From sylvain.desbureaux at orange-ftgroup.com Mon Aug 3 06:47:31 2009 From: sylvain.desbureaux at orange-ftgroup.com (sylvain.desbureaux at orange-ftgroup.com) Date: Mon, 3 Aug 2009 08:47:31 +0200 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: <4A731733.2030609@redfish-group.com> References: <4A7314ED.9050709@redfish-group.com> <4A731733.2030609@redfish-group.com> Message-ID: I also think this problem was a qpid problem. I still have issues with node authentification though -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: Justin Clacherty [mailto:justin at redfish-group.com] Envoy??: vendredi 31 juillet 2009 18:09 ??: DESBUREAUX Sylvain RD-BIZZ-LAN Cc?: ovirt-devel at redhat.com Objet?: Re: [Ovirt-devel] problem with kerberos (I think) Justin Clacherty wrote: > sylvain.desbureaux at orange-ftgroup.com wrote: >> I solved db-omatic crashed with installing ovirt in a __non__ up to >> date F10. But I still have this when I launch libvirt-qpid from the >> physical node: >> 2009-07-31 10:09:32 warning Closing connection due to internal-error: >> Sasl error: SASL(-1): generic failure: GSSAPI Error: Unspecified GSS >> failure. Minor code may provide more information (No credentials >> cache found) (qpid/client/SaslFactory.cpp:226) >> > > I have a problem where an up to date F10 fails to work, looks like > some updates have killed ovirt. Still trying to figure out the actual > problem so any updates on your problem are greatly appreciated. Looks like my problem is the same... I had a running system which failed after running a yum update. From memory there were qpid and qmf updates. Taskomatic seems to stay up once restarted though. [root at ovirt ovirt-server]# /usr/share/ovirt-server/db-omatic/db_omatic.rb -n INFO Sat Aug 01 02:04:29 +1000 2009 (12122) dbomatic started. INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Marking host node3.redfish.local unavailable /usr/share/ovirt-server/db-omatic/db_omatic.rb:252:in `object_props': undefined method `[]' for # (NoMethodError) from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:608:in `handle_content_ind' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1728:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `loop' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:31:in `to_proc' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `call' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `listen' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:87:in `loop' ... 12 levels... from /usr/share/ovirt-server/db-omatic/db_omatic.rb:95:in `initialize' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `new' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `main' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:482 ERROR Sat Aug 01 02:04:31 +1000 2009 (12122) Error in db-omatic: # ERROR Sat Aug 01 02:04:34 +1000 2009 (12122) /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:92:in `sleep'/usr/share/ovirt-server/db-omatic/db_omatic.rb:441:in `check_heartbeats'/usr/share/ovirt-server/db-omatic/db_omatic.rb:478:in `main'/usr/share/ovirt-server/db-omatic/db_omatic.rb:482 [root at ovirt ovirt-server]# /usr/share/ovirt-server/task-omatic/taskomatic.rb -n -d INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Cannot implement tasks, not connected to broker. Sleeping. INFO Sat Aug 01 02:04:50 +1000 2009 (12125) Reconnected, resuming task checking.. From sylvain.desbureaux at orange-ftgroup.com Mon Aug 3 08:44:58 2009 From: sylvain.desbureaux at orange-ftgroup.com (sylvain.desbureaux at orange-ftgroup.com) Date: Mon, 3 Aug 2009 10:44:58 +0200 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: References: <4A7314ED.9050709@redfish-group.com><4A731733.2030609@redfish-group.com> Message-ID: Hi again, I have two questions: 1. In order to make my node working (the node where I've installed ovirt appliance), I need to regularly do a kinit. Is it normal? 2. Other question, When I try to add a storage pool, I have the following error: INFO Mon Aug 03 10:43:27 +0200 2009 (7407) Connecting to amqp://management.open.net:5672 INFO Mon Aug 03 10:43:27 +0200 2009 (7407) Cannot implement tasks, not connected to broker. Sleeping. INFO Mon Aug 03 10:43:33 +0200 2009 (7407) Reconnected, resuming task checking.. INFO Mon Aug 03 10:44:06 +0200 2009 (7407) starting task_refresh_pool INFO Mon Aug 03 10:44:06 +0200 2009 (7407) storage_find_suitable_host: host physical.open.net uuid 44454C4C-3900-1046-8057-B3C04F57334A INFO Mon Aug 03 10:44:06 +0200 2009 (7407) host.is_disabled is 0 ERROR Mon Aug 03 10:44:06 +0200 2009 (7407) Task action processing failed: RuntimeError: Could not find a host within this storage pool to scan the storage server. ERROR Mon Aug 03 10:44:06 +0200 2009 (7407) /usr/share/ovirt-server/task-omatic/taskomatic.rb:587:in `storage_find_suitable_host'/usr/share/ovirt-server/task-omatic/taskomatic.rb:621:in `task_refresh_pool'/usr/share/ovirt-server/task-omatic/taskomatic.rb:906:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:880:in `each'/usr/share/ovirt-server/task-omatic/taskomatic.rb:880:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:847:in `loop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:847:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:947 INFO Mon Aug 03 10:44:06 +0200 2009 (7407) done I took a look at the code and the problem seems to be here (in taskomatic.rb, line 582): 573 def storage_find_suitable_host(hardware_pool) 574 # find all of the hosts in the same pool as the storage 575 hardware_pool.hosts.each do |host| 576 @logger.info "storage_find_suitable_host: host #{host.hostname} uuid #{host.uuid}" 577 @logger.info "host.is_disabled is #{host.is_disabled}" 578 if host.is_disabled.to_i != 0 579 @logger.info "host #{host.hostname} is disabled" 580 next 581 end --> 582 node = @session.object(:class => 'node', 'hostname' => host.hostname) 583 @logger.info "node: #{node.inspect}" if node 584 return node if node 585 end But when I do a qmf-libvirt-example, I have the host and it seems OK: [root at management ~]# ruby /usr/share/ovirt-server/qmf-libvirt-example.rb Connecting to amqp://management.open.net:5672.. node: physical.open.net property: hostname, physical.open.net property: uri, qemu:///system property: libvirtVersion, 0.6.3 property: apiVersion, 0.6.3 property: hypervisorVersion, 0.9.1 property: hypervisorType, QEMU property: model, x86_64 property: memory, 1993688 property: cpus, 2 property: mhz, 1000 property: nodes, 1 property: sockets, 1 property: cores, 2 property: threads, 1 getXMLDesc() status: 0 getXMLDesc() status: OK xml length: 1199 ... Any ideas? -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: ovirt-devel-bounces at redhat.com [mailto:ovirt-devel-bounces at redhat.com] De la part de sylvain.desbureaux at orange-ftgroup.com Envoy??: lundi 3 ao?t 2009 08:48 ??: justin at redfish-group.com Cc?: ovirt-devel at redhat.com Objet?: RE: [Ovirt-devel] problem with kerberos (I think) I also think this problem was a qpid problem. I still have issues with node authentification though -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: Justin Clacherty [mailto:justin at redfish-group.com] Envoy??: vendredi 31 juillet 2009 18:09 ??: DESBUREAUX Sylvain RD-BIZZ-LAN Cc?: ovirt-devel at redhat.com Objet?: Re: [Ovirt-devel] problem with kerberos (I think) Justin Clacherty wrote: > sylvain.desbureaux at orange-ftgroup.com wrote: >> I solved db-omatic crashed with installing ovirt in a __non__ up to >> date F10. But I still have this when I launch libvirt-qpid from the >> physical node: >> 2009-07-31 10:09:32 warning Closing connection due to internal-error: >> Sasl error: SASL(-1): generic failure: GSSAPI Error: Unspecified GSS >> failure. Minor code may provide more information (No credentials >> cache found) (qpid/client/SaslFactory.cpp:226) >> > > I have a problem where an up to date F10 fails to work, looks like > some updates have killed ovirt. Still trying to figure out the actual > problem so any updates on your problem are greatly appreciated. Looks like my problem is the same... I had a running system which failed after running a yum update. From memory there were qpid and qmf updates. Taskomatic seems to stay up once restarted though. [root at ovirt ovirt-server]# /usr/share/ovirt-server/db-omatic/db_omatic.rb -n INFO Sat Aug 01 02:04:29 +1000 2009 (12122) dbomatic started. INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Marking host node3.redfish.local unavailable /usr/share/ovirt-server/db-omatic/db_omatic.rb:252:in `object_props': undefined method `[]' for # (NoMethodError) from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:608:in `handle_content_ind' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1728:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `loop' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:31:in `to_proc' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `call' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `listen' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:87:in `loop' ... 12 levels... from /usr/share/ovirt-server/db-omatic/db_omatic.rb:95:in `initialize' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `new' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `main' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:482 ERROR Sat Aug 01 02:04:31 +1000 2009 (12122) Error in db-omatic: # ERROR Sat Aug 01 02:04:34 +1000 2009 (12122) /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:92:in `sleep'/usr/share/ovirt-server/db-omatic/db_omatic.rb:441:in `check_heartbeats'/usr/share/ovirt-server/db-omatic/db_omatic.rb:478:in `main'/usr/share/ovirt-server/db-omatic/db_omatic.rb:482 [root at ovirt ovirt-server]# /usr/share/ovirt-server/task-omatic/taskomatic.rb -n -d INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Cannot implement tasks, not connected to broker. Sleeping. INFO Sat Aug 01 02:04:50 +1000 2009 (12125) Reconnected, resuming task checking.. _______________________________________________ Ovirt-devel mailing list Ovirt-devel at redhat.com https://www.redhat.com/mailman/listinfo/ovirt-devel From sylvain.desbureaux at orange-ftgroup.com Mon Aug 3 12:30:43 2009 From: sylvain.desbureaux at orange-ftgroup.com (sylvain.desbureaux at orange-ftgroup.com) Date: Mon, 3 Aug 2009 14:30:43 +0200 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: References: <4A7314ED.9050709@redfish-group.com><4A731733.2030609@redfish-group.com> Message-ID: For question 2, It appears to be a problem with the "object" function which doesn't work. I've found a workaround (quite ugly) in order to make it work. I replace line 582 and 583 with: node_found=nil nodes = @session.objects(:class => "node") nodes.each do |node| @logger.info "node: #{node.hostname}" if node.hostname==host.hostname node_found=node @logger.info "node found!" break end end @logger.info "node: #{node_found.hostname}" unless node_found.nil? return node_found if node_found See https://bugzilla.redhat.com/show_bug.cgi?id=515217 PS: I'm not saying we have to include but I'm just showing a way to make it work for people who may encounter the same issue. -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: ovirt-devel-bounces at redhat.com [mailto:ovirt-devel-bounces at redhat.com] De la part de sylvain.desbureaux at orange-ftgroup.com Envoy??: lundi 3 ao?t 2009 10:45 ??: ovirt-devel at redhat.com Objet?: RE: [Ovirt-devel] problem with kerberos (I think) Hi again, I have two questions: 1. In order to make my node working (the node where I've installed ovirt appliance), I need to regularly do a kinit. Is it normal? 2. Other question, When I try to add a storage pool, I have the following error: INFO Mon Aug 03 10:43:27 +0200 2009 (7407) Connecting to amqp://management.open.net:5672 INFO Mon Aug 03 10:43:27 +0200 2009 (7407) Cannot implement tasks, not connected to broker. Sleeping. INFO Mon Aug 03 10:43:33 +0200 2009 (7407) Reconnected, resuming task checking.. INFO Mon Aug 03 10:44:06 +0200 2009 (7407) starting task_refresh_pool INFO Mon Aug 03 10:44:06 +0200 2009 (7407) storage_find_suitable_host: host physical.open.net uuid 44454C4C-3900-1046-8057-B3C04F57334A INFO Mon Aug 03 10:44:06 +0200 2009 (7407) host.is_disabled is 0 ERROR Mon Aug 03 10:44:06 +0200 2009 (7407) Task action processing failed: RuntimeError: Could not find a host within this storage pool to scan the storage server. ERROR Mon Aug 03 10:44:06 +0200 2009 (7407) /usr/share/ovirt-server/task-omatic/taskomatic.rb:587:in `storage_find_suitable_host'/usr/share/ovirt-server/task-omatic/taskomatic.rb:621:in `task_refresh_pool'/usr/share/ovirt-server/task-omatic/taskomatic.rb:906:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:880:in `each'/usr/share/ovirt-server/task-omatic/taskomatic.rb:880:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:847:in `loop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:847:in `mainloop'/usr/share/ovirt-server/task-omatic/taskomatic.rb:947 INFO Mon Aug 03 10:44:06 +0200 2009 (7407) done I took a look at the code and the problem seems to be here (in taskomatic.rb, line 582): 573 def storage_find_suitable_host(hardware_pool) 574 # find all of the hosts in the same pool as the storage 575 hardware_pool.hosts.each do |host| 576 @logger.info "storage_find_suitable_host: host #{host.hostname} uuid #{host.uuid}" 577 @logger.info "host.is_disabled is #{host.is_disabled}" 578 if host.is_disabled.to_i != 0 579 @logger.info "host #{host.hostname} is disabled" 580 next 581 end --> 582 node = @session.object(:class => 'node', 'hostname' => host.hostname) 583 @logger.info "node: #{node.inspect}" if node 584 return node if node 585 end But when I do a qmf-libvirt-example, I have the host and it seems OK: [root at management ~]# ruby /usr/share/ovirt-server/qmf-libvirt-example.rb Connecting to amqp://management.open.net:5672.. node: physical.open.net property: hostname, physical.open.net property: uri, qemu:///system property: libvirtVersion, 0.6.3 property: apiVersion, 0.6.3 property: hypervisorVersion, 0.9.1 property: hypervisorType, QEMU property: model, x86_64 property: memory, 1993688 property: cpus, 2 property: mhz, 1000 property: nodes, 1 property: sockets, 1 property: cores, 2 property: threads, 1 getXMLDesc() status: 0 getXMLDesc() status: OK xml length: 1199 ... Any ideas? -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: ovirt-devel-bounces at redhat.com [mailto:ovirt-devel-bounces at redhat.com] De la part de sylvain.desbureaux at orange-ftgroup.com Envoy??: lundi 3 ao?t 2009 08:48 ??: justin at redfish-group.com Cc?: ovirt-devel at redhat.com Objet?: RE: [Ovirt-devel] problem with kerberos (I think) I also think this problem was a qpid problem. I still have issues with node authentification though -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: Justin Clacherty [mailto:justin at redfish-group.com] Envoy??: vendredi 31 juillet 2009 18:09 ??: DESBUREAUX Sylvain RD-BIZZ-LAN Cc?: ovirt-devel at redhat.com Objet?: Re: [Ovirt-devel] problem with kerberos (I think) Justin Clacherty wrote: > sylvain.desbureaux at orange-ftgroup.com wrote: >> I solved db-omatic crashed with installing ovirt in a __non__ up to >> date F10. But I still have this when I launch libvirt-qpid from the >> physical node: >> 2009-07-31 10:09:32 warning Closing connection due to internal-error: >> Sasl error: SASL(-1): generic failure: GSSAPI Error: Unspecified GSS >> failure. Minor code may provide more information (No credentials >> cache found) (qpid/client/SaslFactory.cpp:226) >> > > I have a problem where an up to date F10 fails to work, looks like > some updates have killed ovirt. Still trying to figure out the actual > problem so any updates on your problem are greatly appreciated. Looks like my problem is the same... I had a running system which failed after running a yum update. From memory there were qpid and qmf updates. Taskomatic seems to stay up once restarted though. [root at ovirt ovirt-server]# /usr/share/ovirt-server/db-omatic/db_omatic.rb -n INFO Sat Aug 01 02:04:29 +1000 2009 (12122) dbomatic started. INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:29 +1000 2009 (12122) Marking host node3.redfish.local unavailable /usr/share/ovirt-server/db-omatic/db_omatic.rb:252:in `object_props': undefined method `[]' for # (NoMethodError) from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:608:in `handle_content_ind' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1728:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `loop' from /usr/lib/ruby/site_ruby/1.8/qpid/qmf.rb:1716:in `reply_cb' from /usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:31:in `to_proc' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `call' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:91:in `listen' from /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:87:in `loop' ... 12 levels... from /usr/share/ovirt-server/db-omatic/db_omatic.rb:95:in `initialize' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `new' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:475:in `main' from /usr/share/ovirt-server/db-omatic/db_omatic.rb:482 ERROR Sat Aug 01 02:04:31 +1000 2009 (12122) Error in db-omatic: # ERROR Sat Aug 01 02:04:34 +1000 2009 (12122) /usr/lib/ruby/site_ruby/1.8/qpid/queue.rb:92:in `sleep'/usr/share/ovirt-server/db-omatic/db_omatic.rb:441:in `check_heartbeats'/usr/share/ovirt-server/db-omatic/db_omatic.rb:478:in `main'/usr/share/ovirt-server/db-omatic/db_omatic.rb:482 [root at ovirt ovirt-server]# /usr/share/ovirt-server/task-omatic/taskomatic.rb -n -d INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Connecting to amqp://ovirt.redfish.local:5672 INFO Sat Aug 01 02:04:44 +1000 2009 (12125) Cannot implement tasks, not connected to broker. Sleeping. INFO Sat Aug 01 02:04:50 +1000 2009 (12125) Reconnected, resuming task checking.. _______________________________________________ Ovirt-devel mailing list Ovirt-devel at redhat.com https://www.redhat.com/mailman/listinfo/ovirt-devel _______________________________________________ Ovirt-devel mailing list Ovirt-devel at redhat.com https://www.redhat.com/mailman/listinfo/ovirt-devel From jboggs at redhat.com Mon Aug 3 14:31:11 2009 From: jboggs at redhat.com (Joey Boggs) Date: Mon, 03 Aug 2009 10:31:11 -0400 Subject: [Ovirt-devel] [PATCH node] update method used to determine boot partition for install In-Reply-To: <4A734C8D.60102@redhat.com> References: <1249053120-4083-1-git-send-email-jboggs@redhat.com> <2be7262f0907311249h4b01dcceu4ec628e0e39273d5@mail.gmail.com> <4A734C8D.60102@redhat.com> Message-ID: <4A76F4AF.50602@redhat.com> Joey Boggs wrote: > Alan Pevec wrote: >> On Fri, Jul 31, 2009 at 5:12 PM, Joey Boggs > > wrote: >> >> This will facilitate the switch to readonly rootfs. The current >> method will not work when symlinking /etc/mtab to /proc/mounts as >> the udev label /BOOT is used rather than /dev/sda1 or similar. The >> same output is generated and has been verified to install correctly. >> >> >> We need this symlink even if /etc/mtab is listed in /etc/rwtab ? > There's some issues with the rwtab in this scenario, for mtab to be in > there the real problem is the lock file created as > /etc/mtab~$randomnumber can't be written in /etc. there are a few > bugzillas concerning this but not likely to change anything until > Fedora 12+ timeframe since it requires updating alot of different > programs mount.* and so forth. If you just add /etc/ then you have > multiple overlaps and mounts including the items in /config/files and > those in /etc/rwtab. /etc is then essentially handled as a separate > partition which is not the most ideal situation and will cause the > machine to hang up on boot since it can't unmount /etc and requires a > manual reboot. Been working on a few scenarios for awhile :). I can > edit the eval statement and make sure that /boot is mounted as well. > >> >> # check that /boot mounted ok and find partition number for GRUB >> - eval $(mount|awk '$3 == "/boot" { >> >> + eval $(readlink -f /dev/disk/by-label/BOOT|awk {' >> >> >> This wouldn't check that /boot actually mounted successfully. >> >> - print "disk=" substr($1,1,length($1)-1); >> - print "disk2=" substr($1,1,length($1)-2); >> - partN=substr($1,length($1),1); partN--; >> - print "partN=" partN; >> - }') >> + print "disk=" substr($1,1,length($1)-1); >> + print "disk2=" substr($1,1,length($1)-2); >> + partN=substr($1,length($1),1); partN--; >> + print "partN=" partN; >> + '}) >> >> This whitespace only change obscures the real change, please put in a >> separate patch, but IMHO fixing indentation is not need here. >> > > _______________________________________________ > Ovirt-devel mailing list > Ovirt-devel at redhat.com > https://www.redhat.com/mailman/listinfo/ovirt-devel Updated version, which checks the /boot mountpoint. Let me know if the my mail client mangles this patch, didn't want to fight git-send-email on replying. -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: 0001-update-method-used-to-determine-bootpartition-for-in.patch URL: From arroy at redhat.com Mon Aug 3 14:35:54 2009 From: arroy at redhat.com (Arjun Roy) Date: Mon, 03 Aug 2009 10:35:54 -0400 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> Message-ID: <4A76F5CA.2020207@redhat.com> On 07/31/2009 04:18 AM, Justin Clacherty wrote: > I'm still having issues with this problem and things are starting to get quite frustrating. I've spent the last two days installing and reinstalling the management server (even tried getting next to work). I've currently got a clean install of F10 in a vm and a fresh install of ovirt from the ovirt repository. This installation is experiencing the exact same problem that caused me to try a reinstall, it's as if one of the updates to fedora has killed it (I think there was a qmf and a qpid update). > > Here's what happens: > - management server is running > - node boots > - node details appear in /var/log/ovirt-server/host-browser.log > - node appears in GUI but appears as "unavailable (enabled)" > - running "ruby /usr/share/ovirt-server/qmf-libvirt-example.rb" yeilds no results > > Has anyone else seen this? I would appreciate it if someone could help me to debug the problem? > > I've put the output of the ruby script, taskomatic.log, and host-browser.log below in case it's of use. > The rpm packages on the ovirt site right now I believe are somewhat out of date. You would need to build the latest version (on the git 'next' tree) on Fedora 11 for things to work (I notice that your host-browser log for example shows that you are using an older revision of ovirt). Note that building on Fedora 10 will not work, since a couple of APIs that ovirt uses changed between 10 and 11. -Arjun From jguiditt at redhat.com Mon Aug 3 14:43:44 2009 From: jguiditt at redhat.com (Jason Guiditta) Date: Mon, 03 Aug 2009 10:43:44 -0400 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: References: <4A7314ED.9050709@redfish-group.com><4A731733.2030609@redfish-group.com> Message-ID: <1249310624.3466.9.camel@lenovo> On Mon, 2009-08-03 at 14:30 +0200, sylvain.desbureaux at orange-ftgroup.com wrote: > For question 2, > It appears to be a problem with the "object" function which doesn't work. > I've found a workaround (quite ugly) in order to make it work. > > I replace line 582 and 583 with: > node_found=nil > nodes = @session.objects(:class => "node") > nodes.each do |node| > @logger.info "node: #{node.hostname}" > if node.hostname==host.hostname > node_found=node > @logger.info "node found!" > break > end > end > @logger.info "node: #{node_found.hostname}" unless node_found.nil? > return node_found if node_found > > > See https://bugzilla.redhat.com/show_bug.cgi?id=515217 > > > PS: I'm not saying we have to include but I'm just showing a way to make it work for people who may encounter the same issue. > > -- > Sylvain Desbureaux > +33 296 051 380 > > Sylvain, sorry for the delayed response, we have been furiously trying to get patches in for the next release. The main issue is that the code you are running is from an old release, and I think some of our deps in F10 have broken ovirt (mostly qpid stuff I think). Because of this, ovirt currently only runs on F11 and rawhide. My suggestion would be either to build from next or try the new f11 release when it is out (never been closer). From sylvain.desbureaux at orange-ftgroup.com Mon Aug 3 15:49:48 2009 From: sylvain.desbureaux at orange-ftgroup.com (sylvain.desbureaux at orange-ftgroup.com) Date: Mon, 3 Aug 2009 17:49:48 +0200 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: <1249310624.3466.9.camel@lenovo> References: <4A7314ED.9050709@redfish-group.com><4A731733.2030609@redfish-group.com> <1249310624.3466.9.camel@lenovo> Message-ID: No problem for delayed response :-) If I want to install it on F11, do you think it's better to install from rpms or to build it? Thanks, -- Sylvain Desbureaux +33 296 051 380 -----Message d'origine----- De?: Jason Guiditta [mailto:jguiditt at redhat.com] Envoy??: lundi 3 ao?t 2009 16:44 ??: DESBUREAUX Sylvain RD-BIZZ-LAN Cc?: ovirt-devel at redhat.com Objet?: RE: [Ovirt-devel] problem with kerberos (I think) On Mon, 2009-08-03 at 14:30 +0200, sylvain.desbureaux at orange-ftgroup.com wrote: > For question 2, > It appears to be a problem with the "object" function which doesn't work. > I've found a workaround (quite ugly) in order to make it work. > > I replace line 582 and 583 with: > node_found=nil > nodes = @session.objects(:class => "node") > nodes.each do |node| > @logger.info "node: #{node.hostname}" > if node.hostname==host.hostname > node_found=node > @logger.info "node found!" > break > end > end > @logger.info "node: #{node_found.hostname}" unless node_found.nil? > return node_found if node_found > > > See https://bugzilla.redhat.com/show_bug.cgi?id=515217 > > > PS: I'm not saying we have to include but I'm just showing a way to make it work for people who may encounter the same issue. > > -- > Sylvain Desbureaux > +33 296 051 380 > > Sylvain, sorry for the delayed response, we have been furiously trying to get patches in for the next release. The main issue is that the code you are running is from an old release, and I think some of our deps in F10 have broken ovirt (mostly qpid stuff I think). Because of this, ovirt currently only runs on F11 and rawhide. My suggestion would be either to build from next or try the new f11 release when it is out (never been closer). From arroy at redhat.com Mon Aug 3 16:32:07 2009 From: arroy at redhat.com (Arjun Roy) Date: Mon, 03 Aug 2009 12:32:07 -0400 Subject: [Ovirt-devel] problem with kerberos (I think) In-Reply-To: References: <4A7314ED.9050709@redfish-group.com><4A731733.2030609@redfish-group.com> <1249310624.3466.9.camel@lenovo> Message-ID: <4A771107.3000803@redhat.com> On 08/03/2009 11:49 AM, sylvain.desbureaux at orange-ftgroup.com wrote: > No problem for delayed response :-) > If I want to install it on F11, do you think it's better to install from rpms or to build it? > Thanks, > > I would build it using these instructions: http://ovirt.org/page/Devel_Quickstart -Arjun From apevec at redhat.com Mon Aug 3 20:29:30 2009 From: apevec at redhat.com (Alan Pevec) Date: Mon, 3 Aug 2009 22:29:30 +0200 Subject: [Ovirt-devel] [PATCH node-image] leave ssh open for devel builds Message-ID: <1249331370-14199-1-git-send-email-apevec@redhat.com> devel builds have empty root password Bug 513633 - Provide a way of setting the root password for oVirt Node --- common-post.ks | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/common-post.ks b/common-post.ks index dcd44d6..0c0a3f4 100644 --- a/common-post.ks +++ b/common-post.ks @@ -115,6 +115,10 @@ sed -i '/rm -f \/etc\/yp.conf/d' $DHSCRIPT if rpm -q --qf '%{release}' ovirt-node | grep -q "^0\." ; then echo "Building in developer mode, leaving root account unlocked" + augtool <<\EOF +set /files/etc/ssh/sshd_config/PermitEmptyPasswords yes +save +EOF else echo "Building in production mode, locking root account" passwd -l root -- 1.6.0.6 From apevec at gmail.com Mon Aug 3 20:41:09 2009 From: apevec at gmail.com (Alan Pevec) Date: Mon, 3 Aug 2009 22:41:09 +0200 Subject: [Ovirt-devel] [PATCH node] update method used to determine boot partition for install In-Reply-To: <4A76F4AF.50602@redhat.com> References: <1249053120-4083-1-git-send-email-jboggs@redhat.com> <2be7262f0907311249h4b01dcceu4ec628e0e39273d5@mail.gmail.com> <4A734C8D.60102@redhat.com> <4A76F4AF.50602@redhat.com> Message-ID: <2be7262f0908031341x61dea813ld0362898e9789c54@mail.gmail.com> ACK -------------- next part -------------- An HTML attachment was scrubbed... URL: From jboggs at redhat.com Mon Aug 3 21:02:12 2009 From: jboggs at redhat.com (Joey Boggs) Date: Mon, 03 Aug 2009 17:02:12 -0400 Subject: [Ovirt-devel] [PATCH node] update method used to determine boot partition for install In-Reply-To: <4A76F4AF.50602@redhat.com> References: <1249053120-4083-1-git-send-email-jboggs@redhat.com> <2be7262f0907311249h4b01dcceu4ec628e0e39273d5@mail.gmail.com> <4A734C8D.60102@redhat.com> <4A76F4AF.50602@redhat.com> Message-ID: <4A775054.7080402@redhat.com> Joey Boggs wrote: > Joey Boggs wrote: >> Alan Pevec wrote: >>> On Fri, Jul 31, 2009 at 5:12 PM, Joey Boggs >> > wrote: >>> >>> This will facilitate the switch to readonly rootfs. The current >>> method will not work when symlinking /etc/mtab to /proc/mounts as >>> the udev label /BOOT is used rather than /dev/sda1 or similar. The >>> same output is generated and has been verified to install >>> correctly. >>> >>> >>> We need this symlink even if /etc/mtab is listed in /etc/rwtab ? >> There's some issues with the rwtab in this scenario, for mtab to be >> in there the real problem is the lock file created as >> /etc/mtab~$randomnumber can't be written in /etc. there are a few >> bugzillas concerning this but not likely to change anything until >> Fedora 12+ timeframe since it requires updating alot of different >> programs mount.* and so forth. If you just add /etc/ then you have >> multiple overlaps and mounts including the items in /config/files and >> those in /etc/rwtab. /etc is then essentially handled as a separate >> partition which is not the most ideal situation and will cause the >> machine to hang up on boot since it can't unmount /etc and requires a >> manual reboot. Been working on a few scenarios for awhile :). I can >> edit the eval statement and make sure that /boot is mounted as well. >> >>> >>> # check that /boot mounted ok and find partition number for >>> GRUB >>> - eval $(mount|awk '$3 == "/boot" { >>> >>> + eval $(readlink -f /dev/disk/by-label/BOOT|awk {' >>> >>> >>> This wouldn't check that /boot actually mounted successfully. >>> >>> - print "disk=" substr($1,1,length($1)-1); >>> - print "disk2=" substr($1,1,length($1)-2); >>> - partN=substr($1,length($1),1); partN--; >>> - print "partN=" partN; >>> - }') >>> + print "disk=" substr($1,1,length($1)-1); >>> + print "disk2=" substr($1,1,length($1)-2); >>> + partN=substr($1,length($1),1); partN--; >>> + print "partN=" partN; >>> + '}) >>> >>> This whitespace only change obscures the real change, please put in >>> a separate patch, but IMHO fixing indentation is not need here. >>> >> >> _______________________________________________ >> Ovirt-devel mailing list >> Ovirt-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/ovirt-devel > > > Updated version, which checks the /boot mountpoint. Let me know if the > my mail client mangles this patch, didn't want to fight git-send-email > on replying. > > ------------------------------------------------------------------------ > > _______________________________________________ > Ovirt-devel mailing list > Ovirt-devel at redhat.com > https://www.redhat.com/mailman/listinfo/ovirt-devel pushed, creating bz for downstream From mmorsi at redhat.com Mon Aug 3 21:37:58 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Mon, 3 Aug 2009 17:37:58 -0400 Subject: [Ovirt-devel] [PATCH node] changed fickle anyterm sysconfig changes to something more rebust Message-ID: <1249335478-23888-1-git-send-email-mmorsi@redhat.com> instead of sed'ing the exact commented line we are looking for, simply add the necessary lines to the end of the anyterm sysconf --- scripts/ovirt-functions | 9 ++++----- 1 files changed, 4 insertions(+), 5 deletions(-) diff --git a/scripts/ovirt-functions b/scripts/ovirt-functions index af3a257..1c7b0c4 100644 --- a/scripts/ovirt-functions +++ b/scripts/ovirt-functions @@ -213,11 +213,10 @@ ovirt_setup_libvirtd() { ovirt_setup_anyterm() { # configure anyterm - sed -i -e 's/^# ANYTERM_CMD=\/usr\/bin\/anyterm-cmd/ANYTERM_CMD="sudo \/usr\/bin\/virsh console %p"/' \ - /etc/sysconfig/anyterm - - sed -i -e 's/^# ANYTERM_LOCAL_ONLY=true/ANYTERM_LOCAL_ONLY=false/' \ - /etc/sysconfig/anyterm + cat >> /etc/sysconfig/anyterm << EOF + ANYTERM_CMD="sudo /usr/bin/virsh console %p" + ANYTERM_LOCAL_ONLY=false + EOF # permit it to run the virsh console echo "anyterm ALL=NOPASSWD: /usr/bin/virsh console *" >> /etc/sudoers -- 1.6.0.6 From justin at redfish-group.com Mon Aug 3 21:39:57 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 07:39:57 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A76F5CA.2020207@redhat.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> Message-ID: <4A77592D.20202@redfish-group.com> Arjun Roy wrote: > The rpm packages on the ovirt site right now I believe are somewhat > out of date. You would need to build the latest version (on the git > 'next' tree) on Fedora 11 for things to work (I notice that your > host-browser log for example shows that you are using an older > revision of ovirt). Yes, I'm using the current release version and installing on Fedora 10. I tried building both the release and next (on Fedora 10) but had no luck with either. The master tree built fine but failed to install properly when I ran ace. The next tree would not build as it complained about unresolved dependencies when downloading rpms (presumabely to make the node image). > Note that building on Fedora 10 will not work, since a couple of APIs > that ovirt uses changed between 10 and 11. Ah, that might explain it. I guess today's job is to reinstall using Fedora 11. Justin. From pmyers at redhat.com Mon Aug 3 21:51:41 2009 From: pmyers at redhat.com (Perry Myers) Date: Mon, 3 Aug 2009 17:51:41 -0400 Subject: [Ovirt-devel] [PATCH node-image] Fix edit-livecd to re-create ext2 filesystem each time Message-ID: <1249336301-28967-1-git-send-email-pmyers@redhat.com> Old edit-livecd just kept adding to the ext2 fs image which meant it never decreased in size even if files were removed. New functionality does multiple passes of ext2fs creation. First pass creates a new ext2fs from the image file inside the squashfs. This first pass is to allow the user to optionally increase the filesystem size. If the -s parameter is omitted the existing size is used and this step is a nop. The new ext2 fs is mounted to the ex-rw directory. After ex-rw is modified by the user (either manually or through supplied script) it is copied again to ex-min. This step is done so that the filesystem is minimized in size (i.e. file removals will actually reduce the size of the ext2 image and therefore reduce the overall size of the livecd) NOTE: There seems to be an approximate 1MB increase in size even on a nop edit-livecd. This is not in the ext2 filesystem itself, it's seen in the squashfs image. Not sure why this happens. Signed-off-by: Perry Myers --- edit-livecd | 123 ++++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 files changed, 110 insertions(+), 13 deletions(-) diff --git a/edit-livecd b/edit-livecd index d69ca9d..e94b1e6 100755 --- a/edit-livecd +++ b/edit-livecd @@ -31,7 +31,7 @@ CD=$NODEIMG_DEFAULT usage() { case $# in 1) warn "$1"; try_h; exit 1;; esac cat < when done, and the script re-packages the ISO. + -s Size of ext2 filesystem in MB (default is existing size) -h display this help and exit EXAMPLES @@ -69,14 +70,16 @@ set -e CODE= OUTPUT_FILE= +SIZE= err=0 help=0 -while getopts :b:hi:o:p: c; do +while getopts :b:hi:o:p:s: c; do case $c in i) CD=$OPTARG;; b) PARAMS=$OPTARG;; o) OUTPUT_FILE=$OPTARG;; p) CODE=$OPTARG;; + s) SIZE=$OPTARG;; h) help=1;; '?') err=1; warn "invalid option: \`-$OPTARG'";; :) err=1; warn "missing argument to \`-$OPTARG' option";; @@ -86,9 +89,9 @@ done test $err = 1 && { try_h; exit 1; } test $help = 1 && { usage; exit 0; } -# Require "-o OUTPUT_FILE" +# Require "-o OUTPUT.iso" test -z "$OUTPUT_FILE" \ - && { warn "no output file specified; use -o FILE.iso"; try_h; exit 1; } + && { warn "no output file specified; use -o OUTPUT.iso"; try_h; exit 1; } # Fail if there are any extra command-line arguments. if test $OPTIND -le $#; then @@ -101,6 +104,8 @@ if [ $( id -u ) -ne 0 ]; then die "Must run as root" fi +test selinuxenabled || { echo "selinux not enabled, aborting" ; exit 1 ; } + # Check for some prerequisites. # "type" prints "PROG not found" if it's not in $PATH. type mkisofs @@ -137,6 +142,71 @@ mnt() { addExit "df | grep $mp > /dev/null 2>&1 && umount -v $mp" } +# Create new ext2 filesystem and copy contents of one into the other +# Size of ext2 filesystem is same as original size unless size parameter +# is passed in +create_ext2fs() { + local input_file=$1 + local input_dir=$2 + local output_file=$3 + local output_dir=$4 + local size_mb=$5 + + local size= + if [[ -n "$size_mb" ]]; then + size=$(( $size_mb * 1024 )) + fi + + echo ">>> Mounting old ext3fs" + mnt "-t ext2 $input_file -o ro,loop" $input_dir + ext_used=$(df $WDIR/$input_dir | tail -1 | awk '{print $3}') + ext_max=$(df $WDIR/$input_dir | tail -1 | awk '{print $2}') + + echo ">>> $input_file max_size=$ext_max current_size=$ext_used" + + local new_size= + if [[ -z "$size" ]]; then + new_size=$ext_max + elif [ $size -lt $ext_used ]; then + echo "New size of $size is less than the current used size $ext_used" + exit 1 + else + new_size=$size + fi + + echo ">>> Size selected for new $output_file: $new_size" + + echo ">>> Creating new ext3fs" + dd if=/dev/zero of=$output_file bs=1024 count=0 seek=${new_size} + mkfs.ext2 -b 4096 -m 1 -L "ovirt-node-image" -F $output_file + tune2fs -i 0 -c 0 -Odir_index -ouser_xattr,acl $output_file + + echo ">>> Mounting new ext3fs" + mnt "-t ext2 $output_file -o rw,loop" $output_dir + + SELINUX_ENFORCING=$(getenforce) + case $SELINUX_ENFORCING in + Enforcing) setenforce Permissive ;; + Permissive) ;; + *) echo "Do not run $0 with selinux in disabled mode" ; exit 1 ;; + esac + + echo ">>> Copying data from old ext3fs to new ext3fs" + rsync -aX $WDIR/$input_dir/ $WDIR/$output_dir/ + + echo -e ">>> $(df $WDIR/$input_dir | tail -1)" + echo -e ">>> $(df $WDIR/$output_dir | tail -1)" + umount $WDIR/$output_dir + umount $WDIR/$input_dir + + if [ "$SELINUX_ENFORCING" = "Enforcing" ]; then + setenforce Enforcing + fi + + e2fsck -f -y $output_file || : + e2fsck -f -y $output_file || : +} + addExit "rm -rf $WDIR" ID_FS_LABEL= # initialize, in case vol_id fails @@ -150,47 +220,73 @@ mnt "-t iso9660 $CD -o loop,ro" cd mnt "-t squashfs $WDIR/cd/LiveOS/squashfs.img -o ro,loop" sq # create writable copy of the new filesystem for the CD -cp -pr $WDIR/cd $WDIR/cd-w +rsync -a $WDIR/cd/ $WDIR/cd-w/ # create writable copy of the filesystem for the new compressed # squashfs filesystem -cp -pr $WDIR/sq $WDIR/sq-w +rsync -a $WDIR/sq/ $WDIR/sq-w/ +echo ">>> Old LiveCD Size: $(du -b $CD)" +echo ">>> Old Squash Size: $(du -b $WDIR/cd/LiveOS/squashfs.img)" +echo ">>> Old ext3fs Size: $(du -b $WDIR/sq-w/LiveOS/ext3fs.img)" -# mount root filesystem -mnt "-t ext2 $WDIR/sq-w/LiveOS/ext3fs.img -o rw,loop" ex +# Create copy of original ext2fs in case user wants to make +# the file system larger than the original +create_ext2fs $WDIR/sq-w/LiveOS/ext3fs.img ex-ro $WDIR/ext3fs-rw.img ex-rw $SIZE + +echo ">>> New ext3fs Size: $(du -b $WDIR/ext3fs-rw.img)" + +echo ">>> Mounting new ext3fs" +mnt "-t ext2 $WDIR/ext3fs-rw.img -o rw,loop" ex-rw echo ">>> Updating CD content" if [ -n "$CODE" ]; then ( - cd $WDIR/ex + cd $WDIR/ex-rw set +e eval "$CODE" set -e ) else echo "***" - echo "*** Pausing to allow manual changes. Press any key to continue." + echo "*** Pausing to allow manual changes." + echo "*** Make changes in $WDIR/ex-rw" + echo "*** Press any key to continue." echo "***" read fi +# cleanup temporary and cache locations in the chroot, rhbz#508043 +rm -rf $WDIR/ex-rw/var/cache/yum +rm -rf $WDIR/ex-rw/tmp +mkdir $WDIR/ex-rw/tmp + # Try to unmount. But this is likely to fail, so let the user retry, -# e.g., if he forgot to "cd" out of $WDIR/ex. +# e.g., if he forgot to "cd" out of $WDIR/ex-rw. while :; do echo ">>> Unmounting ext3fs" - umount $WDIR/ex && break + umount $WDIR/ex-rw && break echo ">>> Unmounting the working file system copy failed" echo "***" - echo "*** Did you forget to 'cd' out of $WDIR/ex?" + echo "*** Did you forget to 'cd' out of $WDIR/ex-rw?" echo "***" echo "*** Press any key to repeat the attempt." echo "***" read done +# Copy the filesystem again to make sure we minimize the image +create_ext2fs $WDIR/ext3fs-rw.img ex-rw $WDIR/ext3fs-min.img ex-min $SIZE + +echo ">>> Min ext3fs Size: $(du -b $WDIR/ext3fs-min.img)" + +# Move the minimized ext2 filesystem into the squash directory +mv -v $WDIR/ext3fs-min.img $WDIR/sq-w/LiveOS/ext3fs.img + echo ">>> Compressing filesystem" mksquashfs $WDIR/sq-w/ $WDIR/cd-w/LiveOS/squashfs.img -noappend +echo ">>> New Squash Size: $(du -b $WDIR/cd-w/LiveOS/squashfs.img)" + echo ">>> Recomputing MD5 sums" ( cd $WDIR/cd-w && find . -type f -not -name md5sum.txt \ -not -path '*/isolinux/*' -print0 | xargs -0 -- md5sum > md5sum.txt ) @@ -215,6 +311,7 @@ mkisofs \ echo ">>> Implanting ISO MD5 Sum" implantisomd5 --force "$OUTPUT_FILE" +echo ">>> New LiveCD Size: $(du -b $OUTPUT_FILE)" # The trap ... callbacks will unmount everything. set +e -- 1.6.2.5 From jguiditt at redhat.com Mon Aug 3 22:08:07 2009 From: jguiditt at redhat.com (Jason Guiditta) Date: Mon, 03 Aug 2009 18:08:07 -0400 Subject: [Ovirt-devel] [PATCH node-image] leave ssh open for devel builds In-Reply-To: <1249331370-14199-1-git-send-email-apevec@redhat.com> References: <1249331370-14199-1-git-send-email-apevec@redhat.com> Message-ID: <1249337287.3466.11.camel@lenovo> On Mon, 2009-08-03 at 22:29 +0200, Alan Pevec wrote: > devel builds have empty root password > Bug 513633 - Provide a way of setting the root password for oVirt Node > --- > common-post.ks | 4 ++++ > 1 files changed, 4 insertions(+), 0 deletions(-) > > diff --git a/common-post.ks b/common-post.ks > index dcd44d6..0c0a3f4 100644 > --- a/common-post.ks > +++ b/common-post.ks > @@ -115,6 +115,10 @@ sed -i '/rm -f \/etc\/yp.conf/d' $DHSCRIPT > > if rpm -q --qf '%{release}' ovirt-node | grep -q "^0\." ; then > echo "Building in developer mode, leaving root account unlocked" > + augtool <<\EOF > +set /files/etc/ssh/sshd_config/PermitEmptyPasswords yes > +save > +EOF > else > echo "Building in production mode, locking root account" > passwd -l root ACK, logged in as root/no pw w/o issue From mmorsi at redhat.com Mon Aug 3 22:18:06 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Mon, 3 Aug 2009 18:18:06 -0400 Subject: [Ovirt-devel] [PATCH node] changed fickle anyterm sysconfig changes to something more robust (revised) Message-ID: <1249337886-25729-1-git-send-email-mmorsi@redhat.com> instead of sed'ing the exact commented line we are looking for, simply add the necessary lines to the end of the anyterm sysconf --- scripts/ovirt-functions | 9 ++++----- 1 files changed, 4 insertions(+), 5 deletions(-) diff --git a/scripts/ovirt-functions b/scripts/ovirt-functions index af3a257..98e88e7 100644 --- a/scripts/ovirt-functions +++ b/scripts/ovirt-functions @@ -213,11 +213,10 @@ ovirt_setup_libvirtd() { ovirt_setup_anyterm() { # configure anyterm - sed -i -e 's/^# ANYTERM_CMD=\/usr\/bin\/anyterm-cmd/ANYTERM_CMD="sudo \/usr\/bin\/virsh console %p"/' \ - /etc/sysconfig/anyterm - - sed -i -e 's/^# ANYTERM_LOCAL_ONLY=true/ANYTERM_LOCAL_ONLY=false/' \ - /etc/sysconfig/anyterm + cat >> /etc/sysconfig/anyterm << EOF +ANYTERM_CMD="sudo /usr/bin/virsh console %p" +ANYTERM_LOCAL_ONLY=false +EOF # permit it to run the virsh console echo "anyterm ALL=NOPASSWD: /usr/bin/virsh console *" >> /etc/sudoers -- 1.6.0.6 From jguiditt at redhat.com Tue Aug 4 01:15:27 2009 From: jguiditt at redhat.com (Jason Guiditta) Date: Mon, 03 Aug 2009 21:15:27 -0400 Subject: [Ovirt-devel] Re: [PATCH node] changed fickle anyterm sysconfig changes to something more robust (revised) In-Reply-To: <1249337886-25729-1-git-send-email-mmorsi@redhat.com> References: <1249337886-25729-1-git-send-email-mmorsi@redhat.com> Message-ID: <1249348527.3466.14.camel@lenovo> On Mon, 2009-08-03 at 18:18 -0400, Mohammed Morsi wrote: > instead of sed'ing the exact commented line we are looking for, simply > add the necessary lines to the end of the anyterm sysconf > --- > scripts/ovirt-functions | 9 ++++----- > 1 files changed, 4 insertions(+), 5 deletions(-) > > diff --git a/scripts/ovirt-functions b/scripts/ovirt-functions > index af3a257..98e88e7 100644 > --- a/scripts/ovirt-functions > +++ b/scripts/ovirt-functions > @@ -213,11 +213,10 @@ ovirt_setup_libvirtd() { > > ovirt_setup_anyterm() { > # configure anyterm > - sed -i -e 's/^# ANYTERM_CMD=\/usr\/bin\/anyterm-cmd/ANYTERM_CMD="sudo \/usr\/bin\/virsh console %p"/' \ > - /etc/sysconfig/anyterm > - > - sed -i -e 's/^# ANYTERM_LOCAL_ONLY=true/ANYTERM_LOCAL_ONLY=false/' \ > - /etc/sysconfig/anyterm > + cat >> /etc/sysconfig/anyterm << EOF > +ANYTERM_CMD="sudo /usr/bin/virsh console %p" > +ANYTERM_LOCAL_ONLY=false > +EOF > > # permit it to run the virsh console > echo "anyterm ALL=NOPASSWD: /usr/bin/virsh console *" >> /etc/sudoers ACK, this now edits the file properly. However, there is still an issue with anyterm. Going directly to :81/?param= gets me: '[sudo] password for anyterm:' So something is still not right with the perms. I'll look into it more with you tomorrow. -j From justin at redfish-group.com Tue Aug 4 01:59:18 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 11:59:18 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A77592D.20202@redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> <4A77592D.20202@redfish-group.com> Message-ID: <4A7795F6.9000502@redfish-group.com> Justin Clacherty wrote: > >> Note that building on Fedora 10 will not work, since a couple of APIs >> that ovirt uses changed between 10 and 11. > > Ah, that might explain it. I guess today's job is to reinstall using > Fedora 11. I've just finished installing a Fedora 11 VM and building next. Ovirt is up and working and the node is seen as available and enabled. However, I am unable to add storage (iSCSI). If I add storage it just times out saying that "No agent responded within the timeout period.". If I ssh to the node and run "iscsiadm --mode discovery --type sendtargets --portal 192.168.50.190" the node returns the available targets so it seems the node can see the san. I can't see any obvious errors in the logs on the management server and there's nothing in the logs on the node. It looks as though the management server isn't sending the request to the node. Can you provide some direction on this? Justin. From justin at redfish-group.com Tue Aug 4 02:44:01 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 12:44:01 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A7795F6.9000502@redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> <4A77592D.20202@redfish-group.com> <4A7795F6.9000502@redfish-group.com> Message-ID: <4A77A071.6050009@redfish-group.com> Justin Clacherty wrote: > However, I am unable to add storage (iSCSI). If I add storage it just > times out saying that "No agent responded within the timeout > period.". If I ssh to the node and run "iscsiadm --mode discovery > --type sendtargets --portal 192.168.50.190" the node returns the > available targets so it seems the node can see the san. I can't see > any obvious errors in the logs on the management server and there's > nothing in the logs on the node. It looks as though the management > server isn't sending the request to the node. Can you provide some > direction on this? Spoke too soon. If you wait long enough then add the storage it magically works :) Justin. From justin at redfish-group.com Tue Aug 4 03:19:16 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 13:19:16 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A77A071.6050009@redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> <4A77592D.20202@redfish-group.com> <4A7795F6.9000502@redfish-group.com> <4A77A071.6050009@redfish-group.com> Message-ID: <4A77A8B4.2020701@redfish-group.com> Justin Clacherty wrote: > Justin Clacherty wrote: >> However, I am unable to add storage (iSCSI). If I add storage it >> just times out saying that "No agent responded within the timeout >> period.". If I ssh to the node and run "iscsiadm --mode discovery >> --type sendtargets --portal 192.168.50.190" the node returns the >> available targets so it seems the node can see the san. I can't see >> any obvious errors in the logs on the management server and there's >> nothing in the logs on the node. It looks as though the management >> server isn't sending the request to the node. Can you provide some >> direction on this? > > Spoke too soon. If you wait long enough then add the storage it > magically works :) > I had to restart the node to get the network bondings to register. Also had to restart the management server because the VM I'd tried to create was marked as unreachable so I couldn't do anything with it (including delete it). After restarting I experienced the same error when I went to start the VM. "No agent responded within the timeout period." Waited 10 minutes and still no luck so I deleted one of the storage pools and re-added it, this seemed to have woken up the node. Now I get: "Error creating virtual machine: Error: Error creating new domain (virDomainCreate). Subsystem qemu: internal error unable to start guest: bind(unix:/var/run/libvirt/qemu/Services.monitor): Permission denied qemu: could not open monitor device 'unix:/var/run/libvirt/qemu/Services.monitor,server,nowait' in DomainWrap.cpp:ManagementMethod:115" When I try to start the VM. Justin. From justin at redfish-group.com Tue Aug 4 03:36:58 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 13:36:58 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A77A8B4.2020701@redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> <4A77592D.20202@redfish-group.com> <4A7795F6.9000502@redfish-group.com> <4A77A071.6050009@redfish-group.com> <4A77A8B4.2020701@redfish-group.com> Message-ID: <4A77ACDA.3030508@redfish-group.com> Justin Clacherty wrote: > "Error creating virtual machine: Error: Error creating new domain > (virDomainCreate). Subsystem qemu: internal error unable to start > guest: bind(unix:/var/run/libvirt/qemu/Services.monitor): Permission > denied qemu: could not open monitor device > 'unix:/var/run/libvirt/qemu/Services.monitor,server,nowait' in > DomainWrap.cpp:ManagementMethod:115" Don't know what's going on with the "No agent responded within the timeout period." problem, but the above error is seen on the node. From /var/log/libvirt/qemu/Services.log LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin /usr/bin/qemu-kvm -S -M pc-0.11 -m 512 -smp 1 -name Services -uuid 06c7406b-4bfc-ac52-26de-00723afd688c -monitor unix:/var/run/libvirt/qemu/Services.monitor,server,nowait -boot c -drive file=/dev/disk/by-id/scsi-149455400000000000000000005000000061080000000d00f,if=ide,index=0,boot=on -net nic,macaddr=00:16:3e:0c:db:ac,vlan=0,name=nic.0 -net tap,fd=28,vlan=0,name=tap.0 -serial pty -parallel none -usb -vnc 0.0.0.0:0 -vga cirrus bind(unix:/var/run/libvirt/qemu/Services.monitor): Permission denied qemu: could not open monitor device 'unix:/var/run/libvirt/qemu/Services.monitor,server,nowait' It appears to be an selinux problem. From /var/log/audit/audit.log type=AVC msg=audit(1249356621.775:115): avc: denied { create } for pid=4611 comm="qemu-kvm" name="Services.monitor" scontext=system_u:system_r:svirt_t:s0:c405,c995 tcontext=system_u:object_r:svirt_var_run_t:s0:c405,c995 tclass=sock_file Justin. From justin at redfish-group.com Tue Aug 4 03:42:50 2009 From: justin at redfish-group.com (Justin Clacherty) Date: Tue, 04 Aug 2009 13:42:50 +1000 Subject: [Ovirt-devel] yum update broke ovirt In-Reply-To: <4A77ACDA.3030508@redfish-group.com> References: <398212241.11841249028315856.JavaMail.root@mail.redfish-group.com> <4A76F5CA.2020207@redhat.com> <4A77592D.20202@redfish-group.com> <4A7795F6.9000502@redfish-group.com> <4A77A071.6050009@redfish-group.com> <4A77A8B4.2020701@redfish-group.com> <4A77ACDA.3030508@redfish-group.com> Message-ID: <4A77AE3A.3090708@redfish-group.com> Justin Clacherty wrote: > It appears to be an selinux problem. From /var/log/audit/audit.log > > type=AVC msg=audit(1249356621.775:115): avc: denied { create } for > pid=4611 comm="qemu-kvm" name="Services.monitor" > scontext=system_u:system_r:svirt_t:s0:c405,c995 > tcontext=system_u:object_r:svirt_var_run_t:s0:c405,c995 tclass=sock_file Running "setenforce Permissive" on the node allows me to start the VM so it seems selinux is the problem here. I don't think it's causing the iscsi problems though. Justin. From mmorsi at redhat.com Tue Aug 4 14:52:47 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Tue, 4 Aug 2009 10:52:47 -0400 Subject: [Ovirt-devel] [PATCH node] add anyterm / virsh sudoers rule in same way other files are appended Message-ID: <1249397567-4273-1-git-send-email-mmorsi@redhat.com> --- scripts/ovirt-functions | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/scripts/ovirt-functions b/scripts/ovirt-functions index 98e88e7..fd86606 100644 --- a/scripts/ovirt-functions +++ b/scripts/ovirt-functions @@ -219,7 +219,9 @@ ANYTERM_LOCAL_ONLY=false EOF # permit it to run the virsh console - echo "anyterm ALL=NOPASSWD: /usr/bin/virsh console *" >> /etc/sudoers + cat >> /etc/sudoers << EOF +anyterm ALL=NOPASSWD: /usr/bin/virsh console * +EOF } md5() { -- 1.6.0.6 From mmorsi at redhat.com Tue Aug 4 16:00:35 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Tue, 4 Aug 2009 12:00:35 -0400 Subject: [Ovirt-devel] [PATCH node] fix anyterm / sudoers setup Message-ID: <1249401635-11513-1-git-send-email-mmorsi@redhat.com> --- scripts/ovirt-functions | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/scripts/ovirt-functions b/scripts/ovirt-functions index 98e88e7..7edb0d2 100644 --- a/scripts/ovirt-functions +++ b/scripts/ovirt-functions @@ -219,7 +219,10 @@ ANYTERM_LOCAL_ONLY=false EOF # permit it to run the virsh console - echo "anyterm ALL=NOPASSWD: /usr/bin/virsh console *" >> /etc/sudoers + cat >> /etc/sudoers << EOF +anyterm ALL=NOPASSWD: /usr/bin/virsh console * +anyterm ALL=NOPASSWD: /usr/bin/virsh console +EOF } md5() { -- 1.6.0.6 From mmorsi at redhat.com Tue Aug 4 19:09:46 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Tue, 04 Aug 2009 15:09:46 -0400 Subject: [Ovirt-devel] [PATCH server] Removed workaround for qmf bug that yielded 0 cpus and 0 nics for nodes from time to time. In-Reply-To: <1249075058.5296.7.camel@lenovo> References: <1249068718-9464-1-git-send-email-arroy@redhat.com> <1249075058.5296.7.camel@lenovo> Message-ID: <4A78877A.3070500@redhat.com> Jason Guiditta wrote: > On Fri, 2009-07-31 at 19:31 +0000, Arjun Roy wrote: > >> The original bug report is filed here: >> http://issues.apache.org/jira/browse/QPID-1973 >> >> The svn revision 975209 and onwards for qpid contains the fix. >> --- >> src/host-browser/host-register.rb | 2 +- >> 1 files changed, 1 insertions(+), 1 deletions(-) >> >> diff --git a/src/host-browser/host-register.rb b/src/host-browser/host-register.rb >> index 57c11b7..06d8553 100755 >> --- a/src/host-browser/host-register.rb >> +++ b/src/host-browser/host-register.rb >> @@ -92,7 +92,7 @@ class HostRegister < Qpid::Qmf::Console >> end >> >> @logger.info "Connecting to amqp://#{server}:#{port}" >> - @session = Qpid::Qmf::Session.new(:console => self, :manage_connections => false) >> + @session = Qpid::Qmf::Session.new(:console => self, :manage_connections => true) >> @broker = @session.add_broker("amqp://#{server}:#{port}", :mechanism => 'GSSAPI') >> >> rescue Exception => ex >> > > ACK. This, combined with new qpid stuff on ovirt repo fixes 'no cpu' > issue for me, seems to work consistently > > _______________________________________________ > Ovirt-devel mailing list > Ovirt-devel at redhat.com > https://www.redhat.com/mailman/listinfo/ovirt-devel > Pushed From mmorsi at redhat.com Tue Aug 4 20:30:16 2009 From: mmorsi at redhat.com (Mohammed Morsi) Date: Tue, 4 Aug 2009 16:30:16 -0400 Subject: [Ovirt-devel] [PATCH server] require at least one vm network if pxe booting Message-ID: <1249417816-23540-1-git-send-email-mmorsi@redhat.com> --- src/app/models/vm.rb | 2 ++ src/app/views/vm/_form.rhtml | 23 +++++++++++++++++++++++ 2 files changed, 25 insertions(+), 0 deletions(-) diff --git a/src/app/models/vm.rb b/src/app/models/vm.rb index 3e0a6fa..9d326ac 100644 --- a/src/app/models/vm.rb +++ b/src/app/models/vm.rb @@ -464,6 +464,8 @@ class Vm < ActiveRecord::Base self.storage_volumes=@storage_volumes_pending @storage_volumes_pending = [] end + errors.add("nics", "must specify at least one network if pxe booting off a network") unless boot_device == BOOT_DEV_NETWORK && nics.size > 0 + end end diff --git a/src/app/views/vm/_form.rhtml b/src/app/views/vm/_form.rhtml index 5f519fc..39ce55e 100644 --- a/src/app/views/vm/_form.rhtml +++ b/src/app/views/vm/_form.rhtml @@ -51,6 +51,9 @@