[ovirt-users] template mystery
Hi, I've got 4 VMs on a 3.6 cluster, and they are listed as based on a template called 'sl6-workstation', yet I cannot find this template anywhere listed in the templates tab or tree. When I go to create a new VM, it gives me the option in the template section to use: sl6-workstation(1) or sl6-workstation(latest) I'm not sure what the difference between these two are, but I'm curious why it doesn't give me the option of not to use a template. It will even be the only option if I select FreeBSD or some other non-sl6 OS. Any ideas? Thanks, Cam ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] backup-engine 3.6.4.1-1 Files Compressor still broken in Centos 6.7? Work around
Of course, I meant "annoying" in a good way ;) Yep, 'xz' was missing, libs were there but not the compressor [root@engine1 ~]# rpm -qa | grep xz xz-libs-4.999.9-0.5.beta.20091007git.el6.x86_64 Installed package xz [root@engine1 ~]# rpm -qa | grep xz xz-4.999.9-0.5.beta.20091007git.el6.x86_64 xz-libs-4.999.9-0.5.beta.20091007git.el6.x86_64 Backup worked like a charm. Cheers On Sun, Apr 17, 2016 at 7:36 AM, Jack Greenewrote: > Been trying to run a complete backup on a fresh install to get rid of > that annoying alert. Well, I hope you prefer that over not be able to restore you engine if it dies :-) http://www.ovirt.org/develop/release-management/features/backupawareness/ Believe me, this was done after real people lost data and time, not because anyone found it fun to do. > > > > engine-backup --scope=all > > > > Kept failing with the error: > > 2016-04-17 04:05:31 8742: Creating temp folder > /tmp/engine-backup.ymrXeYZay1/tar > > 2016-04-17 04:05:31 8742: OUTPUT: - Files > > 2016-04-17 04:05:31 8742: Backing up files to > /tmp/engine-backup.ymrXeYZay1/tar/files > > 2016-04-17 04:05:31 8742: FATAL: Failed backing up /etc/ovirt-engine Perhaps it's because you do not have 'xz' installed, and engine-backup defaults to use xz for the --files-compressor? It will be fixed in 3.6.5, where engine-backup is packaged separately, with proper 'Require:' lines in the spec file: https://bugzilla.redhat.com/show_bug.cgi?id=1320092 https://gerrit.ovirt.org/55045 Please try installing 'xz' and run engine-backup again. If it fails, please report back. Thanks. > > > > Research shows a bug reported on version 3.6.0 related to the tar > options > > https://gerrit.ovirt.org/#/c/48596/3/packaging/bin/engine-backup.sh > > https://bugzilla.redhat.com/show_bug.cgi?id=1282397 > > > > But a grep of the script shows the changes have been made > > [root@engine1 iso]# grep cpSs > /usr/share/ovirt-engine/bin/engine-backup.sh > > [root@engine1 iso]# grep cpS > /usr/share/ovirt-engine/bin/engine-backup.sh > > tar -C "${dir}" -cpS"${ARCHIVE_COMPRESS_OPTION}"f "${file}" . > >> "${tar_log}" 2>&1 > > tar -C / --files-from - > -cpS"${FILES_COMPRESS_OPTION}"f "${target}" || logdie "Failed backing up > ${paths}" > > > > Verified I had a new version (just re-installed ovirt-engine today) > > > > ovirt-engine-3.6.4.1-1.el6.noarch > > Linux engine1.attlocal.net 2.6.32-573.22.1.el6.x86_64 #1 SMP Wed Mar > 23 > 03:35:39 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux > > > > # ovirt-engine-backup - oVirt engine backup and restore utility > > # Copyright (C) 2013-2016 Red Hat, Inc. > > > > Then I noticed there is an option to turn off –files-compressor. > Doing this I was able to get a complete backup. > > > > [root@engine1 iso]# engine-backup --mode=backup --scope=files > --file=/mnt/h97m/backup_files_20160417 > --log=/root/backup_files_20160417.log --files-compressor=None > > Backing up: > > Notifying engine > > - Files > > Packing into file '/mnt/h97m/backup_files_20160417' > > Notifying engine > > Done. That's an option too. And you do not loose much by using it, because the files tar is compressed anyway when the archive is compressed, using gzip by default. 'files' are usually very small, compared to the databases, so the difference is not significant. Best, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] several questions about serial console
On Fri, Apr 15, 2016 at 7:18 PM, Nathanaël Blanchetwrote: > > > Le 15/04/2016 17:27, Nathanaël Blanchet a écrit : > > Hi all, > > About serial console: > > how to get out of a selectionned vm where we are on the login prompt (why > not backing up to the vm menu) rather than killing the ssh process or > closing the terminal? Usual "^] " doesn't work there. > according to > https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.6/html-single/Virtual_Machine_Management_Guide/index.html > > # ssh -t -p ovirt-vmconsole@MANAGER_IP --vm-name vm1 > > should allow to connect directly to a vm on its serial port, and it > is very useful when there are a big number of vm. In reality, we get a SSH > issue : "unknown option -- -" Seems like a bug in the documentation. See also the README file: /usr/share/doc/ovirt-vmconsole/README on your machine, or https://gerrit.ovirt.org/gitweb?p=ovirt-vmconsole.git;a=blob;f=README > > An other question is : Why the vm order is not alphabetic? it could simplify > the search when too many vms are displayed. And a simple "sort -k2" command > should do the stuff... Please open an RFE for this. Thanks. I think you can do that yourself, by editing: /etc/ovirt-vmconsole/ovirt-vmconsole-proxy/conf.d/20-ovirt-vmconsole-proxy-helper.conf > > if we want to add 5 users with UserVmManager role on 150 vms and I can't use > group for this stuff, this means I need to do this with an ovirt-shell > script like : > # for i in $(cat /tmp/ids.ovirt); do for j in $(cat /tmp/list_all);do > ovirt-shell -E "add permission --parent-vm-name $j --user-id $i --role-name > UserVmManager"; done; done > and 5*150 API connections only because I can't add several user ids on the > same "add permission" line ? It's doable, but not not very convinient and > very long if I have many more users to add. > Why can't we add permission by user-name and not user-id ? Adding Francesco. Thanks and best regards, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] backup-engine 3.6.4.1-1 Files Compressor still broken in Centos 6.7? Work around
On Sun, Apr 17, 2016 at 7:36 AM, Jack Greenewrote: > Been trying to run a complete backup on a fresh install to get rid of that > annoying alert. Well, I hope you prefer that over not be able to restore you engine if it dies :-) http://www.ovirt.org/develop/release-management/features/backupawareness/ Believe me, this was done after real people lost data and time, not because anyone found it fun to do. > > > > engine-backup --scope=all > > > > Kept failing with the error: > > 2016-04-17 04:05:31 8742: Creating temp folder > /tmp/engine-backup.ymrXeYZay1/tar > > 2016-04-17 04:05:31 8742: OUTPUT: - Files > > 2016-04-17 04:05:31 8742: Backing up files to > /tmp/engine-backup.ymrXeYZay1/tar/files > > 2016-04-17 04:05:31 8742: FATAL: Failed backing up /etc/ovirt-engine Perhaps it's because you do not have 'xz' installed, and engine-backup defaults to use xz for the --files-compressor? It will be fixed in 3.6.5, where engine-backup is packaged separately, with proper 'Require:' lines in the spec file: https://bugzilla.redhat.com/show_bug.cgi?id=1320092 https://gerrit.ovirt.org/55045 Please try installing 'xz' and run engine-backup again. If it fails, please report back. Thanks. > > > > Research shows a bug reported on version 3.6.0 related to the tar options > > https://gerrit.ovirt.org/#/c/48596/3/packaging/bin/engine-backup.sh > > https://bugzilla.redhat.com/show_bug.cgi?id=1282397 > > > > But a grep of the script shows the changes have been made > > [root@engine1 iso]# grep cpSs /usr/share/ovirt-engine/bin/engine-backup.sh > > [root@engine1 iso]# grep cpS /usr/share/ovirt-engine/bin/engine-backup.sh > > tar -C "${dir}" -cpS"${ARCHIVE_COMPRESS_OPTION}"f "${file}" . >> > "${tar_log}" 2>&1 > > tar -C / --files-from - -cpS"${FILES_COMPRESS_OPTION}"f > "${target}" || logdie "Failed backing up ${paths}" > > > > Verified I had a new version (just re-installed ovirt-engine today) > > > > ovirt-engine-3.6.4.1-1.el6.noarch > > Linux engine1.attlocal.net 2.6.32-573.22.1.el6.x86_64 #1 SMP Wed Mar 23 > 03:35:39 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux > > > > # ovirt-engine-backup - oVirt engine backup and restore utility > > # Copyright (C) 2013-2016 Red Hat, Inc. > > > > Then I noticed there is an option to turn off –files-compressor. Doing this > I was able to get a complete backup. > > > > [root@engine1 iso]# engine-backup --mode=backup --scope=files > --file=/mnt/h97m/backup_files_20160417 > --log=/root/backup_files_20160417.log --files-compressor=None > > Backing up: > > Notifying engine > > - Files > > Packing into file '/mnt/h97m/backup_files_20160417' > > Notifying engine > > Done. That's an option too. And you do not loose much by using it, because the files tar is compressed anyway when the archive is compressed, using gzip by default. 'files' are usually very small, compared to the databases, so the difference is not significant. Best, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] audit_log table performance tuning
On Sun, Apr 17, 2016 at 9:33 AM, Oved Ourfaliwrote: > Juan - we should try to reduce this number consumed by CFME, if possible. > CC-ing Eli for DB related tips. > Reminds me of https://gerrit.ovirt.org/#/c/55743/ . Y. > > On Fri, Apr 15, 2016 at 1:12 AM, Marina Kalinin > wrote: > >> Hi, >> >> Any suggestions or maybe already available features in the pipeline for >> tuning the database, and specifically the audit_log table? >> >> The problem today is that with multiple applications accessing the engine >> through the RestAPI, especially deployments with CloudForms, create huge >> amount of login records in the audit_table. Which, in turns, consumes most >> of the available memory on the machine running the engine and the database >> and results in a terrible performance of engine and inaccessible Web UI. >> >> The solution today is to delete those records from the table [1]: >> => delete from audit_log where message like '%logged%'; >> >> >> Are there any current tunings we can apply to the database? >> And if not - do we have any RFEs on limiting the records entered to the >> database or a way to delete/filter those records somehow from the WebUI? >> All I could find was RFE#1120659 [2], but it does not describe the exact >> issue. >> >> >> -- >> Thanks, >> Marina. >> >> >> [1] https://access.redhat.com/solutions/2110011 >> [2] https://bugzilla.redhat.com/show_bug.cgi?id=1120659 >> ___ >> Users mailing list >> Users@ovirt.org >> http://lists.ovirt.org/mailman/listinfo/users >> >> >> > > ___ > Users mailing list > Users@ovirt.org > http://lists.ovirt.org/mailman/listinfo/users > > ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] hosted-engine stuck "failed liveliness check" "detail": "up"
On Fri, Apr 15, 2016 at 2:33 PM, Paul Groeneweg | Pazionwrote: > Thanks for the help! > > I managed to fix it :-) > > I made a device from the hosted engine file with losetup and for the LVM > again. > From there I was able to fsck my partition. And answered a lot of y to the > list of inconsistences. > I removed the loop back devices. > I turned off maintenance mode and hosted engine started :-) > Few hours later , restore of the pgsql db and scratching my head, very happy > it is running again without any VMS being offline meanwhile :) > > I will defintely save the way to start hosted-engine with cd iso, as this > seems less error sensitive and easier to do. > It was recently mentioned [1] here that in 3.6, you can't start the vm when you are in maintenance. Was this changed? Adding also Martin. [1] http://lists.ovirt.org/pipermail/users/2016-January/036993.html Best, > > > Thanks! > > > Op vr 15 apr. 2016 om 11:35 schreef Simone Tiraboschi : >> >> On Fri, Apr 15, 2016 at 11:02 AM, Paul Groeneweg | Pazion >> wrote: >> > Thanks! >> > >> > I managed to get the console through: >> > >> > hosted-engine --add-console-password >> > /bin/remote-viewer vnc://localhost:5900 >> > >> > Turns out, there seems to be some corruption on the partition: >> > http://screencast.com/t/6iR0U3QuI >> > >> > Is there a way to boot from CD, so I can start rescue mode? >> >> Yes, >> you have to set global maintenance mode to avoid VM start attempts on >> other hosts. >> >> You have to poweroff the engine VM with: >> hosted-engine --vm-shutdown/--vm-poweroff (probably the second one in >> your case) >> Then you have to create a copy of >> /var/run/ovirt-hosted-engine-ha/vm.conf and you have to manually edit >> it in order to attach the ISO image and change the boot order. >> At that point you can start the engine VM with your custom configuration >> with: >> hosted-engine --vm-start --vm-conf=my_custom_vm.conf >> >> Good luck! >> >> > Op vr 15 apr. 2016 om 10:58 schreef Martin Sivak : >> >> >> >> Hi, >> >> >> >> you can access the console using vnc or use virsh to get access to the >> >> serial console. >> >> >> >> Check the following commands on the host where the VM is currently >> >> running: >> >> >> >> virsh -r list >> >> virsh -r console HostedEngine >> >> virsh -r vncdisplay HostedEngine >> >> >> >> Those should give you enough pointers to connect to the VM. >> >> >> >> Regards >> >> >> >> Martin Sivak >> >> >> >> >> >> >> >> On Fri, Apr 15, 2016 at 10:14 AM, Paul Groeneweg | Pazion >> >> wrote: >> >> > >> >> > Tonight my server with NFS hosted-engine mount crashed. >> >> > Now all is back online ,except the hosted engine. I can't ping or >> >> > ssh >> >> > the >> >> > machine >> >> > >> >> > when I do hosted-engine --vm-status, I get: >> >> > >> >> > .. >> >> > --== Host 2 status ==-- >> >> > >> >> > Status up-to-date : True >> >> > Hostname : geisha-3.pazion.nl >> >> > Host ID: 2 >> >> > Engine status : {"reason": "failed liveliness >> >> > check", >> >> > "health": "bad", "vm": "up", "detail": "up"} >> >> > Score : 3400 >> >> > stopped: False >> >> > Local maintenance : False >> >> > crc32 : d71d7c6b >> >> > Host timestamp : 4404 >> >> > >> >> > >> >> > I tried restarting all services/nfs mounts, start hosted engine on >> >> > other >> >> > hosts, but all the same host up, but liveliness failed and unable to >> >> > access >> >> > the network/IP. >> >> > >> >> > I imagine it is stuck at the console requiring a fsck check maybe? >> >> > Is there a way to access the boot display directly? >> >> > >> >> > Any help is highly appreciated! >> >> > >> >> > >> >> > >> >> > >> >> > >> >> > >> >> > ___ >> >> > Users mailing list >> >> > Users@ovirt.org >> >> > http://lists.ovirt.org/mailman/listinfo/users >> >> > >> > >> > >> > ___ >> > Users mailing list >> > Users@ovirt.org >> > http://lists.ovirt.org/mailman/listinfo/users >> > > > > ___ > Users mailing list > Users@ovirt.org > http://lists.ovirt.org/mailman/listinfo/users > -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] audit_log table performance tuning
Juan - we should try to reduce this number consumed by CFME, if possible. CC-ing Eli for DB related tips. On Fri, Apr 15, 2016 at 1:12 AM, Marina Kalininwrote: > Hi, > > Any suggestions or maybe already available features in the pipeline for > tuning the database, and specifically the audit_log table? > > The problem today is that with multiple applications accessing the engine > through the RestAPI, especially deployments with CloudForms, create huge > amount of login records in the audit_table. Which, in turns, consumes most > of the available memory on the machine running the engine and the database > and results in a terrible performance of engine and inaccessible Web UI. > > The solution today is to delete those records from the table [1]: > => delete from audit_log where message like '%logged%'; > > > Are there any current tunings we can apply to the database? > And if not - do we have any RFEs on limiting the records entered to the > database or a way to delete/filter those records somehow from the WebUI? > All I could find was RFE#1120659 [2], but it does not describe the exact > issue. > > > -- > Thanks, > Marina. > > > [1] https://access.redhat.com/solutions/2110011 > [2] https://bugzilla.redhat.com/show_bug.cgi?id=1120659 > ___ > Users mailing list > Users@ovirt.org > http://lists.ovirt.org/mailman/listinfo/users > > > ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Move hosted storage
On Fri, Apr 15, 2016 at 11:25 AM, Marcin Michtawrote: > Hi, > > I created hosted engine on the NFS storage at server1. I wonder - can I move > (migrate) hosted storage to server2 (also NFS) and how to make work after > all. Are both on the same LAN? If so, you can just add the other as an additional host, migrate the VM to it, then remove the first from the engine (after moving it to maintenance). IIRC HA will forget about it after a week or so (in the output of 'hosted-engine --vm-status'). In principle you can remove it manually using 'hosted-engine --clean-metadata --host-id=1'. See also: https://www.ovirt.org/documentation/how-to/hosted-engine/ If on different LANs, you can try using backup/restore: https://www.ovirt.org/documentation/admin-guide/hosted-engine-backup-and-restore/ Best, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users