So, in oVirt if I want a single storage, without high availability, what is the best solution?
Can I use gluster replica 1 for a single storage?

by the way, cluster.min-free-disk: 1% not 10% the default value. And I still cannot remove the disk.
I think I will destroy the volume and configure it again.

Thanks

José


De: "Strahil Nikolov" <hunter86_bg@yahoo.com>
Para: suporte@logicworks.pt
Cc: "users" <users@ovirt.org>
Enviadas: Terça-feira, 27 De Outubro de 2020 17:11:14
Assunto: Re: [ovirt-users] Re: Gluster Domain Storage full

Nope,

oficially oVirt supports only replica 3 (replica 3 arbiter 1) or replica 1 (which actually is a single brick distributed ) volumes.
If you have issues related to the Gluster Volume - liek this case , the community support will be "best effort".

Best Regards,
Strahil Nikolov






В вторник, 27 октомври 2020 г., 18:24:07 Гринуич+2, <suporte@logicworks.pt> написа:





But there are also distributed volume types, right?
Replicated is if you want high availability, that is not the case.

Thanks

José

________________________________
De: "Strahil Nikolov" <hunter86_bg@yahoo.com>
Para: suporte@logicworks.pt
Cc: "users" <users@ovirt.org>
Enviadas: Terça-feira, 27 De Outubro de 2020 15:49:22
Assunto: Re: [ovirt-users] Re: Gluster Domain Storage full

You have exactly 90% used space.
The Gluster's default protection value is exactly 10%:


Option: cluster.min-free-disk
Default Value: 10%
Description: Percentage/Size of disk space, after which the process starts balancing out the cluster, and logs will appear in log files

I would recommend you to temporarily drop that value till you do the cleanup.

gluster volume set data cluster.min-free-disk 5%

To restore the default values of any option:
gluster volume reset data cluster.min-free-disk


P.S.: Keep in mind that if you have sparse disks for your VMs , they will be unpaused and start eating the last space you got left ... so be quick :)

P.S.2: I hope you know that the only supported volume types are 'distributed-replicated' and 'replicated' :)

Best Regards,
Strahil Nikolov








В вторник, 27 октомври 2020 г., 11:02:10 Гринуич+2, suporte@logicworks.pt <suporte@logicworks.pt> написа:





This is a simple installation with one storage, 3 hosts. One volume with 2 bricks, the second brick was to get more space to try to remove the disk, but without success

]# df -h
Filesystem                      Size  Used Avail Use% Mounted on
/dev/md127                       50G  2.5G   48G   5% /
devtmpfs                        7.7G     0  7.7G   0% /dev
tmpfs                           7.7G   16K  7.7G   1% /dev/shm
tmpfs                           7.7G   98M  7.6G   2% /run
tmpfs                           7.7G     0  7.7G   0% /sys/fs/cgroup
/dev/md126                     1016M  194M  822M  20% /boot
/dev/md125                      411G  407G  4.1G 100% /home
gfs2.domain.com:/data     461G  414G   48G  90% /rhev/data-center/mnt/glusterSD/gfs2.domain.com:_data
tmpfs                           1.6G     0  1.6G   0% /run/user/0


# gluster volume status data
Status of volume: data
Gluster process                             TCP Port  RDMA Port  Online  Pid
------------------------------------------------------------------------------
Brick gfs2.domain.com:/home/brick1    49154     0          Y       8908
Brick gfs2.domain.com:/brickx         49155     0          Y       8931

Task Status of Volume data
------------------------------------------------------------------------------
There are no active volume tasks

# gluster volume info data

Volume Name: data
Type: Distribute
Volume ID: 2d3ea533-aca3-41c4-8cb6-239fe4f82bc3
Status: Started
Snapshot Count: 0
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: gfs2.domain.com:/home/brick1
Brick2: gfs2.domain.com:/brickx
Options Reconfigured:
cluster.min-free-disk: 1%
cluster.data-self-heal-algorithm: full
performance.low-prio-threads: 32
features.shard-block-size: 512MB
features.shard: on
storage.owner-gid: 36
storage.owner-uid: 36
transport.address-family: inet
nfs.disable: on

________________________________
De: "Strahil Nikolov" <hunter86_bg@yahoo.com>
Para: suporte@logicworks.pt
Cc: "users" <users@ovirt.org>
Enviadas: Terça-feira, 27 De Outubro de 2020 1:00:08
Assunto: Re: [ovirt-users] Re: Gluster Domain Storage full

So what is the output of "df" agianst :
- all bricks in the volume (all nodes)
- on the mount point in /rhev/mnt/....

Usually adding a new brick (per host) in replica 3 volume should provide you more space.
Also what is the status of the volume:

gluster volume status <VOL>
gluster volume info <VOL> 


Best Regards,
Strahil Nikolov






В четвъртък, 15 октомври 2020 г., 16:55:27 Гринуич+3, <suporte@logicworks.pt> написа:





Hello,

I just add a second brick to the volume. Now I have 10% free, but still cannot delete the disk. Still the same message:

VDSM command DeleteImageGroupVDS failed: Could not remove all image's volumes: (u'b6165676-a6cd-48e2-8925-43ed49bc7f8e [Errno 28] No space left on device',)

Any idea?
Thanks

José

________________________________
De: "Strahil Nikolov" <hunter86_bg@yahoo.com>
Para: suporte@logicworks.pt
Cc: "users" <users@ovirt.org>
Enviadas: Terça-feira, 22 De Setembro de 2020 13:36:27
Assunto: Re: [ovirt-users] Re: Gluster Domain Storage full

Any option to extend the Gluster Volume ?

Other approaches are quite destructive. I guess , you can obtain the VM's xml via virsh and then copy the disks to another pure-KVM host.
Then you can start the VM , while you are recovering from the situation.

virsh -c qemu:///system?authfile=/etc/ovirt-hosted-engine/virsh_auth.conf dumpxml <VM_NAME> > /some/path/<VM_NAME>.xml

Once you got the VM running on a pure-KVM host , you can go to oVirt and try to wipe the VM from the UI. 


Usually those 10% reserve is just in case something like this one has happened, but Gluster doesn't check it every second (or the overhead will be crazy).

Maybe you can extend the Gluster volume temporarily , till you manage to move away the VM to a bigger storage. Then you can reduce the volume back to original size.

Best Regards,
Strahil Nikolov



В вторник, 22 септември 2020 г., 14:53:53 Гринуич+3, suporte@logicworks.pt <suporte@logicworks.pt> написа:





Hello Strahil,

I just set cluster.min-free-disk to 1%:
# gluster volume info data

Volume Name: data
Type: Distribute
Volume ID: 2d3ea533-aca3-41c4-8cb6-239fe4f82bc3
Status: Started
Snapshot Count: 0
Number of Bricks: 1
Transport-type: tcp
Bricks:
Brick1: node2.domain.com:/home/brick1
Options Reconfigured:
cluster.min-free-disk: 1%
cluster.data-self-heal-algorithm: full
performance.low-prio-threads: 32
features.shard-block-size: 512MB
features.shard: on
storage.owner-gid: 36
storage.owner-uid: 36
transport.address-family: inet
nfs.disable: on

But still get the same error: Error while executing action: Cannot move Virtual Disk. Low disk space on Storage Domain
I restarted the glusterfs volume.
But I can not do anything with the VM disk.


I know that filling the bricks is very bad, we lost access to the VM. I think there should be a mechanism to prevent stopping the VM.
we should continue to have access to the VM to free some space.

If you have a VM with a Thin Provision disk, if the VM fills the entire disk, we got the same problem.

Any idea?

Thanks

José



________________________________
De: "Strahil Nikolov" <hunter86_bg@yahoo.com>
Para: "users" <users@ovirt.org>, suporte@logicworks.pt
Enviadas: Segunda-feira, 21 De Setembro de 2020 21:28:10
Assunto: Re: [ovirt-users] Gluster Domain Storage full

Usually gluster has a 10% reserver defined in 'cluster.min-free-disk' volume option.
You can power off the VM , then set cluster.min-free-disk
to 1% and immediately move any of the VM's disks to another storage domain.

Keep in mind that filling your bricks is bad and if you eat that reserve , the only option would be to try to export the VM as OVA and then wipe from current storage and import in a bigger storage domain.

Of course it would be more sensible to just expand the gluster volume (either scale-up the bricks -> add more disks, or scale-out -> adding more servers with disks on them), but I guess that is not an option - right ?

Best Regards,
Strahil Nikolov








В понеделник, 21 септември 2020 г., 15:58:01 Гринуич+3, suporte@logicworks.pt <suporte@logicworks.pt> написа:





Hello,

I'm running oVirt Version 4.3.4.3-1.el7.
I have a small GlusterFS Domain storage brick on a dedicated filesystem serving only one VM.
The VM filled all the Domain storage.
The Linux filesystem has 4.1G available and 100% used, the mounted brick has 0GB available and 100% used

I can not do anything with this disk, for example, if I try to move it to another Gluster Domain Storage get the message:

Error while executing action: Cannot move Virtual Disk. Low disk space on Storage Domain

Any idea?

Thanks

--
________________________________
Jose Ferradeira
http://www.logicworks.pt
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/privacy-policy.html
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/WFN2VOQZPPVCGXAIFEYVIDEVJEUCSWY7/
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/privacy-policy.html
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/AIJUP2HZIWRSQHN4XU3BGGT2ZDKEVJZ3/