network information network --bootproto=dhcp --device=eth0 network --bootproto=static --ip=10.0.2.1 --netmask=255.255.255.0 --gat eway=10.0. 2.254 --nameserver=10.0.2.254 --device=eth1 #Firewall configuration firewall --disabled #Do not configure XWindows skipx #Package install information %packages --resolvedeps %post From phr at doc.ic.ac.uk Wed Aug 31 21:33:07 2005 From: phr at doc.ic.ac.uk (Philip Rowlands) Date: Wed, 31 Aug 2005 22:33:07 +0100 (BST) Subject: Installing a (compressed) image over nfs In-Reply-To: References: Message-ID: On Wed, 31 Aug 2005, Uwe E. Bilger wrote: >On recovery, the post section is supposed to overwrite the whole HD >with this particular image again: > >[...] >#Firewall configuration >firewall --disabled >#Do not configure XWindows >skipx >#Package install information >%packages --resolvedeps >%post >mkdir /mnt/server >mount -t nfs 192.168.2.254:/backup/images /mnt/vitels >gunzip -c /mnt/vitels/host2-hda.gz | dd of=/dev/hda > > > >And this aint working at all. > >Any ideas why? A few thoughts: What did happen? Was there any useful output on the virtual consoles? Why is /mnt/server created then not used? Do you have any HDD partitions still mounted when %post is reached? Cheers, Phil