From rmcdaniel at indata.us Tue Nov 4 03:15:47 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Mon, 03 Nov 2008 20:15:47 -0700 Subject: single interface setup Message-ID: <20081103201547.d7061e97b78b017ac15395d64f2ce134.84f3dbddce.wbe@email.secureserver.net> I finally got the setup working with just a single interface. All that I did was configure eth0 as my main interface and change the dhcpd.conf file with the correct subnet information. I just disabled eth1 and left the bridge interface as it was setup from the initial install. I will probably get rid of the bridge interface just to remove any possible problems. Everything is working GREAT! Has anyone tried using some of the scripts that were included with K12LTSP, ie. "Add/Remove Icons to Desktops"? Thanks, Ron McDaniel rmcdaniel at indata.us From monteslu at cox.net Tue Nov 4 05:10:28 2008 From: monteslu at cox.net (Luis Montes) Date: Mon, 03 Nov 2008 22:10:28 -0700 Subject: single interface setup In-Reply-To: <20081103201547.d7061e97b78b017ac15395d64f2ce134.84f3dbddce.wbe@email.secureserver.net> References: <20081103201547.d7061e97b78b017ac15395d64f2ce134.84f3dbddce.wbe@email.secureserver.net> Message-ID: <490FD944.4090306@cox.net> I though I did just that and it didn't work for me. Could you post your exact configuration changes? Thanks, Luis rmcdaniel at indata.us wrote: > I finally got the setup working with just a single interface. All that > I did was configure eth0 as my main interface and change the dhcpd.conf > file with the correct subnet information. I just disabled eth1 and left > the bridge interface as it was setup from the initial install. I will > probably get rid of the bridge interface just to remove any possible > problems. Everything is working GREAT! Has anyone tried using some of > the scripts that were included with K12LTSP, ie. "Add/Remove Icons to > Desktops"? > > > Thanks, > > Ron McDaniel > rmcdaniel at indata.us > > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > From peter at scheie.homedns.org Tue Nov 4 13:20:49 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Tue, 04 Nov 2008 07:20:49 -0600 Subject: single interface setup In-Reply-To: <20081103201547.d7061e97b78b017ac15395d64f2ce134.84f3dbddce.wbe@email.secureserver.net> References: <20081103201547.d7061e97b78b017ac15395d64f2ce134.84f3dbddce.wbe@email.secureserver.net> Message-ID: <49104C31.3080607@scheie.homedns.org> The 'Add/Remove Icons' script should work just fine on K12Linux because the underlying user accounts and directories hasn't changed. Just copy the push_icons.desktop file and the /usr/sbin/push-icons-to-users-desktops script from from a K12LTSP server and you should be good to go. Peter rmcdaniel at indata.us wrote: > I finally got the setup working with just a single interface. All that > I did was configure eth0 as my main interface and change the dhcpd.conf > file with the correct subnet information. I just disabled eth1 and left > the bridge interface as it was setup from the initial install. I will > probably get rid of the bridge interface just to remove any possible > problems. Everything is working GREAT! Has anyone tried using some of > the scripts that were included with K12LTSP, ie. "Add/Remove Icons to > Desktops"? > > > Thanks, > > Ron McDaniel > rmcdaniel at indata.us > > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > From odin at gnuskole.no Tue Nov 4 19:03:11 2008 From: odin at gnuskole.no (=?ISO-8859-1?Q?Odin_N=F8sen?=) Date: Tue, 4 Nov 2008 20:03:11 +0100 Subject: vncts In-Reply-To: <4909BF58.3080608@redhat.com> References: <6befb72f0810300417q7d1f31e2p881cf893a7340300@mail.gmail.com> <4909BF58.3080608@redhat.com> Message-ID: <20081104190051.M35269@gnuskole.no> I can't get vncts to work in fedora 9 (og 10). I only get the gray screen when I try to "vncviewer mymachine:5903". Any hints? I can't find a Xsession-file to edit - which is what I'm supposed to according to the vncts-file in xinetd.d Odin From rmcdaniel at indata.us Tue Nov 4 21:24:29 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Tue, 04 Nov 2008 14:24:29 -0700 Subject: single interface setup Message-ID: <20081104142429.d7061e97b78b017ac15395d64f2ce134.a31ccde70b.wbe@email.secureserver.net> Here are the contents of the relevant files with explanation of our IP scheme: My subnet is 192.168.5.X/24 My gateway off of the subnet is 192.168.5.1/24 (Cisco Router) My LTSP server address is 192.168.5.3/24 I have disabled eth1 from starting up I completely left the bridge interface as it was setup from the installation I loaded the image from a flash drive and selected all of the defaults, except I chose eth0 as my main interface and gave it the static settings shown above. I also inputted our DNS settings during the NIC configuration step. When the server first booted, I tested Internet connectivity. I downloaded/installed Webmin, www.webmin.com. Using Webmin, I shutdown unnecessary services, ie. bluetooth, isdn, and IP Tables. I restarted the server and then booted our thin clients. Clients booted just fine with the 192.168.5.X subnet. /etc/sysconfig/network-scripts/ifcfg-eth0 # Broadcom Corporation NetXtreme II BCM5708 Gigabit Ethernet DEVICE=eth0 BOOTPROTO=static BROADCAST=192.168.5.255 HWADDR=00:1c:c4:ed:27:b4 IPADDR=192.168.5.3 NETMASK=255.255.255.0 NETWORK=192.168.5.0 ONBOOT=yes DNS1=129.66.76.4 DNS2=129.66.95.3 NM_CONTROLLED= /etc/sysconfig/network-scripts/ifcfg-eth1 # Broadcom Corporation NetXtreme II BCM5708 Gigabit Ethernet DEVICE=eth1 BOOTPROTO=dhcp HWADDR=00:1c:c4:ed:27:b2 IPADDR=dhcp ONBOOT=no DHCP_HOSTNAME=LJH-LTSP DNS1=129.66.76.4 DNS2=129.66.95.3 NM_CONTROLLED= /etc/sysconfig/network-scripts/ifcfg-lo DEVICE=lo IPADDR=127.0.0.1 NETMASK=255.0.0.0 NETWORK=127.0.0.0 # If you're having problems with gated making 127.0.0.0/8 a martian, # you can change this to something else (255.255.255.255, for example) BROADCAST=127.255.255.255 ONBOOT=yes NAME=loopback /etc/sysconfig/network-scripts/ifcfg-ltspbr0 # Sample Configuration for Initscript Driven Bridge # Put this into /etc/sysconfig/network-scripts/ifcfg-ltspbr0 # Note that this network range does not match the defaults in dhcpd.conf DEVICE=ltspbr0 TYPE=Bridge ONBOOT=yes BOOTPROTO=static STP=off DELAY=0.1 USERCTL=no IPV6INIT=no PEERDNS=yes NETMASK=255.255.255.0 IPADDR=172.31.100.254 ARP=yes /etc/ltsp/dhcpd.conf # Sample configuration file for LTSP dhcpd # default-lease-time 21600; max-lease-time 21600; ddns-update-style none; allow booting; allow bootp; option subnet-mask 255.255.255.0; option broadcast-address 192.168.5.255; option routers 192.168.5.1; option domain-name-servers 129.66.76.4,129.66.95.3; next-server 192.168.5.3; option domain-name "ltsp"; option option-128 code 128 = string; option option-129 code 129 = text; option option-221 code 221 = text; shared-network WORKSTATIONS { subnet 192.168.5.0 netmask 255.255.255.0 { range dynamic-bootp 192.168.5.50 192.168.5.253; use-host-decl-names on; option log-servers 192.168.5.3; # trick from Peter Rundle # newer PPC Macs if substring (option vendor-class-identifier, 0, 9) = "AAPLBSDPC" { filename "yaboot"; option vendor-class-identifier "AAPLBSDPC"; } # really old ppc iMacs elsif substring (option option-221, 0, 5) = "Apple" { filename "yaboot"; option vendor-class-identifier "AAPLBSDPC"; } # Etherboot ELF (only 5.4), should work with Coreboot elsif substring (option vendor-class-identifier, 0, 13) = "Etherboot-5.4" { filename "/ltsp/i386/elf.ltsp"; } # Etherboot NBI (older clients) elsif substring (option vendor-class-identifier, 0, 9) = "Etherboot" { filename "/ltsp/i386/wraplinux-nbi.ltsp"; } # PXE elsif substring (option vendor-class-identifier, 0, 9) = "PXEClient" { # NOTE: kernels are specified in /tftpboot/ltsp/i386/pxelinux.cfg/ filename "/ltsp/i386/pxelinux.0"; } # if all else fails (likely BOOTP), default to an NBI image else { filename "/ltsp/i386/wraplinux-nbi.ltsp"; } if substring (option vendor-class-identifier, 20, 3) = "ppc" { option root-path "192.168.5.3:/opt/ltsp/ppc"; } else { option root-path "192.168.5.3:/opt/ltsp/i386"; #option root-path "nbd:192.168.5.3:2000:squashfs:ro"; } } } ## example configurations for specifying specific kernels to specific clients #group { # use-host-decl-names on; # option log-servers 192.168.5.3; # # host ws001 { # hardware ethernet 00:E0:06:E8:00:84; # fixed-address 172.31.100.1; # filename "/ltsp/vmlinuz.ltsp"; # option option-128 e4:45:74:68:00:00; # option option-129 "NIC=3c509"; # } # host ws002 { # hardware ethernet 00:D0:09:30:6A:1C; # fixed-address 172.31.100.2; # filename "/ltsp/vmlinuz.ltsp"; # option option-128 e4:45:74:68:00:00; # option option-129 "NIC=ne"; # } # host ws003 { # hardware ethernet 00:D0:09:30:28:B2; # fixed-address 172.31.100.3; # # kernels are specified in /tftpboot/ltsp/i386/pxelinux.cfg/ # filename "/ltsp/i386/pxelinux.0"; # } # # Apple Specific Settings # host ws007 { # hardware ethernet 00:30:65:69:23:60; # fixed-address 172.31.100.4; # option root-path "192.168.5.3:/opt/ltsp/ppc"; # filename "yaboot"; # option vendor-class-identifier "AAPLBSDPC"; # } #} I hope that this is useful! Ron Ron McDaniel rmcdaniel at indata.us Conecuh County Schools Technology Coordinator -------- Original Message -------- Subject: Re: single interface setup From: Luis Montes Date: Mon, November 03, 2008 11:10 pm To: Development discussion of K12Linux I though I did just that and it didn't work for me. Could you post your exact configuration changes? Thanks, Luis rmcdaniel at indata.us wrote: > I finally got the setup working with just a single interface. All that > I did was configure eth0 as my main interface and change the dhcpd.conf > file with the correct subnet information. I just disabled eth1 and left > the bridge interface as it was setup from the initial install. I will > probably get rid of the bridge interface just to remove any possible > problems. Everything is working GREAT! Has anyone tried using some of > the scripts that were included with K12LTSP, ie. "Add/Remove Icons to > Desktops"? > > > Thanks, > > Ron McDaniel > rmcdaniel at indata.us > > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From william at fragakis.com Wed Nov 5 15:26:51 2008 From: william at fragakis.com (William Fragakis) Date: Wed, 05 Nov 2008 10:26:51 -0500 Subject: Migrating legacy k12ltsp to k12linux In-Reply-To: <20081104170040.A6B188E0210@hormel.redhat.com> References: <20081104170040.A6B188E0210@hormel.redhat.com> Message-ID: <1225898811.7571.34.camel@server.ltsp> Hope this helps someone. Please forgive any errors. I've been able to move a F9 (originally K12LTSP based on f7) to k12linux, back and again forward to k12linux based on these steps. Excuse the length but I wanted to be thorough. Comments, corrections, etc. welcome. Regards, William --- Moving Fedora 9 K12LTSP (ltsp 4.2) to k12linux/ltsp 5 Edits/Corrections/Flames to william at fragakis.com This uses legacy network configuration of eth0 facing clients and eth1 facing outside world. This simplifies certain network setup issues (ie uses existing network profile), can use existing dhcpd and existing iptable rules. This is meant to be reversable; if you need to, you can revert to your previous configuration quickly and without restarting the server. The process doesn't take long - building the new ltsp image takes the longest - and doesn't require restarting your server. Test system started as F7 and has been updated, over time, to F9. (Because I'm lazy and didn't want to figure out everthing that i need to reinstall) A familiarity with https://fedorahosted.org/k12linux/wiki/InstallGuide will be a big plus as well as having set up a K12Linux system from scratch just to become comfortable with any differences and nuances in the new setup. Backup to safe location: /etc/cron.daily/ltsp-swap-delete (or will be duplicate file from new package) /etc/xinetd.d/tftp (xinetd will see any file in /etc/xinetd/ as a service, you'll end up with multiple, conflicting tftp configs if you leave this file in. e.g. tftp and tftp.rpmnew) Rename from /opt/ltsp to /opt/ltsp4 (or something suitable, for our purposes here, we'll say /opt/ltsp4) Install new components ## yum install ltsp-server will fail because /etc/cron.daily/ltsp-swap-delete is part of ## the old ltsp package. We could bomb ## the old ltsp package but we are trying to keep the old setup "just in case". # yum install ltspfs nbd. bridge-utils livecd-tools squashfs-tools pykickstart live555 Download latest ltsp-server from http://togami.com/~k12linux-temporary/fedora/9/ depending upon architecture # wget http://togami.com/~k12linux-temporary/fedora/9/i386/ltsp-server-5.1.32-1.fc9.i386.rpm # rpm -ivh /ltsp-server-5.1.32-1.fc9.i386.rpm --replacefile transfer any machine specific settings - especially if using thin clients as print servers transfer from: /opt/ltsp4/i386/etc/lts.conf to: /var/lib/tftpboot/ltsp/i386/lts.conf simply replacing the file is not recommended because certain settings in LTSP4 are not recognized in 5 and ver 5 does a good job of recognizing a lot of hardware that ver. 4 struggled with (e.g. my ISA soundcards and PCI video cards). see: http://www.ltsp.org/~sbalneav/LTSPManual.html#lts-conf for more At this point, if you wish you can migrate to new style of bridge and virtual network interface serving the clients. Please consult K12Linux docs on advanced network setup. https://fedorahosted.org/k12linux/wiki/NetworkSetup ## Ugly simplification if you are using only PXE boot thin clients and have a bunch of fixed ips you don't want to have to copy over in /etc/dhcpd-k12ltsp.conf verify that this line is correct for your system option root-path "192.168.0.254:/opt/ltsp/i386"; and change # NOTE: kernels are specified in /tftpboot/lts/pxe/pxelinux.cfg/ filename "/lts/pxe/pxelinux.0"; to # NOTE: kernels are specified in /tftpboot/lts/pxe/pxelinux.cfg/ filename "/ltsp/i386/pxelinux.0"; ##better: option rename /etc/dhcpd-k12ltsp.conf to etc/dhcpd-k12ltsp.conf.bk and copy to /etc/ltsp/dhcpd.conf to /etc/dhcpd-k12ltsp.conf you may want to replace all instances of 172.31.100 with 192.168.0 if you are fond of the previous ip range. Also, make any appropriate changes if you are using a static ip addressing scheme. Restart existing services # service dhcpd restart # service xinetd restart Start new services # for server in ldminfod nbdrootd nbdswapd; do chkconfig $server on; done Note that other services mentioned in the k12linux docs should already be running. Lastly, one of the reasons you did this is to take advantage of newer versions of flash. Go to adobe.com and download the latest version. If you already have the adobe repositories in your yum configuration - # yum update -y flash-plugin Consult http://macromedia.mplug.org/ for any tweaks. This one is highly recommended: Create /etc/adobe/mms.cfg containing this line WindowlessDisable=true and remove libflashsupport as it is no longer needed for anything with Flash 10, and it might even be the cause of crashes. # yum remove -y libflashsupport ### Undoing the damage If you need to revert, for whatever reason, to using ltsp ver. 4: change in /etc/xinetd.d/tftp from server_args = -s /var/lib/tftpboot to server_args = -s /tftpboot If you wish to keep two ltsp directories, you can do this: change in /etc/exports from /opt/ltsp/i386 192.168.0.0/255.255.255.0(ro,no_root_squash,sync) to /opt/ltsp4/ i 386 192.168.0.0/255.255.255.0(ro,no_root_squash,sync) and from /opt/ltsp/ppc 192.168.0.0/255.255.255.0(ro,no_root_squash,sync) to /opt/ltsp4/ppc 192.168.0.0/255.255.255.0(ro,no_root_squash,sync) or rename from /opt/ltsp to /opt/ltsp5 and /opt/ltsp4 to /opt/ltsp switch dhcpd-k12ltsp.conf files (make .bk the current one and the current one .bk or .new as you wish) or aim pxe filename to lts (not ltsp) directory - depending whether you swapped .conf files or hacked the original Don't forget to replace all instances of /opt/ltsp with /opt/ltsp4, or whatever you renamed your original ltsp directory! restart xinetd, nfs nfslock and dhcpd you should be able to boot back in to ltsp 4 ## Troubleshooting tftp fails Did you restart xinetd? Is the correct tftp file in /etc/xinetd.d/ ? If using your previous dhcpd file, is the correct filename and path set for your new configuration (e.g. ilename "/ltsp/i386/pxelinux.0"; )? pxelinux boots with old/incorrect vmlinuz tftp points to wrong location. New location is /var/lib/tftpboot. Ver. 4.2 uses /tftpboot boot hangs after starting redhat nash Does the directory holding the ltsp files match your settings elsewhere? Check that option root-path points to correct ip and directory e.g. option root-path "192.168.0.254:/opt/ltsp/i386"; After reverting to previous ltsp, you get nfs errors Make sure you restart nfs and nfslock if you've edited /etc/exports Make sure that /etc/exports has the correct path (e.g. /opt/ltsp4/i386 192.168.0.0/255.255.255.0(ro,no_root_squash,sync) ) Make sure that your dhcpd file (e.g. dhcpd-k12ltsp.conf) points option root-path to the correct location. From monteslu at cox.net Wed Nov 5 16:01:37 2008 From: monteslu at cox.net (monteslu at cox.net) Date: Wed, 5 Nov 2008 8:01:37 -0800 Subject: single interface setup In-Reply-To: <20081104142429.d7061e97b78b017ac15395d64f2ce134.a31ccde70b.wbe@email.secureserver.net> Message-ID: <20081105110137.YAMPQ.259732.imail@fed1rmwml38> This helps a ton, thanks for posting it. I think the main difference in our setups is that you IP tables disabled. I'm going to give that a shot this weekend. I'll follow up on the list and try to update the networking section of the wiki if I can get it to work. Luis ---- rmcdaniel at indata.us wrote: > Here are the contents of the relevant files with explanation of our IP > scheme: > > My subnet is 192.168.5.X/24 > My gateway off of the subnet is 192.168.5.1/24 (Cisco Router) > My LTSP server address is 192.168.5.3/24 > I have disabled eth1 from starting up > I completely left the bridge interface as it was setup from the > installation > > I loaded the image from a flash drive and selected all of the defaults, > except I chose eth0 as my main interface and gave it the static settings > shown above. I also inputted our DNS settings during the NIC > configuration step. When the server first booted, I tested Internet > connectivity. I downloaded/installed Webmin, www.webmin.com. Using > Webmin, I shutdown unnecessary services, ie. bluetooth, isdn, and IP > Tables. I restarted the server and then booted our thin clients. > Clients booted just fine with the 192.168.5.X subnet. > > > > /etc/sysconfig/network-scripts/ifcfg-eth0 > > # Broadcom Corporation NetXtreme II BCM5708 Gigabit Ethernet > DEVICE=eth0 > BOOTPROTO=static > BROADCAST=192.168.5.255 > HWADDR=00:1c:c4:ed:27:b4 > IPADDR=192.168.5.3 > NETMASK=255.255.255.0 > NETWORK=192.168.5.0 > ONBOOT=yes > DNS1=129.66.76.4 > DNS2=129.66.95.3 > NM_CONTROLLED= > > /etc/sysconfig/network-scripts/ifcfg-eth1 > > # Broadcom Corporation NetXtreme II BCM5708 Gigabit Ethernet > DEVICE=eth1 > BOOTPROTO=dhcp > HWADDR=00:1c:c4:ed:27:b2 > IPADDR=dhcp > ONBOOT=no > DHCP_HOSTNAME=LJH-LTSP > DNS1=129.66.76.4 > DNS2=129.66.95.3 > NM_CONTROLLED= > > /etc/sysconfig/network-scripts/ifcfg-lo > > DEVICE=lo > IPADDR=127.0.0.1 > NETMASK=255.0.0.0 > NETWORK=127.0.0.0 > # If you're having problems with gated making 127.0.0.0/8 a martian, > # you can change this to something else (255.255.255.255, for example) > BROADCAST=127.255.255.255 > ONBOOT=yes > NAME=loopback > > /etc/sysconfig/network-scripts/ifcfg-ltspbr0 > > # Sample Configuration for Initscript Driven Bridge > # Put this into /etc/sysconfig/network-scripts/ifcfg-ltspbr0 > # Note that this network range does not match the defaults in dhcpd.conf > DEVICE=ltspbr0 > TYPE=Bridge > ONBOOT=yes > BOOTPROTO=static > STP=off > DELAY=0.1 > USERCTL=no > IPV6INIT=no > PEERDNS=yes > NETMASK=255.255.255.0 > IPADDR=172.31.100.254 > ARP=yes > > /etc/ltsp/dhcpd.conf > > # Sample configuration file for LTSP dhcpd > # > > default-lease-time 21600; > max-lease-time 21600; > ddns-update-style none; > allow booting; > allow bootp; > > option subnet-mask 255.255.255.0; > option broadcast-address 192.168.5.255; > option routers 192.168.5.1; > option domain-name-servers 129.66.76.4,129.66.95.3; > next-server 192.168.5.3; > option domain-name "ltsp"; > option option-128 code 128 = string; > option option-129 code 129 = text; > option option-221 code 221 = text; > > shared-network WORKSTATIONS { > subnet 192.168.5.0 netmask 255.255.255.0 { > range dynamic-bootp 192.168.5.50 192.168.5.253; > use-host-decl-names on; > option log-servers 192.168.5.3; > > # trick from Peter Rundle > # newer PPC Macs > if substring (option vendor-class-identifier, 0, 9) = "AAPLBSDPC" > { > filename "yaboot"; > option vendor-class-identifier "AAPLBSDPC"; > } > # really old ppc iMacs > elsif substring (option option-221, 0, 5) = "Apple" > { > filename "yaboot"; > option vendor-class-identifier "AAPLBSDPC"; > } > # Etherboot ELF (only 5.4), should work with Coreboot > elsif substring (option vendor-class-identifier, 0, 13) = > "Etherboot-5.4" > { > filename "/ltsp/i386/elf.ltsp"; > } > # Etherboot NBI (older clients) > elsif substring (option vendor-class-identifier, 0, 9) = > "Etherboot" > { > filename "/ltsp/i386/wraplinux-nbi.ltsp"; > } > # PXE > elsif substring (option vendor-class-identifier, 0, 9) = > "PXEClient" > { > # NOTE: kernels are specified in > /tftpboot/ltsp/i386/pxelinux.cfg/ > filename "/ltsp/i386/pxelinux.0"; > } > # if all else fails (likely BOOTP), default to an NBI image > else > { > filename "/ltsp/i386/wraplinux-nbi.ltsp"; > } > > if substring (option vendor-class-identifier, 20, 3) = "ppc" { > option root-path "192.168.5.3:/opt/ltsp/ppc"; > } else { > option root-path "192.168.5.3:/opt/ltsp/i386"; > #option root-path "nbd:192.168.5.3:2000:squashfs:ro"; > } > } > } > > ## example configurations for specifying specific kernels to specific > clients > #group { > # use-host-decl-names on; > # option log-servers 192.168.5.3; > # > # host ws001 { > # hardware ethernet 00:E0:06:E8:00:84; > # fixed-address 172.31.100.1; > # filename "/ltsp/vmlinuz.ltsp"; > # option option-128 e4:45:74:68:00:00; > # option option-129 "NIC=3c509"; > # } > # host ws002 { > # hardware ethernet 00:D0:09:30:6A:1C; > # fixed-address 172.31.100.2; > # filename "/ltsp/vmlinuz.ltsp"; > # option option-128 e4:45:74:68:00:00; > # option option-129 "NIC=ne"; > # } > # host ws003 { > # hardware ethernet 00:D0:09:30:28:B2; > # fixed-address 172.31.100.3; > # # kernels are specified in /tftpboot/ltsp/i386/pxelinux.cfg/ > # filename "/ltsp/i386/pxelinux.0"; > # } > # > # Apple Specific Settings > # host ws007 { > # hardware ethernet 00:30:65:69:23:60; > # fixed-address 172.31.100.4; > # option root-path "192.168.5.3:/opt/ltsp/ppc"; > # filename "yaboot"; > # option vendor-class-identifier "AAPLBSDPC"; > # } > #} > > > > I hope that this is useful! > > Ron > > > Ron McDaniel > rmcdaniel at indata.us > Conecuh County Schools > Technology Coordinator > > > -------- Original Message -------- > Subject: Re: single interface setup > From: Luis Montes > Date: Mon, November 03, 2008 11:10 pm > To: Development discussion of K12Linux > > I though I did just that and it didn't work for me. > > Could you post your exact configuration changes? > > Thanks, > Luis > > > rmcdaniel at indata.us wrote: > > I finally got the setup working with just a single interface. All that > > I did was configure eth0 as my main interface and change the dhcpd.conf > > file with the correct subnet information. I just disabled eth1 and left > > the bridge interface as it was setup from the initial install. I will > > probably get rid of the bridge interface just to remove any possible > > problems. Everything is working GREAT! Has anyone tried using some of > > the scripts that were included with K12LTSP, ie. "Add/Remove Icons to > > Desktops"? > > > > > > Thanks, > > > > Ron McDaniel > > rmcdaniel at indata.us > > > > > > > > > > > > _______________________________________________ > > K12Linux-devel-list mailing list > > K12Linux-devel-list at redhat.com > > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list From monteslu at cox.net Wed Nov 5 20:12:04 2008 From: monteslu at cox.net (Luis Montes) Date: Wed, 05 Nov 2008 13:12:04 -0700 Subject: very slow login times In-Reply-To: <4901FC16.4010605@cox.net> References: <4901FC16.4010605@cox.net> Message-ID: <4911FE14.8020804@cox.net> This has really become a problem. Waiting 3 to 4 minutes from login is too long. Any one have any ideas what script is eating up the thin client's cpu? I could really use some help with this. Thanks, Luis Luis Montes wrote: > Boot up seems normal, but time from login to having gnome up is about > 4 and a half minutes. > > If I jump over to another tty on the terminal it looks like sh is > eating up most of the cpu and ram for that duration, though the > resources aren't completely tapped. > > The thin clients are disklessworkstations.com t150s. (32bit 500Mhz > via, 128MB ram). Top shows that during the login period I'm using > about 80MB of ram. > > The server is Fedora 9 64bit. > > Nothing particularly interesting either /var/log/messages on the > server or the client that I can tell, but I'm not sure what to look for. > > Any idea which script this is that could be taking up all the cpu time? > > Thanks, > > Luis > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > From rmcdaniel at indata.us Wed Nov 5 21:19:38 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Wed, 05 Nov 2008 14:19:38 -0700 Subject: very slow login times Message-ID: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> This is interesting. I am using a HP Proliant server with Dual Quad Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing very slow bootup and login times. I am suspecting that it has something to do with my network. When I run the resource monitor on the server, nothing looks like it is getting close to maxing out. I am running the 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. Tomorrow I am going to look at my network and see if I have any problems. Ron McDaniel rmcdaniel at indata.us -------- Original Message -------- Subject: Re: very slow login times From: Luis Montes Date: Wed, November 05, 2008 2:12 pm To: Development discussion of K12Linux This has really become a problem. Waiting 3 to 4 minutes from login is too long. Any one have any ideas what script is eating up the thin client's cpu? I could really use some help with this. Thanks, Luis Luis Montes wrote: > Boot up seems normal, but time from login to having gnome up is about > 4 and a half minutes. > > If I jump over to another tty on the terminal it looks like sh is > eating up most of the cpu and ram for that duration, though the > resources aren't completely tapped. > > The thin clients are disklessworkstations.com t150s. (32bit 500Mhz > via, 128MB ram). Top shows that during the login period I'm using > about 80MB of ram. > > The server is Fedora 9 64bit. > > Nothing particularly interesting either /var/log/messages on the > server or the client that I can tell, but I'm not sure what to look for. > > Any idea which script this is that could be taking up all the cpu time? > > Thanks, > > Luis > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From monteslu at cox.net Wed Nov 5 21:48:02 2008 From: monteslu at cox.net (Luis Montes) Date: Wed, 05 Nov 2008 14:48:02 -0700 Subject: very slow login times In-Reply-To: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> References: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> Message-ID: <49121492.9000909@cox.net> My guess right now is that its neither the network or the server. After you type in a user/password on the login screen, switch over to a different tty like ctl-alt-f8 I have this in my lts.conf: [default] SCREEN_07=ldm LDM_DIRECTX=True SCREEN_08=shell run top during the slow login time and let me know if you also have a pegged thin client. This slow login time has to be a bug. I've been using ltsp for years and have never seen this before. Luis rmcdaniel at indata.us wrote: > This is interesting. I am using a HP Proliant server with Dual Quad > Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing > very slow bootup and login times. I am suspecting that it has something > to do with my network. When I run the resource monitor on the server, > nothing looks like it is getting close to maxing out. I am running the > 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. > Tomorrow I am going to look at my network and see if I have any > problems. > > > Ron McDaniel > rmcdaniel at indata.us > > -------- Original Message -------- > Subject: Re: very slow login times > From: Luis Montes > Date: Wed, November 05, 2008 2:12 pm > To: Development discussion of K12Linux > > This has really become a problem. Waiting 3 to 4 minutes from login is > too long. > > Any one have any ideas what script is eating up the thin client's cpu? > > I could really use some help with this. > > Thanks, > > Luis > > > > > > Luis Montes wrote: > >> Boot up seems normal, but time from login to having gnome up is about >> 4 and a half minutes. >> >> If I jump over to another tty on the terminal it looks like sh is >> eating up most of the cpu and ram for that duration, though the >> resources aren't completely tapped. >> >> The thin clients are disklessworkstations.com t150s. (32bit 500Mhz >> via, 128MB ram). Top shows that during the login period I'm using >> about 80MB of ram. >> >> The server is Fedora 9 64bit. >> >> Nothing particularly interesting either /var/log/messages on the >> server or the client that I can tell, but I'm not sure what to look for. >> >> Any idea which script this is that could be taking up all the cpu time? >> >> Thanks, >> >> Luis >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > From dyoung at mesd.k12.or.us Wed Nov 5 21:54:50 2008 From: dyoung at mesd.k12.or.us (Dan Young) Date: Wed, 5 Nov 2008 13:54:50 -0800 Subject: very slow login times In-Reply-To: <4901FC16.4010605@cox.net> References: <4901FC16.4010605@cox.net> Message-ID: <994441ae0811051354ja2108b4pdb0a46b84aa68651@mail.gmail.com> On Fri, Oct 24, 2008 at 8:47 AM, Luis Montes wrote: > Boot up seems normal, but time from login to having gnome up is about 4 and > a half minutes. > > If I jump over to another tty on the terminal it looks like sh is eating up > most of the cpu and ram for that duration, though the resources aren't > completely tapped. When you see in top that sh has the CPU pegged, can you hit "c" to display the full command? -- Dan Young Multnomah ESD - Technology Services 503-257-1562 From monteslu at cox.net Wed Nov 5 21:59:34 2008 From: monteslu at cox.net (Luis Montes) Date: Wed, 05 Nov 2008 14:59:34 -0700 Subject: very slow login times In-Reply-To: <49121492.9000909@cox.net> References: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> <49121492.9000909@cox.net> Message-ID: <49121746.6040009@cox.net> copy & paste got wierd on me there. it should have been: [default] SCREEN_07=ldm LDM_DIRECTX=True SCREEN_08=shell (each on its own line of course, in case this one doesn't work either) Luis Montes wrote: > My guess right now is that its neither the network or the server. > After you type in a user/password on the login screen, switch over to > a different tty like ctl-alt-f8 > > I have this in my lts.conf: > [default] > SCREEN_07=ldm LDM_DIRECTX=True > SCREEN_08=shell > > run top during the slow login time and let me know if you also have a > pegged thin client. > > This slow login time has to be a bug. I've been using ltsp for years > and have never seen this before. > > Luis > > > rmcdaniel at indata.us wrote: >> This is interesting. I am using a HP Proliant server with Dual Quad >> Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing >> very slow bootup and login times. I am suspecting that it has something >> to do with my network. When I run the resource monitor on the server, >> nothing looks like it is getting close to maxing out. I am running the >> 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. >> Tomorrow I am going to look at my network and see if I have any >> problems. >> >> >> Ron McDaniel >> rmcdaniel at indata.us >> >> -------- Original Message -------- >> Subject: Re: very slow login times >> From: Luis Montes >> Date: Wed, November 05, 2008 2:12 pm >> To: Development discussion of K12Linux >> >> This has really become a problem. Waiting 3 to 4 minutes from login >> is too long. >> >> Any one have any ideas what script is eating up the thin client's cpu? >> >> I could really use some help with this. >> >> Thanks, >> >> Luis >> >> >> >> >> >> Luis Montes wrote: >> >>> Boot up seems normal, but time from login to having gnome up is >>> about 4 and a half minutes. >>> >>> If I jump over to another tty on the terminal it looks like sh is >>> eating up most of the cpu and ram for that duration, though the >>> resources aren't completely tapped. >>> >>> The thin clients are disklessworkstations.com t150s. (32bit 500Mhz >>> via, 128MB ram). Top shows that during the login period I'm using >>> about 80MB of ram. >>> >>> The server is Fedora 9 64bit. >>> >>> Nothing particularly interesting either /var/log/messages on the >>> server or the client that I can tell, but I'm not sure what to look >>> for. >>> >>> Any idea which script this is that could be taking up all the cpu time? >>> >>> Thanks, >>> >>> Luis >>> >>> _______________________________________________ >>> K12Linux-devel-list mailing list >>> K12Linux-devel-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >>> >>> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> >> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> > > From rmcdaniel at indata.us Wed Nov 5 22:36:30 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Wed, 05 Nov 2008 15:36:30 -0700 Subject: very slow login times Message-ID: <20081105153630.d7061e97b78b017ac15395d64f2ce134.78f1869ccb.wbe@email.secureserver.net> I will give that a shout tomorrow when I am out at the school. Thanks for the information! Ron -------- Original Message -------- Subject: Re: very slow login times From: Luis Montes Date: Wed, November 05, 2008 3:59 pm To: Development discussion of K12Linux copy & paste got wierd on me there. it should have been: [default] SCREEN_07=ldm LDM_DIRECTX=True SCREEN_08=shell (each on its own line of course, in case this one doesn't work either) Luis Montes wrote: > My guess right now is that its neither the network or the server. > After you type in a user/password on the login screen, switch over to > a different tty like ctl-alt-f8 > > I have this in my lts.conf: > [default] > SCREEN_07=ldm LDM_DIRECTX=True > SCREEN_08=shell > > run top during the slow login time and let me know if you also have a > pegged thin client. > > This slow login time has to be a bug. I've been using ltsp for years > and have never seen this before. > > Luis > > > rmcdaniel at indata.us wrote: >> This is interesting. I am using a HP Proliant server with Dual Quad >> Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing >> very slow bootup and login times. I am suspecting that it has something >> to do with my network. When I run the resource monitor on the server, >> nothing looks like it is getting close to maxing out. I am running the >> 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. >> Tomorrow I am going to look at my network and see if I have any >> problems. >> >> >> Ron McDaniel >> rmcdaniel at indata.us >> >> -------- Original Message -------- >> Subject: Re: very slow login times >> From: Luis Montes >> Date: Wed, November 05, 2008 2:12 pm >> To: Development discussion of K12Linux >> >> This has really become a problem. Waiting 3 to 4 minutes from login >> is too long. >> >> Any one have any ideas what script is eating up the thin client's cpu? >> >> I could really use some help with this. >> >> Thanks, >> >> Luis >> >> >> >> >> >> Luis Montes wrote: >> >>> Boot up seems normal, but time from login to having gnome up is >>> about 4 and a half minutes. >>> >>> If I jump over to another tty on the terminal it looks like sh is >>> eating up most of the cpu and ram for that duration, though the >>> resources aren't completely tapped. >>> >>> The thin clients are disklessworkstations.com t150s. (32bit 500Mhz >>> via, 128MB ram). Top shows that during the login period I'm using >>> about 80MB of ram. >>> >>> The server is Fedora 9 64bit. >>> >>> Nothing particularly interesting either /var/log/messages on the >>> server or the client that I can tell, but I'm not sure what to look >>> for. >>> >>> Any idea which script this is that could be taking up all the cpu time? >>> >>> Thanks, >>> >>> Luis >>> >>> _______________________________________________ >>> K12Linux-devel-list mailing list >>> K12Linux-devel-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >>> >>> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> >> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> > > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From wtogami at redhat.com Wed Nov 5 23:19:54 2008 From: wtogami at redhat.com (Warren Togami) Date: Wed, 05 Nov 2008 18:19:54 -0500 Subject: very slow login times In-Reply-To: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> References: <20081105141938.d7061e97b78b017ac15395d64f2ce134.ea3b7e5d15.wbe@email.secureserver.net> Message-ID: <49122A1A.6010601@redhat.com> rmcdaniel at indata.us wrote: > This is interesting. I am using a HP Proliant server with Dual Quad > Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing > very slow bootup and login times. I am suspecting that it has something > to do with my network. When I run the resource monitor on the server, > nothing looks like it is getting close to maxing out. I am running the > 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. > Tomorrow I am going to look at my network and see if I have any > problems. > > > Ron McDaniel > rmcdaniel at indata.us > Have you tried to plug a thin client directly into the 1GB ethernet interface on the server instead of going through switches? I've often seen bad switches or cables cause slow bootup and slow login. Warren From rmcdaniel at indata.us Thu Nov 6 03:31:09 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Wed, 05 Nov 2008 20:31:09 -0700 Subject: very slow login times Message-ID: <20081105203109.d7061e97b78b017ac15395d64f2ce134.855d8cc738.wbe@email.secureserver.net> I ran a new cable today from the lab of 30 thin clients to the server. I updated the lab switch which is a HP Procurve 2650. I haven't tried the thin clients after I updated the switch. I am going to give it a shot tomorrow. Thanks for the recommendations. Ron -------- Original Message -------- Subject: Re: very slow login times From: Warren Togami Date: Wed, November 05, 2008 5:19 pm To: Development discussion of K12Linux rmcdaniel at indata.us wrote: > This is interesting. I am using a HP Proliant server with Dual Quad > Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing > very slow bootup and login times. I am suspecting that it has something > to do with my network. When I run the resource monitor on the server, > nothing looks like it is getting close to maxing out. I am running the > 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. > Tomorrow I am going to look at my network and see if I have any > problems. > > > Ron McDaniel > rmcdaniel at indata.us > Have you tried to plug a thin client directly into the 1GB ethernet interface on the server instead of going through switches? I've often seen bad switches or cables cause slow bootup and slow login. Warren _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From rmcdaniel at indata.us Thu Nov 6 17:19:36 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Thu, 06 Nov 2008 10:19:36 -0700 Subject: very slow login times Message-ID: <20081106101936.d7061e97b78b017ac15395d64f2ce134.ff27abd7d0.wbe@email.secureserver.net> Here's the scoop this morning. I have run all updates on the server and rebuilt the client /opt/ltsp/i386 to make sure that I had all current updates applied. I am still seeing excessive bandwidth used on the switch with just 1 thin client attached and booting off of the server. To eliminate any network suspicion, I have a regular pc with XP running off of the same switch in the lab. When I access a youtube video with the thin client, I max my switch port at 100% utilization with 100% unicast packets. When I access the same youtube video with the pc, the switch utilization is very low 2%. It seems like some great inefficiencies in the networking protocols on the thin clients. I can load K12LTSP centos 5 and compare the results. The only problem with that is that there isn't any sound in flash... Ron McDaniel rmcdaniel at indata.us -------- Original Message -------- Subject: Re: very slow login times From: Warren Togami Date: Wed, November 05, 2008 5:19 pm To: Development discussion of K12Linux rmcdaniel at indata.us wrote: > This is interesting. I am using a HP Proliant server with Dual Quad > Core cpu's, 4 GB memory, and about 210GB of storage. I am also seeing > very slow bootup and login times. I am suspecting that it has something > to do with my network. When I run the resource monitor on the server, > nothing looks like it is getting close to maxing out. I am running the > 32 bit version of the OS and my clients are HP T5530 with 128 MB of mem. > Tomorrow I am going to look at my network and see if I have any > problems. > > > Ron McDaniel > rmcdaniel at indata.us > Have you tried to plug a thin client directly into the 1GB ethernet interface on the server instead of going through switches? I've often seen bad switches or cables cause slow bootup and slow login. Warren _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From robark at gmail.com Thu Nov 6 18:13:27 2008 From: robark at gmail.com (Robert Arkiletian) Date: Thu, 6 Nov 2008 10:13:27 -0800 Subject: very slow login times In-Reply-To: <20081106101936.d7061e97b78b017ac15395d64f2ce134.ff27abd7d0.wbe@email.secureserver.net> References: <20081106101936.d7061e97b78b017ac15395d64f2ce134.ff27abd7d0.wbe@email.secureserver.net> Message-ID: On 11/6/08, rmcdaniel at indata.us wrote: > Here's the scoop this morning. I have run all updates on the server and > rebuilt the client /opt/ltsp/i386 to make sure that I had all current > updates applied. I am still seeing excessive bandwidth used on the > switch with just 1 thin client attached and booting off of the server. > To eliminate any network suspicion, I have a regular pc with XP running > off of the same switch in the lab. When I access a youtube video with > the thin client, I max my switch port at 100% utilization with 100% > unicast packets. When I access the same youtube video with the pc, the > switch utilization is very low 2%. It seems like some great > inefficiencies in the networking protocols on the thin clients. I can Ron this is normal. Youtube is not a good test to see if your setup is over utilizing network bandwidth. Youtube will saturate a 100Mbps connection. I have tried running a youtube video as a local app and the network usage drops dramatically. Here's why. Flash is a highly compressed video format. The SERVER decompresses it and THEN pushes the decompressed stream to the X server on the client. This is inefficient for 2 reasons. 1) the server is taking the cpu load to do the decompression 2) your lan is being saturated with decompressed video image data Try running firefox as a local app and watch the network utilization on gnome-system-monitor This is a big reason why local apps is so important. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ From rmcdaniel at indata.us Thu Nov 6 19:04:37 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Thu, 06 Nov 2008 12:04:37 -0700 Subject: very slow login times Message-ID: <20081106120436.d7061e97b78b017ac15395d64f2ce134.fbd494fc00.wbe@email.secureserver.net> Robert, Thanks for that helpful info. I remember you addressing this back in the K12LTSP days. Do you have a link to assist in setting up local apps? Are there any disadvantages to setting up local apps on clients? Thanks, Ron -------- Original Message -------- Subject: Re: very slow login times From: "Robert Arkiletian" Date: Thu, November 06, 2008 12:13 pm To: "Development discussion of K12Linux" On 11/6/08, rmcdaniel at indata.us wrote: > Here's the scoop this morning. I have run all updates on the server and > rebuilt the client /opt/ltsp/i386 to make sure that I had all current > updates applied. I am still seeing excessive bandwidth used on the > switch with just 1 thin client attached and booting off of the server. > To eliminate any network suspicion, I have a regular pc with XP running > off of the same switch in the lab. When I access a youtube video with > the thin client, I max my switch port at 100% utilization with 100% > unicast packets. When I access the same youtube video with the pc, the > switch utilization is very low 2%. It seems like some great > inefficiencies in the networking protocols on the thin clients. I can Ron this is normal. Youtube is not a good test to see if your setup is over utilizing network bandwidth. Youtube will saturate a 100Mbps connection. I have tried running a youtube video as a local app and the network usage drops dramatically. Here's why. Flash is a highly compressed video format. The SERVER decompresses it and THEN pushes the decompressed stream to the X server on the client. This is inefficient for 2 reasons. 1) the server is taking the cpu load to do the decompression 2) your lan is being saturated with decompressed video image data Try running firefox as a local app and watch the network utilization on gnome-system-monitor This is a big reason why local apps is so important. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From rmcdaniel at indata.us Thu Nov 6 19:13:05 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Thu, 06 Nov 2008 12:13:05 -0700 Subject: very slow login times Message-ID: <20081106121305.d7061e97b78b017ac15395d64f2ce134.13ab84a74e.wbe@email.secureserver.net> I found one of your post addressing this! Thanks. http://www.mail-archive.com/ltsp-discuss at lists.sourceforge.net/msg35351.html -------- Original Message -------- Subject: Re: very slow login times From: "Robert Arkiletian" Date: Thu, November 06, 2008 12:13 pm To: "Development discussion of K12Linux" On 11/6/08, rmcdaniel at indata.us wrote: > Here's the scoop this morning. I have run all updates on the server and > rebuilt the client /opt/ltsp/i386 to make sure that I had all current > updates applied. I am still seeing excessive bandwidth used on the > switch with just 1 thin client attached and booting off of the server. > To eliminate any network suspicion, I have a regular pc with XP running > off of the same switch in the lab. When I access a youtube video with > the thin client, I max my switch port at 100% utilization with 100% > unicast packets. When I access the same youtube video with the pc, the > switch utilization is very low 2%. It seems like some great > inefficiencies in the networking protocols on the thin clients. I can Ron this is normal. Youtube is not a good test to see if your setup is over utilizing network bandwidth. Youtube will saturate a 100Mbps connection. I have tried running a youtube video as a local app and the network usage drops dramatically. Here's why. Flash is a highly compressed video format. The SERVER decompresses it and THEN pushes the decompressed stream to the X server on the client. This is inefficient for 2 reasons. 1) the server is taking the cpu load to do the decompression 2) your lan is being saturated with decompressed video image data Try running firefox as a local app and watch the network utilization on gnome-system-monitor This is a big reason why local apps is so important. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From robark at gmail.com Fri Nov 7 04:53:41 2008 From: robark at gmail.com (Robert Arkiletian) Date: Thu, 6 Nov 2008 20:53:41 -0800 Subject: Idea to solve Fl_TeacherTool security issue Message-ID: Is it possible to use the local-app plumbing to start x11vnc on the client while passing it a rfbauth password file from a path OUTSIDE the chroot? Can a process which runs as a local app see outside the chroot? Or is this impossible by definition? This would solve the issue for k12linux but not k12ltsp since k12ltsp doesn't have local-app support. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ From monteslu at cox.net Fri Nov 7 15:15:18 2008 From: monteslu at cox.net (Luis Montes) Date: Fri, 07 Nov 2008 08:15:18 -0700 Subject: very slow login times In-Reply-To: <994441ae0811051354ja2108b4pdb0a46b84aa68651@mail.gmail.com> References: <4901FC16.4010605@cox.net> <994441ae0811051354ja2108b4pdb0a46b84aa68651@mail.gmail.com> Message-ID: <49145B86.70404@cox.net> With only 80 chars I'm not able to see the full command. All I'm getting is /bin/sh /usr/share But I did get a little more with ps -Af Looks like the culprit is /bin/sh /usr/share/ldm/ldm-script What exactly in there is killing the CPU for 3 or more minutes I'm not sure of. Luis Dan Young wrote: > On Fri, Oct 24, 2008 at 8:47 AM, Luis Montes wrote: > >> Boot up seems normal, but time from login to having gnome up is about 4 and >> a half minutes. >> >> If I jump over to another tty on the terminal it looks like sh is eating up >> most of the cpu and ram for that duration, though the resources aren't >> completely tapped. >> > > When you see in top that sh has the CPU pegged, can you hit "c" to > display the full command? > > -- > Dan Young > Multnomah ESD - Technology Services > 503-257-1562 > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > From rmcdaniel at indata.us Fri Nov 7 17:44:10 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 10:44:10 -0700 Subject: localapps Message-ID: <20081107104410.d7061e97b78b017ac15395d64f2ce134.b59c5b08c4.wbe@email.secureserver.net> Robert, I have firefox running as a local app and it is working great. I am trying to setup an icon that the kids can click on to take them to a specific site, www.kclogin.com. This is a flash based program that we have been testing. I can't simply make a shotcut to the site like I would typically do with Firefox (drag and drop the address from the toolbar to the desktop). I have tried several switches from the command line to start localapp firefox with a specific URL. The switches don't seem to work. The exact command that I was trying is /usr/bin/ltsp-localapps /usr/bin/firefox -url http://www.kclogin.com. Do you know of any way that this can be accomplished. This is for the young kids, the older kids can just simply type the site in the address bar. Thanks, Ron -------- Original Message -------- Subject: RE: very slow login times From: Date: Thu, November 06, 2008 1:13 pm To: "Development discussion of K12Linux" I found one of your post addressing this! Thanks. http://www.mail-archive.com/ltsp-discuss at lists.sourceforge.net/msg35351.html -------- Original Message -------- Subject: Re: very slow login times From: "Robert Arkiletian" Date: Thu, November 06, 2008 12:13 pm To: "Development discussion of K12Linux" On 11/6/08, rmcdaniel at indata.us wrote: > Here's the scoop this morning. I have run all updates on the server and > rebuilt the client /opt/ltsp/i386 to make sure that I had all current > updates applied. I am still seeing excessive bandwidth used on the > switch with just 1 thin client attached and booting off of the server. > To eliminate any network suspicion, I have a regular pc with XP running > off of the same switch in the lab. When I access a youtube video with > the thin client, I max my switch port at 100% utilization with 100% > unicast packets. When I access the same youtube video with the pc, the > switch utilization is very low 2%. It seems like some great > inefficiencies in the networking protocols on the thin clients. I can Ron this is normal. Youtube is not a good test to see if your setup is over utilizing network bandwidth. Youtube will saturate a 100Mbps connection. I have tried running a youtube video as a local app and the network usage drops dramatically. Here's why. Flash is a highly compressed video format. The SERVER decompresses it and THEN pushes the decompressed stream to the X server on the client. This is inefficient for 2 reasons. 1) the server is taking the cpu load to do the decompression 2) your lan is being saturated with decompressed video image data Try running firefox as a local app and watch the network utilization on gnome-system-monitor This is a big reason why local apps is so important. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From monteslu at cox.net Fri Nov 7 17:54:56 2008 From: monteslu at cox.net (Luis Montes) Date: Fri, 07 Nov 2008 10:54:56 -0700 Subject: localapps In-Reply-To: <20081107104410.d7061e97b78b017ac15395d64f2ce134.b59c5b08c4.wbe@email.secureserver.net> References: <20081107104410.d7061e97b78b017ac15395d64f2ce134.b59c5b08c4.wbe@email.secureserver.net> Message-ID: <491480F0.1090202@cox.net> Just a stab in the dark since I don't have localapps working yet, try quoting the argument to ltsp-localapps /usr/bin/ltsp-localapps "/usr/bin/firefox -url http://www.kclogin.com" also, i dont think you need -url if the only parameter to firefox is the url. Luis rmcdaniel at indata.us wrote: > Robert, > > I have firefox running as a local app and it is working great. I am > trying to setup an icon that the kids can click on to take them to a > specific site, www.kclogin.com. This is a flash based program that we > have been testing. I can't simply make a shotcut to the site like I > would typically do with Firefox (drag and drop the address from the > toolbar to the desktop). I have tried several switches from the command > line to start localapp firefox with a specific URL. The switches don't > seem to work. The exact command that I was trying is > /usr/bin/ltsp-localapps /usr/bin/firefox -url http://www.kclogin.com. > Do you know of any way that this can be accomplished. This is for the > young kids, the older kids can just simply type the site in the address > bar. > > Thanks, > > Ron > > > -------- Original Message -------- > Subject: RE: very slow login times > From: > Date: Thu, November 06, 2008 1:13 pm > To: "Development discussion of K12Linux" > > > I found one of your post addressing this! Thanks. > > http://www.mail-archive.com/ltsp-discuss at lists.sourceforge.net/msg35351.html > > -------- Original Message -------- > Subject: Re: very slow login times > From: "Robert Arkiletian" > Date: Thu, November 06, 2008 12:13 pm > To: "Development discussion of K12Linux" > > > On 11/6/08, rmcdaniel at indata.us wrote: > >> Here's the scoop this morning. I have run all updates on the server and >> rebuilt the client /opt/ltsp/i386 to make sure that I had all current >> updates applied. I am still seeing excessive bandwidth used on the >> switch with just 1 thin client attached and booting off of the server. >> To eliminate any network suspicion, I have a regular pc with XP running >> off of the same switch in the lab. When I access a youtube video with >> the thin client, I max my switch port at 100% utilization with 100% >> unicast packets. When I access the same youtube video with the pc, the >> switch utilization is very low 2%. It seems like some great >> inefficiencies in the networking protocols on the thin clients. I can >> > > Ron this is normal. Youtube is not a good test to see if your setup is > over utilizing network bandwidth. Youtube will saturate a 100Mbps > connection. I have tried running a youtube video as a local app and > the network usage drops dramatically. Here's why. > > Flash is a highly compressed video format. The SERVER decompresses it > and THEN pushes the decompressed stream to the X server on the client. > This is inefficient for 2 reasons. > > 1) the server is taking the cpu load to do the decompression > 2) your lan is being saturated with decompressed video image data > > Try running firefox as a local app and watch the network utilization > on gnome-system-monitor > > This is a big reason why local apps is so important. > > From rmcdaniel at indata.us Fri Nov 7 18:09:07 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 11:09:07 -0700 Subject: localapps Message-ID: <20081107110907.d7061e97b78b017ac15395d64f2ce134.08d72dc778.wbe@email.secureserver.net> nice stab! Thanks, Ron -------- Original Message -------- Subject: Re: localapps From: Luis Montes Date: Fri, November 07, 2008 11:54 am To: Development discussion of K12Linux Just a stab in the dark since I don't have localapps working yet, try quoting the argument to ltsp-localapps /usr/bin/ltsp-localapps "/usr/bin/firefox -url http://www.kclogin.com" also, i dont think you need -url if the only parameter to firefox is the url. Luis rmcdaniel at indata.us wrote: > Robert, > > I have firefox running as a local app and it is working great. I am > trying to setup an icon that the kids can click on to take them to a > specific site, www.kclogin.com. This is a flash based program that we > have been testing. I can't simply make a shotcut to the site like I > would typically do with Firefox (drag and drop the address from the > toolbar to the desktop). I have tried several switches from the command > line to start localapp firefox with a specific URL. The switches don't > seem to work. The exact command that I was trying is > /usr/bin/ltsp-localapps /usr/bin/firefox -url http://www.kclogin.com. > Do you know of any way that this can be accomplished. This is for the > young kids, the older kids can just simply type the site in the address > bar. > > Thanks, > > Ron > > > -------- Original Message -------- > Subject: RE: very slow login times > From: > Date: Thu, November 06, 2008 1:13 pm > To: "Development discussion of K12Linux" > > > I found one of your post addressing this! Thanks. > > http://www.mail-archive.com/ltsp-discuss at lists.sourceforge.net/msg35351.html > > -------- Original Message -------- > Subject: Re: very slow login times > From: "Robert Arkiletian" > Date: Thu, November 06, 2008 12:13 pm > To: "Development discussion of K12Linux" > > > On 11/6/08, rmcdaniel at indata.us wrote: > >> Here's the scoop this morning. I have run all updates on the server and >> rebuilt the client /opt/ltsp/i386 to make sure that I had all current >> updates applied. I am still seeing excessive bandwidth used on the >> switch with just 1 thin client attached and booting off of the server. >> To eliminate any network suspicion, I have a regular pc with XP running >> off of the same switch in the lab. When I access a youtube video with >> the thin client, I max my switch port at 100% utilization with 100% >> unicast packets. When I access the same youtube video with the pc, the >> switch utilization is very low 2%. It seems like some great >> inefficiencies in the networking protocols on the thin clients. I can >> > > Ron this is normal. Youtube is not a good test to see if your setup is > over utilizing network bandwidth. Youtube will saturate a 100Mbps > connection. I have tried running a youtube video as a local app and > the network usage drops dramatically. Here's why. > > Flash is a highly compressed video format. The SERVER decompresses it > and THEN pushes the decompressed stream to the X server on the client. > This is inefficient for 2 reasons. > > 1) the server is taking the cpu load to do the decompression > 2) your lan is being saturated with decompressed video image data > > Try running firefox as a local app and watch the network utilization > on gnome-system-monitor > > This is a big reason why local apps is so important. > > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From rmcdaniel at indata.us Fri Nov 7 18:45:18 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 11:45:18 -0700 Subject: localapps Message-ID: <20081107114518.d7061e97b78b017ac15395d64f2ce134.b9bf3434db.wbe@email.secureserver.net> Localapps seem to run slower than a similar app on the server. My thin clinets are HP T5530 with 256MB. Is there a setting that can be changed to improve the performance? For example when I start Firefox as an localapp, it takes about 5 seconds to load. Running Firefox from the server takes about 2 seconds. It also seems a little slugish when the localapp is running. Ron -------- Original Message -------- Subject: RE: localapps From: Date: Fri, November 07, 2008 12:09 pm To: "Development discussion of K12Linux" nice stab! Thanks, Ron -------- Original Message -------- Subject: Re: localapps From: Luis Montes Date: Fri, November 07, 2008 11:54 am To: Development discussion of K12Linux Just a stab in the dark since I don't have localapps working yet, try quoting the argument to ltsp-localapps /usr/bin/ltsp-localapps "/usr/bin/firefox -url http://www.kclogin.com" also, i dont think you need -url if the only parameter to firefox is the url. Luis rmcdaniel at indata.us wrote: > Robert, > > I have firefox running as a local app and it is working great. I am > trying to setup an icon that the kids can click on to take them to a > specific site, www.kclogin.com. This is a flash based program that we > have been testing. I can't simply make a shotcut to the site like I > would typically do with Firefox (drag and drop the address from the > toolbar to the desktop). I have tried several switches from the command > line to start localapp firefox with a specific URL. The switches don't > seem to work. The exact command that I was trying is > /usr/bin/ltsp-localapps /usr/bin/firefox -url http://www.kclogin.com. > Do you know of any way that this can be accomplished. This is for the > young kids, the older kids can just simply type the site in the address > bar. > > Thanks, > > Ron > > > -------- Original Message -------- > Subject: RE: very slow login times > From: > Date: Thu, November 06, 2008 1:13 pm > To: "Development discussion of K12Linux" > > > I found one of your post addressing this! Thanks. > > http://www.mail-archive.com/ltsp-discuss at lists.sourceforge.net/msg35351.html > > -------- Original Message -------- > Subject: Re: very slow login times > From: "Robert Arkiletian" > Date: Thu, November 06, 2008 12:13 pm > To: "Development discussion of K12Linux" > > > On 11/6/08, rmcdaniel at indata.us wrote: > >> Here's the scoop this morning. I have run all updates on the server and >> rebuilt the client /opt/ltsp/i386 to make sure that I had all current >> updates applied. I am still seeing excessive bandwidth used on the >> switch with just 1 thin client attached and booting off of the server. >> To eliminate any network suspicion, I have a regular pc with XP running >> off of the same switch in the lab. When I access a youtube video with >> the thin client, I max my switch port at 100% utilization with 100% >> unicast packets. When I access the same youtube video with the pc, the >> switch utilization is very low 2%. It seems like some great >> inefficiencies in the networking protocols on the thin clients. I can >> > > Ron this is normal. Youtube is not a good test to see if your setup is > over utilizing network bandwidth. Youtube will saturate a 100Mbps > connection. I have tried running a youtube video as a local app and > the network usage drops dramatically. Here's why. > > Flash is a highly compressed video format. The SERVER decompresses it > and THEN pushes the decompressed stream to the X server on the client. > This is inefficient for 2 reasons. > > 1) the server is taking the cpu load to do the decompression > 2) your lan is being saturated with decompressed video image data > > Try running firefox as a local app and watch the network utilization > on gnome-system-monitor > > This is a big reason why local apps is so important. > > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From rmcdaniel at indata.us Fri Nov 7 19:17:33 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 12:17:33 -0700 Subject: K12LTSP centos 5 vs. K12Linux Message-ID: <20081107121733.d7061e97b78b017ac15395d64f2ce134.8add042619.wbe@email.secureserver.net> The more that I try and get a lab running with K12Linux, the more frustrated I become. Why does it seem that K12LTSP Centos5 greatly outperforms K12Linux. Is the difference due to LTSP 4.2 vs. LTSP 5? I have also given Edubuntu 8.1 a try and end up experiencing all of the crazy DHCP problems that seem to plague certain installations. Is this really progression??? Sorry for the rant, but I have been working on a single lab with 30 clients for two weeks. The K12LTSP install would have been done in about 2 hours... Ron McDaniel rmcdaniel at indata.us From peter at scheie.homedns.org Fri Nov 7 19:42:11 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Fri, 07 Nov 2008 13:42:11 -0600 Subject: K12LTSP centos 5 vs. K12Linux In-Reply-To: <20081107121733.d7061e97b78b017ac15395d64f2ce134.8add042619.wbe@email.secureserver.net> References: <20081107121733.d7061e97b78b017ac15395d64f2ce134.8add042619.wbe@email.secureserver.net> Message-ID: <49149A13.5000501@scheie.homedns.org> rmcdaniel at indata.us wrote: > The more that I try and get a lab running with K12Linux, the more > frustrated I become. Why does it seem that K12LTSP Centos5 greatly > outperforms K12Linux. Is the difference due to LTSP 4.2 vs. LTSP 5? I > have also given Edubuntu 8.1 a try and end up experiencing all of the > crazy DHCP problems that seem to plague certain installations. Is this > really progression??? Sorry for the rant, but I have been working on a > single lab with 30 clients for two weeks. The K12LTSP install would > have been done in about 2 hours... > > > Ron McDaniel > rmcdaniel at indata.us > > Performance-wise, yes, the demands of LTSP 5 on the client are greater than those of LTSP4.2. If your clients are slower than, say 400mhz and have less than 128MB of RAM, they'll probably seem slower than when running under the 4.2-based K12LTSP. OTOH, 4.2 was a highly optimized configuration that did not lend itself to updating with new drivers as they became available. LTSP 5 also much better suited for leveraging the work of the distros, which is paying off well as demonstrated the packages available now for multiple distros. Not to mention local apps. The performance issue *is* of concern to people who, like me, like to use old hardware for clients. But I also recognize that as time goes by, the low end 'used PC' client has become faster & faster. LSP 4.2 isn't being developed any more, and the development model used for it was not sustainable long-term. But if K12LTSP-EL5 does was you want, don't bother with the upgrade. Go with what works for you. The K12Linux installations I've done so far have been quite easy, especially if you start with the Live USB version. But it's only been with a small number of clients. Is it just performance issues you're having or something else? Peter From robark at gmail.com Fri Nov 7 20:14:43 2008 From: robark at gmail.com (Robert Arkiletian) Date: Fri, 7 Nov 2008 12:14:43 -0800 Subject: localapps In-Reply-To: <20081107114518.d7061e97b78b017ac15395d64f2ce134.b9bf3434db.wbe@email.secureserver.net> References: <20081107114518.d7061e97b78b017ac15395d64f2ce134.b9bf3434db.wbe@email.secureserver.net> Message-ID: On Fri, Nov 7, 2008 at 10:45 AM, wrote: > Localapps seem to run slower than a similar app on the server. My thin > clinets are HP T5530 with 256MB. Is there a setting that can be changed > to improve the performance? For example when I start Firefox as an > localapp, it takes about 5 seconds to load. Running Firefox from the > server takes about 2 seconds. It also seems a little slugish when the > localapp is running. Your clients are not very powerful so they cannot run firefox very fast. Plus it will always take longer to launch a local app cause the binary executable has to be copied to local ram over the network. However, the more ram you have in the client the better because your local kernel will cache application data. So the second time you launch firefox it will be faster. I tested this and found you need a min. of 512MB on the client to have firefox cached. I did not see any increase in launching firefox a second time with only 256MB. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ From rmcdaniel at indata.us Fri Nov 7 20:38:53 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 13:38:53 -0700 Subject: K12LTSP centos 5 vs. K12Linux Message-ID: <20081107133852.d7061e97b78b017ac15395d64f2ce134.f8d0a9e35b.wbe@email.secureserver.net> The issues that I have been having with K12Linux have been performance related, some of the problems are certainly the result of the flash intensive app (kids college - learning through sports). I am running another lab with duplicate hardware/setup that is utilizing K12LTSP Centos 5. When the K12LTSP is booted up (30 clients), the clients all load quite fast. The K12Linux lab boots considerably slower and many of the clients seem to be waiting their turn to boot. I don't consider my clients, HPT5530, to be low end. They are much beefier than others that I have been running for 4 years on K12LTSP without any performance issues. I would stick with K12LTSP Centos 5 if I didn't have the frigging sound problem with flash that is so common. The K12Linux installation doesn't have any sound issues. We use another program, Practice Planet, that is a mix of java and flash and IMO better developed for a lab environment. btw, I do appreciate all of the hard work that is being devoted to K12Linux:) Ron -------- Original Message -------- Subject: Re: K12LTSP centos 5 vs. K12Linux From: Peter Scheie Date: Fri, November 07, 2008 1:42 pm To: Development discussion of K12Linux rmcdaniel at indata.us wrote: > The more that I try and get a lab running with K12Linux, the more > frustrated I become. Why does it seem that K12LTSP Centos5 greatly > outperforms K12Linux. Is the difference due to LTSP 4.2 vs. LTSP 5? I > have also given Edubuntu 8.1 a try and end up experiencing all of the > crazy DHCP problems that seem to plague certain installations. Is this > really progression??? Sorry for the rant, but I have been working on a > single lab with 30 clients for two weeks. The K12LTSP install would > have been done in about 2 hours... > > > Ron McDaniel > rmcdaniel at indata.us > > Performance-wise, yes, the demands of LTSP 5 on the client are greater than those of LTSP4.2. If your clients are slower than, say 400mhz and have less than 128MB of RAM, they'll probably seem slower than when running under the 4.2-based K12LTSP. OTOH, 4.2 was a highly optimized configuration that did not lend itself to updating with new drivers as they became available. LTSP 5 also much better suited for leveraging the work of the distros, which is paying off well as demonstrated the packages available now for multiple distros. Not to mention local apps. The performance issue *is* of concern to people who, like me, like to use old hardware for clients. But I also recognize that as time goes by, the low end 'used PC' client has become faster & faster. LSP 4.2 isn't being developed any more, and the development model used for it was not sustainable long-term. But if K12LTSP-EL5 does was you want, don't bother with the upgrade. Go with what works for you. The K12Linux installations I've done so far have been quite easy, especially if you start with the Live USB version. But it's only been with a small number of clients. Is it just performance issues you're having or something else? Peter _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From odin at gnuskole.no Fri Nov 7 20:46:06 2008 From: odin at gnuskole.no (=?ISO-8859-1?Q?Odin_N=F8sen?=) Date: Fri, 7 Nov 2008 21:46:06 +0100 Subject: localapps In-Reply-To: References: <20081107114518.d7061e97b78b017ac15395d64f2ce134.b9bf3434db.wbe@email.secureserver.net> Message-ID: <20081107203707.M57992@gnuskole.no> > launch firefox it will be faster. I tested this and found you need a > min. of 512MB on the client to have firefox cached. I did not see any > increase in launching firefox a second time with only 256MB. In my experience 256MB is just enough to make firefox run without swaping, but it's running fine (5 sec startup on 900mhz klients). I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could run 40 thin clients on 4GB RAM with K12LTSP, but with K12Linux you should run more than 25 thin clients on 8GB RAM. RAM is cheap, so the only problem is old cheap servers that doesn't support more than 4GB RAM. Flash also seems to demand more, but I think that is linked to flash movies and 25 pupils watching flash movies as the same time :-) Localapps is the future! Odin From rmcdaniel at indata.us Fri Nov 7 20:50:38 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 13:50:38 -0700 Subject: localapps Message-ID: <20081107135038.d7061e97b78b017ac15395d64f2ce134.a065ac4f02.wbe@email.secureserver.net> "Localapps is the future" If that is the case, why not just use fat clients (comptuers)? I am a huge advocate of centralized computing, especially in a school environment... Ron -------- Original Message -------- Subject: Re: localapps From: "Odin_N?sen" Date: Fri, November 07, 2008 2:46 pm To: Development discussion of K12Linux > launch firefox it will be faster. I tested this and found you need a > min. of 512MB on the client to have firefox cached. I did not see any > increase in launching firefox a second time with only 256MB. In my experience 256MB is just enough to make firefox run without swaping, but it's running fine (5 sec startup on 900mhz klients). I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could run 40 thin clients on 4GB RAM with K12LTSP, but with K12Linux you should run more than 25 thin clients on 8GB RAM. RAM is cheap, so the only problem is old cheap servers that doesn't support more than 4GB RAM. Flash also seems to demand more, but I think that is linked to flash movies and 25 pupils watching flash movies as the same time :-) Localapps is the future! Odin _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From peter at scheie.homedns.org Fri Nov 7 21:56:25 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Fri, 07 Nov 2008 15:56:25 -0600 Subject: localapps In-Reply-To: <20081107135038.d7061e97b78b017ac15395d64f2ce134.a065ac4f02.wbe@email.secureserver.net> References: <20081107135038.d7061e97b78b017ac15395d64f2ce134.a065ac4f02.wbe@email.secureserver.net> Message-ID: <4914B989.4010302@scheie.homedns.org> Managing a bunch of fat clients is a lot more work than a bunch of thin clients. I think a more accurate statement, as I was once told, would be "Apps run where they should run." rmcdaniel at indata.us wrote: > "Localapps is the future" > > If that is the case, why not just use fat clients (comptuers)? I am a > huge advocate of centralized computing, especially in a school > environment... > > > Ron > > -------- Original Message -------- > Subject: Re: localapps > From: "Odin_N?sen" > Date: Fri, November 07, 2008 2:46 pm > To: Development discussion of K12Linux > > >> launch firefox it will be faster. I tested this and found you need a >> min. of 512MB on the client to have firefox cached. I did not see any >> increase in launching firefox a second time with only 256MB. >> > > In my experience 256MB is just enough to make firefox run without > swaping, but it's > running fine (5 sec startup on 900mhz klients). > > I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could > run 40 thin > clients on 4GB RAM with K12LTSP, but with K12Linux you should run more > than 25 thin > clients on 8GB RAM. RAM is cheap, so the only problem is old cheap > servers that doesn't > support more than 4GB RAM. Flash also seems to demand more, but I think > that is linked > to flash movies and 25 pupils watching flash movies as the same time :-) > Localapps is > the future! > > > Odin > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > From rmcdaniel at indata.us Fri Nov 7 22:20:35 2008 From: rmcdaniel at indata.us (rmcdaniel at indata.us) Date: Fri, 07 Nov 2008 15:20:35 -0700 Subject: localapps Message-ID: <20081107152035.d7061e97b78b017ac15395d64f2ce134.44b429e0c9.wbe@email.secureserver.net> "Apps run where they should run." I can see the logic in that statement. Thanks for the replies. Ron -------- Original Message -------- Subject: Re: localapps From: Peter Scheie Date: Fri, November 07, 2008 3:56 pm To: Development discussion of K12Linux Managing a bunch of fat clients is a lot more work than a bunch of thin clients. I think a more accurate statement, as I was once told, would be "Apps run where they should run." rmcdaniel at indata.us wrote: > "Localapps is the future" > > If that is the case, why not just use fat clients (comptuers)? I am a > huge advocate of centralized computing, especially in a school > environment... > > > Ron > > -------- Original Message -------- > Subject: Re: localapps > From: "Odin_N?sen" > Date: Fri, November 07, 2008 2:46 pm > To: Development discussion of K12Linux > > >> launch firefox it will be faster. I tested this and found you need a >> min. of 512MB on the client to have firefox cached. I did not see any >> increase in launching firefox a second time with only 256MB. >> > > In my experience 256MB is just enough to make firefox run without > swaping, but it's > running fine (5 sec startup on 900mhz klients). > > I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could > run 40 thin > clients on 4GB RAM with K12LTSP, but with K12Linux you should run more > than 25 thin > clients on 8GB RAM. RAM is cheap, so the only problem is old cheap > servers that doesn't > support more than 4GB RAM. Flash also seems to demand more, but I think > that is linked > to flash movies and 25 pupils watching flash movies as the same time :-) > Localapps is > the future! > > > Odin > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > _______________________________________________ K12Linux-devel-list mailing list K12Linux-devel-list at redhat.com https://www.redhat.com/mailman/listinfo/k12linux-devel-list From monteslu at cox.net Fri Nov 7 22:28:09 2008 From: monteslu at cox.net (Luis Montes) Date: Fri, 07 Nov 2008 15:28:09 -0700 Subject: localapps In-Reply-To: <20081107152035.d7061e97b78b017ac15395d64f2ce134.44b429e0c9.wbe@email.secureserver.net> References: <20081107152035.d7061e97b78b017ac15395d64f2ce134.44b429e0c9.wbe@email.secureserver.net> Message-ID: <4914C0F9.3020003@cox.net> I see that too. I just wish that I had more RAM on the 40 or so t150s I got from disklessworkstations a couple of years ago. They have 128 but I think they max out at 256. If I can get them to all to 512 it would be well worth it, I think. The next problem will be to specify icons to launch local apps only on specific machines. I have about 30 newer thin clients with 800Mhz VIAs and 512M ram that I'd like to use localapps, but just those machines specifically for now. Luis rmcdaniel at indata.us wrote: > "Apps run where they should run." > > > I can see the logic in that statement. > > Thanks for the replies. > > Ron > > -------- Original Message -------- > Subject: Re: localapps > From: Peter Scheie > Date: Fri, November 07, 2008 3:56 pm > To: Development discussion of K12Linux > > Managing a bunch of fat clients is a lot more work than a bunch of thin > clients. I think a more accurate statement, as I was once told, would > be "Apps run where they should run." > > rmcdaniel at indata.us wrote: > >> "Localapps is the future" >> >> If that is the case, why not just use fat clients (comptuers)? I am a >> huge advocate of centralized computing, especially in a school >> environment... >> >> >> Ron >> >> -------- Original Message -------- >> Subject: Re: localapps >> From: "Odin_N?sen" >> Date: Fri, November 07, 2008 2:46 pm >> To: Development discussion of K12Linux >> >> >> >>> launch firefox it will be faster. I tested this and found you need a >>> min. of 512MB on the client to have firefox cached. I did not see any >>> increase in launching firefox a second time with only 256MB. >>> >>> >> In my experience 256MB is just enough to make firefox run without >> swaping, but it's >> running fine (5 sec startup on 900mhz klients). >> >> I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could >> run 40 thin >> clients on 4GB RAM with K12LTSP, but with K12Linux you should run more >> than 25 thin >> clients on 8GB RAM. RAM is cheap, so the only problem is old cheap >> servers that doesn't >> support more than 4GB RAM. Flash also seems to demand more, but I think >> that is linked >> to flash movies and 25 pupils watching flash movies as the same time :-) >> Localapps is >> the future! >> >> >> Odin >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> >> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > > From peter at scheie.homedns.org Sat Nov 8 17:40:51 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Sat, 08 Nov 2008 11:40:51 -0600 Subject: localapps In-Reply-To: <4914C0F9.3020003@cox.net> References: <20081107152035.d7061e97b78b017ac15395d64f2ce134.44b429e0c9.wbe@email.secureserver.net> <4914C0F9.3020003@cox.net> Message-ID: <4915CF23.1090500@scheie.homedns.org> I think the idea with local apps is that you'll set a flag in lts.conf for those clients that can support them, and then it shows up as a submenu, under the Applications menu, called Local Applications. If a given client doesn't have the horsepower, then the submenu won't be there (because you've disabled in lts.conf). Peter Luis Montes wrote: > I see that too. I just wish that I had more RAM on the 40 or so t150s > I got from disklessworkstations a couple of years ago. > They have 128 but I think they max out at 256. If I can get them to > all to 512 it would be well worth it, I think. > > The next problem will be to specify icons to launch local apps only on > specific machines. I have about 30 newer thin clients with 800Mhz VIAs > and 512M ram that I'd like to use localapps, but just those machines > specifically for now. > > Luis > > > rmcdaniel at indata.us wrote: >> "Apps run where they should run." >> >> >> I can see the logic in that statement. >> >> Thanks for the replies. >> >> Ron >> >> -------- Original Message -------- >> Subject: Re: localapps >> From: Peter Scheie >> Date: Fri, November 07, 2008 3:56 pm >> To: Development discussion of K12Linux >> >> Managing a bunch of fat clients is a lot more work than a bunch of >> thin clients. I think a more accurate statement, as I was once told, >> would be "Apps run where they should run." >> >> rmcdaniel at indata.us wrote: >> >>> "Localapps is the future" >>> >>> If that is the case, why not just use fat clients (comptuers)? I am a >>> huge advocate of centralized computing, especially in a school >>> environment... >>> >>> >>> Ron >>> >>> -------- Original Message -------- >>> Subject: Re: localapps >>> From: "Odin_N?sen" >>> Date: Fri, November 07, 2008 2:46 pm >>> To: Development discussion of K12Linux >>> >>> >>> >>>> launch firefox it will be faster. I tested this and found you need a >>>> min. of 512MB on the client to have firefox cached. I did not see any >>>> increase in launching firefox a second time with only 256MB. >>>> >>>> >>> In my experience 256MB is just enough to make firefox run without >>> swaping, but it's >>> running fine (5 sec startup on 900mhz klients). >>> >>> I must agree about K12Linux been "slower" than K12LTSP 5.2EL. We could >>> run 40 thin >>> clients on 4GB RAM with K12LTSP, but with K12Linux you should run more >>> than 25 thin >>> clients on 8GB RAM. RAM is cheap, so the only problem is old cheap >>> servers that doesn't >>> support more than 4GB RAM. Flash also seems to demand more, but I think >>> that is linked >>> to flash movies and 25 pupils watching flash movies as the same time >>> :-) >>> Localapps is >>> the future! >>> >>> >>> Odin >>> >>> _______________________________________________ >>> K12Linux-devel-list mailing list >>> K12Linux-devel-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >>> >>> >>> >>> >>> _______________________________________________ >>> K12Linux-devel-list mailing list >>> K12Linux-devel-list at redhat.com >>> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >>> >>> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> >> >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> >> > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list From wtogami at redhat.com Sat Nov 8 17:58:37 2008 From: wtogami at redhat.com (Warren Togami) Date: Sat, 08 Nov 2008 12:58:37 -0500 Subject: localapps In-Reply-To: <4915CF23.1090500@scheie.homedns.org> References: <20081107152035.d7061e97b78b017ac15395d64f2ce134.44b429e0c9.wbe@email.secureserver.net> <4914C0F9.3020003@cox.net> <4915CF23.1090500@scheie.homedns.org> Message-ID: <4915D34D.7000908@redhat.com> Peter Scheie wrote: > I think the idea with local apps is that you'll set a flag in lts.conf > for those clients that can support them, and then it shows up as a > submenu, under the Applications menu, called Local Applications. If a > given client doesn't have the horsepower, then the submenu won't be > there (because you've disabled in lts.conf). > We might also add some kind of auto-detection during client login, to disregard the lts.conf localapps option if the client has less than 512MB RAM, for example. Warren From balmquist at mindfirestudios.com Tue Nov 11 22:34:39 2008 From: balmquist at mindfirestudios.com (Almquist Burke) Date: Tue, 11 Nov 2008 16:34:39 -0600 Subject: k12linux wikki updates Message-ID: <65E6C79A-3D9D-4472-AD0E-75A8C8ADDB53@mindfirestudios.com> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I just made a couple updates to the install guide, networking setup, and advanced networking setup pages on the wikki at fedorahosted.org/ k12linux, now that the K12LTSP FC9 RC1 respin has been out for a while. A couple people here on the list have talked about how to do a single NIC setup. I've set aside the AdvancedNetworkSetup page for information about the Single NIC setup. If anyone wants to write up their experience doing a single NIC setup as a guide and put it up on the wikki, that would be greatly appreciated. If you don't have wikki access, you can certainly email it to me and I'd be happy to put it up. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (Darwin) iEYEARECAAYFAkkaCH8ACgkQxWV7OPa/g5FMRwCfUv2HpRgVd6xe9Kq+IvjIIc/P c+EAn0EVZSF4lcRtneuAztYwvP4ReQEc =fm5B -----END PGP SIGNATURE----- From andrew at mobileitsolutions.co.uk Wed Nov 12 09:59:31 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Wed, 12 Nov 2008 09:59:31 +0000 Subject: ltsp-dhcpd not starting at start up Message-ID: <6befb72f0811120159k5b5b18d4x7c3cd947dad8c817@mail.gmail.com> ltsp-dhcpd wont start at start-up, i seem to have to start it in services on ever boot, was wondering if I had missed something in the setup instructions? -- Andrew M Osborne -------------- next part -------------- An HTML attachment was scrubbed... URL: From peter at scheie.homedns.org Wed Nov 12 13:31:16 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Wed, 12 Nov 2008 07:31:16 -0600 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <6befb72f0811120159k5b5b18d4x7c3cd947dad8c817@mail.gmail.com> References: <6befb72f0811120159k5b5b18d4x7c3cd947dad8c817@mail.gmail.com> Message-ID: <491ADAA4.5070507@scheie.homedns.org> Run 'chkconfig --list ltsp-dhcpd' to see whether it's currently set to startup at boot. Run 'chkconfig ltsp-dhcpd on' to set it to startup at boot if it's not. Peter Andrew Osborne wrote: > ltsp-dhcpd wont start at start-up, i seem to have to start it in > services on ever boot, was wondering if I had missed something in the > setup instructions? > > -- > Andrew M Osborne > > > > ------------------------------------------------------------------------ > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list From andrew at mobileitsolutions.co.uk Wed Nov 12 13:55:28 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Wed, 12 Nov 2008 13:55:28 +0000 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <491ADAA4.5070507@scheie.homedns.org> References: <6befb72f0811120159k5b5b18d4x7c3cd947dad8c817@mail.gmail.com> <491ADAA4.5070507@scheie.homedns.org> Message-ID: <6befb72f0811120555i21b3ee48h44e54cc61a21b792@mail.gmail.com> Correction, the ltsp-dhcpd service starts but says 'this service is been refreshed' but never does, not until I start it in Services. Andrew 2008/11/12 Peter Scheie > Run 'chkconfig --list ltsp-dhcpd' to see whether it's currently set to > startup at boot. Run 'chkconfig ltsp-dhcpd on' to set it to startup at boot > if it's not. > > Peter > > Andrew Osborne wrote: > >> ltsp-dhcpd wont start at start-up, i seem to have to start it in services >> on ever boot, was wondering if I had missed something in the setup >> instructions? >> >> -- >> Andrew M Osborne >> >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> K12Linux-devel-list mailing list >> K12Linux-devel-list at redhat.com >> https://www.redhat.com/mailman/listinfo/k12linux-devel-list >> > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > -- Andrew M Osborne www.mobileitsolutions.co.uk -------------- next part -------------- An HTML attachment was scrubbed... URL: From william at fragakis.com Wed Nov 12 18:08:07 2008 From: william at fragakis.com (William Fragakis) Date: Wed, 12 Nov 2008 13:08:07 -0500 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <20081112170050.6C4718E00E4@hormel.redhat.com> References: <20081112170050.6C4718E00E4@hormel.redhat.com> Message-ID: <1226513287.29431.62.camel@server.ltsp> I, too, encountered the same issue, necessitating the same fix. ltsp-dhcpd was configured to run at boot and was showing enabled but wouldn't run unless restarted. wf On Wed, 2008-11-12 at 12:00 -0500, k12linux-devel-list-request at redhat.com wrote: > Message: 4 > Date: Wed, 12 Nov 2008 13:55:28 +0000 > From: "Andrew Osborne" > Subject: Re: ltsp-dhcpd not starting at start up > To: k12linux-devel-list at redhat.com > Message-ID: > <6befb72f0811120555i21b3ee48h44e54cc61a21b792 at mail.gmail.com> > Content-Type: text/plain; charset="iso-8859-1" > > Correction, the ltsp-dhcpd service starts but says 'this service is > been > refreshed' but never does, not until I start it in Services. > > > Andrew > From robark at gmail.com Wed Nov 12 18:15:12 2008 From: robark at gmail.com (Robert Arkiletian) Date: Wed, 12 Nov 2008 10:15:12 -0800 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <1226513287.29431.62.camel@server.ltsp> References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> Message-ID: On Wed, Nov 12, 2008 at 10:08 AM, William Fragakis wrote: > I, too, encountered the same issue, necessitating the same fix. > > ltsp-dhcpd was configured to run at boot and was showing enabled but > wouldn't run unless restarted. I also noticed this also. I noticed my bridged network device was always down upon boot so I would bring it up before running ltsp-dhcpd. -- Robert Arkiletian Eric Hamber Secondary, Vancouver, Canada Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/ C++ GUI tutorial http://www3.telus.net/public/robark/ From wtogami at redhat.com Wed Nov 12 19:01:18 2008 From: wtogami at redhat.com (Warren Togami) Date: Wed, 12 Nov 2008 14:01:18 -0500 Subject: ltsp-dhcpd not starting at start up In-Reply-To: References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> Message-ID: <491B27FE.7060109@redhat.com> Robert Arkiletian wrote: > On Wed, Nov 12, 2008 at 10:08 AM, William Fragakis wrote: >> I, too, encountered the same issue, necessitating the same fix. >> >> ltsp-dhcpd was configured to run at boot and was showing enabled but >> wouldn't run unless restarted. > > I also noticed this also. I noticed my bridged network device was > always down upon boot so I would bring it up before running > ltsp-dhcpd. > chkconfig network on NetworkManager does not bring up bridges, but network does. You can enable both services without apparent trouble. Warren From andrew at mobileitsolutions.co.uk Thu Nov 13 18:21:49 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Thu, 13 Nov 2008 18:21:49 +0000 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <491B27FE.7060109@redhat.com> References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> <491B27FE.7060109@redhat.com> Message-ID: <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> Ran through the suggestions, but still will not start without going into service after boot? Andrew 2008/11/12 Warren Togami > Robert Arkiletian wrote: > >> On Wed, Nov 12, 2008 at 10:08 AM, William Fragakis >> wrote: >> >>> I, too, encountered the same issue, necessitating the same fix. >>> >>> ltsp-dhcpd was configured to run at boot and was showing enabled but >>> wouldn't run unless restarted. >>> >> >> I also noticed this also. I noticed my bridged network device was >> always down upon boot so I would bring it up before running >> ltsp-dhcpd. >> >> > chkconfig network on > > NetworkManager does not bring up bridges, but network does. You can enable > both services without apparent trouble. > > Warren > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > -- Andrew M Osborne www.mobileitsolutions.co.uk -------------- next part -------------- An HTML attachment was scrubbed... URL: From burke at thealmquists.net Thu Nov 13 23:33:37 2008 From: burke at thealmquists.net (Almquist Burke) Date: Thu, 13 Nov 2008 17:33:37 -0600 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> <491B27FE.7060109@redhat.com> <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Nov 13, 2008, at 12:21 PM, Andrew Osborne wrote: > Ran through the suggestions, but still will not start without going > into service after boot? > > Crazy question, is the plain DHCPD service disabled? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (Darwin) iEYEARECAAYFAkkcuVIACgkQxWV7OPa/g5HrFwCfYoVDdGJ2ZX84jAo1BXK0+C5U n14AnR5Ioa1wieg+vm2MPGDh074D5SVJ =/Ru+ -----END PGP SIGNATURE----- From andrew at mobileitsolutions.co.uk Fri Nov 14 08:50:35 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Fri, 14 Nov 2008 08:50:35 +0000 Subject: ltsp-dhcpd not starting at start up In-Reply-To: References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> <491B27FE.7060109@redhat.com> <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> Message-ID: <6befb72f0811140050q371860a3ye51d1fa8b8a408e8@mail.gmail.com> Please forgive the confusion. ltsp-dhcpd wont start at start-up, i seem to have to start it in services on every boot. *Run 'chkconfig --list ltsp-dhcpd' to see whether it's currently set to startup at boot. Run 'chkconfig ltsp-dhcpd on' to set it to startup at boot if it's not.* ** thank you, but still not working, ** tried this, the service is enabled but will not run, still needs manual click in services, this is a Fedora 9 64bit installation, network service is enabled and running, this problem does not happen with the F9 LIVE USB install. this problem is on two servers of different makes. this problem does not happen on Fedora 10. the system is running all K12Linux defaults. thanks in advance. Plain DHCPD service is disabled. Andrew. 2008/11/13 Almquist Burke > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Nov 13, 2008, at 12:21 PM, Andrew Osborne wrote: > > Ran through the suggestions, but still will not start without going into >> service after boot? >> >> >> > Crazy question, is the plain DHCPD service disabled? > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.7 (Darwin) > > iEYEARECAAYFAkkcuVIACgkQxWV7OPa/g5HrFwCfYoVDdGJ2ZX84jAo1BXK0+C5U > n14AnR5Ioa1wieg+vm2MPGDh074D5SVJ > =/Ru+ > -----END PGP SIGNATURE----- > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > -- Andrew M Osborne www.mobileitsolutions.co.uk -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew at mobileitsolutions.co.uk Fri Nov 14 08:50:35 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Fri, 14 Nov 2008 08:50:35 +0000 Subject: ltsp-dhcpd not starting at start up In-Reply-To: References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> <491B27FE.7060109@redhat.com> <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> Message-ID: <6befb72f0811140050q371860a3ye51d1fa8b8a408e8@mail.gmail.com> Please forgive the confusion. ltsp-dhcpd wont start at start-up, i seem to have to start it in services on every boot. *Run 'chkconfig --list ltsp-dhcpd' to see whether it's currently set to startup at boot. Run 'chkconfig ltsp-dhcpd on' to set it to startup at boot if it's not.* ** thank you, but still not working, ** tried this, the service is enabled but will not run, still needs manual click in services, this is a Fedora 9 64bit installation, network service is enabled and running, this problem does not happen with the F9 LIVE USB install. this problem is on two servers of different makes. this problem does not happen on Fedora 10. the system is running all K12Linux defaults. thanks in advance. Plain DHCPD service is disabled. Andrew. 2008/11/13 Almquist Burke > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Nov 13, 2008, at 12:21 PM, Andrew Osborne wrote: > > Ran through the suggestions, but still will not start without going into >> service after boot? >> >> >> > Crazy question, is the plain DHCPD service disabled? > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.7 (Darwin) > > iEYEARECAAYFAkkcuVIACgkQxWV7OPa/g5HrFwCfYoVDdGJ2ZX84jAo1BXK0+C5U > n14AnR5Ioa1wieg+vm2MPGDh074D5SVJ > =/Ru+ > -----END PGP SIGNATURE----- > > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > -- Andrew M Osborne www.mobileitsolutions.co.uk -------------- next part -------------- An HTML attachment was scrubbed... URL: From burke at thealmquists.net Fri Nov 14 23:08:27 2008 From: burke at thealmquists.net (Almquist Burke) Date: Fri, 14 Nov 2008 17:08:27 -0600 Subject: ltsp-dhcpd not starting at start up In-Reply-To: <6befb72f0811140050q371860a3ye51d1fa8b8a408e8@mail.gmail.com> References: <20081112170050.6C4718E00E4@hormel.redhat.com> <1226513287.29431.62.camel@server.ltsp> <491B27FE.7060109@redhat.com> <6befb72f0811131021p60b4c5f5w9a68119c01506c80@mail.gmail.com> <6befb72f0811140050q371860a3ye51d1fa8b8a408e8@mail.gmail.com> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 What does the log say when it tries to start DHCPD? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (Darwin) iEYEARECAAYFAkkeBOsACgkQxWV7OPa/g5G4WwCfYsS0tON298i6ORV6vldzyVAg F54An2tZ52cb+7jxz8SxkWBvdTd9IIZx =iutr -----END PGP SIGNATURE----- From mcosta at fc.up.pt Mon Nov 17 12:02:05 2008 From: mcosta at fc.up.pt (Miguel Dias Costa) Date: Mon, 17 Nov 2008 12:02:05 +0000 Subject: K12Linux without dhcp Message-ID: <49215D3D.9020309@fc.up.pt> Hello all. I'm new to ltsp, I've only tried K12Linux with the virtual machine client, and it seems great, but I do have one question - is it possible/simple to deploy clients that instead of booting from pxe would boot from a floppy/cd/usb and connect to a specific ip address? Cheers, Miguel From stretchem at gmail.com Mon Nov 17 17:37:23 2008 From: stretchem at gmail.com (Mike Rathburn) Date: Mon, 17 Nov 2008 12:37:23 -0500 Subject: New Fedora9 Installation and Networking Message-ID: <7cee901e0811170937l68826422ge3901537a8f0f19b@mail.gmail.com> I'm attempting to test out LTSP5. Installed Fedora 9 i386. Used default settings throughout installation, except for the network interfaces use: eth0 - Static - 192.168.0.254/255.255.255.0/192.168.1.1 eth1 - Static - 192.168.1.25/255.255.255.0/192.168.1.1 Followed instructions exactly as posted here: http://www.ltsp.org/~sbalneav/LTSPManual.html#AEN537 and then here for 'A Permanent Network Setup': https://fedorahosted.org/k12linux/wiki/NetworkSetup Note that during this entire install process networking is fine, internet updates are being successfully made. Rebooted. When the machine boots up, there is no networking. Log into the GUI console, notice that Network Manager icon in the upper right is inactive. From what I've read that's a good thing (?). I navigate to System -> Administration -> Network, and notice that all three network devices are 'Inactive'. eth0 - Inactive eth1 - Inactive ltspbr0 - Inactive If I click on eth0, then choose Activate, nothing happens. If I choose eth1 and Activate, it works. If I choose ltspbr0, it activates. After doing the above, networking now works in the sense that the machine can see the internet. Plugged in a thin client (Diskless Workstations T1420 PXE) into what I'm thinking is the 192.168.0.x network switch configured for the TC's, and it's not getting DHCP. Looking at /var/log/messages, I see: Nov 17 11:15:59 ltsp avahi-daemon[2175]: Joining mDNS multicast group on interface ltspbr0.IPv4 with address 172.31.100.254. Nov 17 11:15:59 ltsp avahi-daemon[2175]: New relevant interface ltspbr0.IPv4 for mDNS. Nov 17 11:15:59 ltsp avahi-daemon[2175]: Registering new address record for 172.31.100.254 on ltspbr0.IPv4. So I've obviously missed something. 1. The interfaces aren't starting at system startup. 2. Can't activate eth0 at all. 3. TC not getting DHCP to boot up. Where did I go wrong? Thanks! -M Rathburn Children First Florida From andrew at mobileitsolutions.co.uk Mon Nov 17 17:54:11 2008 From: andrew at mobileitsolutions.co.uk (Andrew Osborne) Date: Mon, 17 Nov 2008 17:54:11 +0000 Subject: New Fedora9 Installation and Networking In-Reply-To: <7cee901e0811170937l68826422ge3901537a8f0f19b@mail.gmail.com> References: <7cee901e0811170937l68826422ge3901537a8f0f19b@mail.gmail.com> Message-ID: <6befb72f0811170954t79333572wd0a957ad21615b59@mail.gmail.com> this may help in service on th gui, enable and start the NETWORK service 2008/11/17 Mike Rathburn > I'm attempting to test out LTSP5. Installed Fedora 9 i386. Used > default settings throughout installation, except for the network > interfaces use: > > eth0 - Static - 192.168.0.254/255.255.255.0/192.168.1.1 > eth1 - Static - > 192.168.1.25/255.255.255.0/192.168.1.1 > > Followed instructions exactly as posted here: > http://www.ltsp.org/~sbalneav/LTSPManual.html#AEN537 > > and then here for 'A Permanent Network Setup': > https://fedorahosted.org/k12linux/wiki/NetworkSetup > > Note that during this entire install process networking is fine, > internet updates are being successfully made. > > Rebooted. > > When the machine boots up, there is no networking. Log into the GUI > console, notice that Network Manager icon in the upper right is > inactive. From what I've read that's a good thing (?). > > I navigate to System -> Administration -> Network, and notice that all > three network devices are 'Inactive'. > > eth0 - Inactive > eth1 - Inactive > ltspbr0 - Inactive > > If I click on eth0, then choose Activate, nothing happens. If I > choose eth1 and Activate, it works. If I choose ltspbr0, it > activates. > > After doing the above, networking now works in the sense that the > machine can see the internet. Plugged in a thin client (Diskless > Workstations T1420 PXE) into what I'm thinking is the 192.168.0.x > network switch configured for the TC's, and it's not getting DHCP. > Looking at /var/log/messages, I see: > > Nov 17 11:15:59 ltsp avahi-daemon[2175]: Joining mDNS multicast group > on interface ltspbr0.IPv4 with address 172.31.100.254. > Nov 17 11:15:59 ltsp avahi-daemon[2175]: New relevant interface > ltspbr0.IPv4 for mDNS. > Nov 17 11:15:59 ltsp avahi-daemon[2175]: Registering new address > record for 172.31.100.254 on ltspbr0.IPv4. > > So I've obviously missed something. > > 1. The interfaces aren't starting at system startup. > 2. Can't activate eth0 at all. > 3. TC not getting DHCP to boot up. > > Where did I go wrong? Thanks! > > -M Rathburn > Children First Florida > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > -- Andrew M Osborne www.mobileitsolutions.co.uk -------------- next part -------------- An HTML attachment was scrubbed... URL: From william at fragakis.com Mon Nov 17 19:37:55 2008 From: william at fragakis.com (William Fragakis) Date: Mon, 17 Nov 2008 14:37:55 -0500 Subject: K12LTSP5 EL + K12Linux Sound working and localdev (sort of) In-Reply-To: <20081117170032.0B4096196B0@hormel.redhat.com> References: <20081117170032.0B4096196B0@hormel.redhat.com> Message-ID: <1226950675.7530.65.camel@server.ltsp> Hi I've been able to modify a K12LTSP5 EL installation to use LTSP5 with working sound and (sort of, see below) working local devices. I hope the following helps. Let me know if you run into any problems following what I have below because I was never the best at taking notes. It requires the use of some f8 and f9 rpms and you are definitely skiing off-piste here - you are on your own. But it does work in my limited testing. Unfortunately, I had to reinvent Warren Togami's wheel a bit because I had googled for Centos instead of RHEL and ended up sorting out what he quite elegantly details in his nice guide into getting a base install of LTSP 5 onto RHEL5 here: https://fedorahosted.org/k12linux/wiki/RHEL5Server I began with a vanilla install of K12LTSP 5 EL (Centos 5) 32 bit and ran yum update which brought the installation up to ver. 5.2. Lacking Warren's tutorial, I copied the relevant files over from a working F9 K12Linux install and edit configuration files accordingly. I had to copy the /opt/ltsp/i386 directly via ssh because uncompressing or burning the directory kept generating errors because of the symlinks that were beyond my proficiency to correct. I retained the legacy networking scheme. If you are unfamiliar with installing k12Linux, I highly recommend proceeding from Warren's tutorial. For those of you who want to skip to the good part... from Warren - rpm -Uhv http://download.fedora.redhat.com/pub/epel/5/i386/epel-release-5-2.noarch.rpm yum install ltsp-server copy in new client root (copy from working installation or Warren's tarball run ltsp-update-sshkeys and ltsp-update-kernels set up networking and enable new services post (after) Warren's tutorial for sound: yum remove ltsp_esound* (this may obsolete the next step but I came from a different direction. Please let me know) rpm -ev esound --nodeps from http://www.gtlib.gatech.edu/pub/fedora.redhat/linux/releases/8/Everything/i386/os/Packages/ or your favorite f8 mirror, install the following: these are require to install alsa-lib, iirc glibc-devel-2.7-2.i386.rpm glibc-2.7-2.i686.rpm glibc-headers-2.7-2.i386.rpm glibc-common-2.7-2.i386.rpm glibc-utils-2.7-2.i386.rpm ncurses-5.6-12.20070812.fc8.i386.rpm next install alsa-lib-1.0.15-1.fc8.i386.rpm alsa-plugins-pulseaudio-1.0.14-5.fc8.i386.rpm alsa-utils-1.0.15-1.fc8.i386.rpm alsa-plugins-pulseaudio.i386 1.0.15-3.fc8.1 the pulse packages come from rhel5 repository, iirc gstreamer-plugins-pulse.i386 0.9.5-0.5.svn20070924 pulseaudio.i386 0.9.10-1.el5.1 pulseaudio-core-libs.i386 0.9.10-1.el5.1 pulseaudio-esound-compat.i386 0.9.10-1.el5.1 i pulseaudio-libs.i386 0.9.10-1.el5.1 pulseaudio-libs-glib2.i386 0.9.10-1.el5.1 pulseaudio-libs-zeroconf.i386 0.9.10-1.el5.1 pulseaudio-module-gconf.i386 0.9.10-1.el5.1 pulseaudio-module-x11.i386 0.9.10-1.el5.1 pulseaudio-module-zeroconf.i386 0.9.10-1.el5.1 pulseaudio-utils.i386 0.9.10-1.el5.1 while you are adding stuff: afaik, you'll need this to support nbd swap; nbd-2.9.7-3.fc8.i386.rpm add /etc/asound.conf with the following contents #Generated by system-config-soundcard, do not edit by hand #SWCONF #DEV 0 defaults.pcm.card 0 defaults.pcm.device 0 defaults.ctl.card 0 pcm.pulse { type pulse } ctl.pulse { type pulse } pcm.!default { type pulse } ctl.!default { type pulse } compare/create /etc/alsa/pulse-default.conf so it looks like this: # PulseAudio plugin configuration # $Id: pulse-default.conf,v 1.3 2008/03/09 15:50:49 lkundrak Exp $ # Let's create a virtual device "pulse" for mixer and PCM pcm.pulse { type pulse hint { description "PulseAudio Sound Server" } } ctl.pulse { type pulse hint { description "PulseAudio Sound Server" } } # Let's make it the default! pcm.!default { type pulse hint { description "Default" } } ctl.!default { type pulse hint { description "Default" } } go to System -> Preferences -> Sound and choose Alsa for your output. If you get gsink errors, make sure that your pulse-default.conf and asound.conf files are correct. You should be able to play test tones on your client. update or install flash-plugin (Flash 10) Now, you should be able to watch something fun like Snow Patrol on youtube: http://www.youtube.com/watch?v=RZCeFvM-5yc&feature=dir _______________ for local devices - install ltspfs-05.5-1.fc9.i386.rpm update fuse.i386.2.7.3-2.el5 dkms-fuse.noarch.2.7.3-0 install fuse-libs.i386.2.7.3-2.el5 for using local devices, I could only get usb sticks to work. After upgrading to the most recent ltspfs package, I couldn't generate /dev/cdrom. Anyone have any ideas? I don't know if it's an issue with ltspfs or cdpinger, etc. I could at least get an cdrom icon before the update. Double check this file or fuse will create incorrect permissions. /etc/udev/rules.d/99-fuse.rules (correct Group from root to fuse) KERNEL=="fuse", NAME="%k", MODE="0666",OWNER="root",GROUP="fuse" add your regular user to the fuse group. Your usb stick will show up in /media/foouser/ K12Linux uses gvfs to automatically have an icon appear for you. Perhaps someone can wrestle with ltspfsmounter (?) and generate an icon/link on the user desktop, etc. post mortem openal timidity icewm fail dep resolution - require lib.esd usb stick appears in obscure location. Sound works great. My thanks to Peter Scheie who allowed me to bounce some ideas off of him. Regards, William Fragakis From william at fragakis.com Mon Nov 17 19:47:54 2008 From: william at fragakis.com (William Fragakis) Date: Mon, 17 Nov 2008 14:47:54 -0500 Subject: K12LTSP5 EL + K12Linux Sound working and localdev (sort of)] Message-ID: <1226951274.7530.70.camel@server.ltsp> Hello all, skipped one important detail > from Warren - > rpm -Uhv > http://download.fedora.redhat.com/pub/epel/5/i386/epel-release-5-2.noarch.rpm install temporary RHEL5 k12linux repo. see step two on his page or ### Add the temporary RHEL5 k12linux repository # wget http://people.redhat.com/wtogami/temp/fedora9-ltsp-chroot-20080912.tar.bz2 # rpm -Uhv k12linux-temp-release-5-1.el5.noarch.rpm > > yum install ltsp-server > [snip] Thanks, William From burke at thealmquists.net Tue Nov 18 00:08:17 2008 From: burke at thealmquists.net (Almquist Burke) Date: Mon, 17 Nov 2008 18:08:17 -0600 Subject: New Fedora9 Installation and Networking In-Reply-To: <6befb72f0811170954t79333572wd0a957ad21615b59@mail.gmail.com> References: <7cee901e0811170937l68826422ge3901537a8f0f19b@mail.gmail.com> <6befb72f0811170954t79333572wd0a957ad21615b59@mail.gmail.com> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Nov 17, 2008, at 11:54 AM, Andrew Osborne wrote: > this may help > > in service on th gui, enable and start the NETWORK service I updated the wikki last week, and that step is included. Just make sure you didn't miss anything under the NetworkSetup on the wikki. If you miss something, it likely won't work. You may want to restart the server at this point; I always do just to make sure that it will come up cleanly on reboot without manual starting/stopping of services afterwards. If you don't, you MAY need to cycle some of the network services (I don't know cause I always restart when I get everything configured, per the reason above). Oh, and don't worry about eth0 being up, it's attached to the ltspbr0 bridge device, so it's not supposed to be active. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (Darwin) iEYEARECAAYFAkkiB3EACgkQxWV7OPa/g5FPDwCbBgGjPTEKJRKRPTpO8MuyIViW qGEAnjoj/tdHwngHeZd8bPxEtyykvmtt =/h4Q -----END PGP SIGNATURE----- From william at fragakis.com Wed Nov 19 19:47:36 2008 From: william at fragakis.com (william at fragakis.com) Date: Wed, 19 Nov 2008 11:47:36 -0800 (PST) Subject: latest client update killed printing Message-ID: <457cb1edd9a89f23ca34c58503cfdafc.squirrel@webmail.fragakis.com> Hi, I ran yum update yesterday in my client chroot and something has killed printing to clients (separate clients with different printers). Cups reports: "recoverable: Network host 'ws152' is busy; will retry in 30 seconds..." printers are set up via app socket/jet direct socket://client:9100 I've seen this error before when the printer_0 setting wasn't set correctly in lts.conf. (yes, it worked before and I've checked the settings) No changes were made on the server, just the client chroot. Additionally, reverting to an earlier image before the update restores printing. I encountered this issue with a second server, too, that had just run an update. My last update included the following packages: Nov 17 13:31:35 Updated: elfutils-libelf-0.137-3.fc9.i386 Nov 17 13:31:35 Updated: gnutls-2.0.4-4.fc9.i386 Nov 17 13:31:35 Updated: xorg-x11-server-common-1.5.2-3.fc9.i386 Nov 17 13:31:36 Installed: kernel-firmware-2.6.27.5-37.fc9.noarch Nov 17 13:31:37 Updated: xorg-x11-server-Xorg-1.5.2-3.fc9.i386 Nov 17 13:31:37 Updated: xorg-x11-drv-i810-2.3.2-3.fc9.i386 Nov 17 13:31:58 Installed: kernel-2.6.27.5-37.fc9.i586 Nov 17 13:31:58 Updated: ldm-2.0.18-1.fc9.i386 Nov 17 13:31:59 Updated: ltsp-client-5.1.34-2.fc9.i386 Nov 17 13:32:00 Updated: 2:tar-1.19-5.fc9.i386 Nov 17 13:32:01 Installed: kernel-2.6.27.5-37.fc9.i586 Regards, William From wtogami at redhat.com Thu Nov 20 02:22:11 2008 From: wtogami at redhat.com (Warren Togami) Date: Wed, 19 Nov 2008 21:22:11 -0500 Subject: latest client update killed printing In-Reply-To: <457cb1edd9a89f23ca34c58503cfdafc.squirrel@webmail.fragakis.com> References: <457cb1edd9a89f23ca34c58503cfdafc.squirrel@webmail.fragakis.com> Message-ID: <4924C9D3.3050006@redhat.com> william at fragakis.com wrote: > Hi, > I ran yum update yesterday in my client chroot and something has killed > printing to clients (separate clients with different printers). Cups > reports: > "recoverable: Network host 'ws152' is busy; will retry in 30 seconds..." > > printers are set up via app socket/jet direct socket://client:9100 > I've seen this error before when the printer_0 setting wasn't set > correctly in lts.conf. (yes, it worked before and I've checked the > settings) > > No changes were made on the server, just the client chroot. Additionally, > reverting to an earlier image before the update restores > printing. I encountered this issue with a second server, too, that had > just run an update. > In your lts.conf please enable shells on the other VT's. After logging in as a user, switch to a VT and check if jetpipe is running at all. Is it? Warren From wtogami at redhat.com Tue Nov 25 03:40:39 2008 From: wtogami at redhat.com (Warren Togami) Date: Mon, 24 Nov 2008 22:40:39 -0500 Subject: Help Needed: jetpipe testing Message-ID: <492B73B7.2040907@redhat.com> Is anyone successfully using jetpipe with F9 K12Linux? William Fragakis indicated that jetpipe was working for him until he updated his client chroot with the latest ltsp update. I personally have never used jetpipe and lack local printing hardware so I cannot test this easily. If you are currently successfully using jetpipe, could you please update your client chroot to the latest F9 updates and see if it still works? Peter Scheie and I want to go ahead with a K12Linux F9 final relesae, but I want to be sure this jetpipe issue is fixed before we go ahead. Warren Togami wtogami at redhat.com From odin at gnuskole.no Tue Nov 25 09:32:33 2008 From: odin at gnuskole.no (=?ISO-8859-1?Q?Odin_N=F8sen?=) Date: Tue, 25 Nov 2008 10:32:33 +0100 Subject: vncts doesn't work Message-ID: <20081125091550.M53161@gnuskole.no> I can't get vncts to work in fedora 9 (or fedora 10). I only get the grey screen when I try to "vncviewer mymachine:5900" (or :5903). Any hints? I can't find a Xaccess-file to edit - which is what I'm supposed to according to the vncts-file in xinetd.d. I've enabled XDMCP in /etc/gdm/custom.conf, restarted GDM, but nothing works. I always get the grey screen. Has anyone got it working? "help me!" from Odin From peter at scheie.homedns.org Tue Nov 25 12:22:58 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Tue, 25 Nov 2008 06:22:58 -0600 Subject: K12Linux Quick Start Guide Message-ID: <492BEE22.4000503@scheie.homedns.org> The latest version of the K12Linux Quick Start Guide ver. 3, formerly known as the README LTSP Server Setup, can be seen at http://petre.homedns.org/k12linux/k12linux-quick-start-guide/k12linux-quick-start-guide.html It is also available in an RPM at http://petre.homedns.org/k12linux/k12linux-quick-start-guide-0.0.3-1.fc9.noarch.rpm This QSG contains the instructions for enabling the LTSP networking on K12Linux. It is intended for use in both LiveUSB and installed-to-hard-disk situations. Please have a look at it and let me know if you see any mistakes or other issues. Thanks. Peter From monteslu at cox.net Tue Nov 25 15:41:22 2008 From: monteslu at cox.net (Luis Montes) Date: Tue, 25 Nov 2008 08:41:22 -0700 Subject: Flash working with sound on f9 64bit server Message-ID: <492C1CA2.20207@cox.net> Though I was able to get flash to work with nspluginwrapper on my 64bit server, I was never actually able to get sound to work. This morning I just gave the flash 64 bit beta a shot and it worked! http://blogs.adobe.com/penguin.swf/ I removed both 32bit & 64bit nspluginwrappers as well at the 32bit flash plugin itself. Then I downloaded the 64bit plugin from the adobe site, then extracted the libflashplayer.so into /usr/lib64/mozilla/plugins/ I made sure that the permissions for that binary are the same as the other plugins. Launched firefox and sound was working in flash, even youtube. I realize that the best way to run firefox/flash would be as a localapp, but my thin clients are way underpowered for that. This whole time I thought adobe was ignoring the 64bit request, turns out they were actually working on it :) Now I can't think of a reason to run a 32 bit OS on the server if your CPU has x86_64 chips. Luis From peter at scheie.homedns.org Wed Nov 26 02:33:44 2008 From: peter at scheie.homedns.org (Peter Scheie) Date: Tue, 25 Nov 2008 20:33:44 -0600 Subject: vncts doesn't work In-Reply-To: <20081125091550.M53161@gnuskole.no> References: <20081125091550.M53161@gnuskole.no> Message-ID: <492CB588.8070701@scheie.homedns.org> Hmm. I've got it working on my F9/K12Linux box. I wrote some notes to myself a few years ago about setting it up. You can find them at http://petre.homedns.org/unix/vnc_via_xinetd.html HTH Peter Odin N?sen wrote: > I can't get vncts to work in fedora 9 (or fedora 10). I only get the grey screen when I > try to "vncviewer mymachine:5900" (or :5903). Any hints? > > I can't find a Xaccess-file to edit - which is what I'm supposed to according to the > vncts-file in xinetd.d. I've enabled XDMCP in /etc/gdm/custom.conf, restarted GDM, but > nothing works. I always get the grey screen. > > Has anyone got it working? > > > "help me!" from Odin > > _______________________________________________ > K12Linux-devel-list mailing list > K12Linux-devel-list at redhat.com > https://www.redhat.com/mailman/listinfo/k12linux-devel-list > From stretchem at gmail.com Wed Nov 26 04:21:14 2008 From: stretchem at gmail.com (M Rathburn) Date: Tue, 25 Nov 2008 23:21:14 -0500 Subject: vncts doesn't work In-Reply-To: <492CB588.8070701@scheie.homedns.org> Message-ID: <492cceba.1d1e640a.67fd.3bbd@mx.google.com> > Hmm. I've got it working on my F9/K12Linux box. I wrote > some notes to myself a few years ago about setting it up. > You can find them at http://petre.homedns.org/unix/vnc_via_xinetd.html > > HTH > > Peter > > Odin N?sen wrote: > > I can't get vncts to work in fedora 9 (or fedora 10). I > only get the > > grey screen when I try to "vncviewer mymachine:5900" (or > :5903). Any hints? > > > > I can't find a Xaccess-file to edit - which is what I'm supposed to > > according to the vncts-file in xinetd.d. I've enabled XDMCP in > > /etc/gdm/custom.conf, restarted GDM, but nothing works. I > always get the grey screen. > > > > Has anyone got it working? Getting the same result here. Fresh install of F9 and LTSP v5. Installed via 'yum install vnc-server*'. After installation, tried the instructions from here: http://www.ltsp.org/twiki/bin/view/Ltsp/VNC This will 'seem' like it's connecting, but then never does. /var/log/messages shows a connection then a disconnection: Nov 25 23:17:12 ltsp1 xinetd[3413]: START: xvnc1280 pid=3960 from=::ffff:192.168.3.105 Nov 25 23:17:17 ltsp1 xinetd[3413]: EXIT: xvnc1280 status=1 pid=3960 duration=5(sec) Followed Peter's instructions above, installed via 'yum install vnc-ltsp*', and after installation get the grey screen and mouse pointer only. Again /var/log/messages shows a connection and disconnection: Nov 25 23:15:46 ltsp1 xinetd[3413]: START: vnc-1024x768x16 pid=3956 from=::ffff:192.168.3.105 Nov 25 23:16:04 ltsp1 xinetd[3413]: EXIT: vnc-1024x768x16 status=0 pid=3956 duration=18(sec) Followed instructions here: http://www.g-loaded.eu/2005/11/10/configure-vnc-server-in-fedora/ Set up a user on their own display port, and this actually works perfectly. From odin at gnuskole.no Wed Nov 26 11:34:53 2008 From: odin at gnuskole.no (=?ISO-8859-1?Q?Odin_N=F8sen?=) Date: Wed, 26 Nov 2008 12:34:53 +0100 Subject: vncts doesn't work In-Reply-To: <492CB588.8070701@scheie.homedns.org> References: <20081125091550.M53161@gnuskole.no> <492CB588.8070701@scheie.homedns.org> Message-ID: <20081126112233.M72425@gnuskole.no> > Hmm. I've got it working on my F9/K12Linux box. I wrote some notes to myself > a few years ago about setting it up. You can find them at http://petre.homedns.org/unix/vnc_via_xinetd.html I read through your webpage and though "I have done all that several times!". But I sat down did it all over again - and suddentlig - it did work. I must have forgotten something on the way (all those times)... Odin, feelin' a bit stupid (like a real newbie)