Current status of Ceph support in oVirt (2019)?
Hi, I currently have a 3-node HA cluster running Proxmox (with integrated Ceph). oVirt looks pretty neat, however, and I'm excited to check it out. One of the things I love about Proxmox is the integrated Ceph support. I saw on the mailing lists that there is some talk of Ceph support earlier, but it was via OpenStack/Cinder. What exactly does this mean? 1. Does this require you to install OpenStack, or will a vanilla Ceph installation work? 2. Is it possible to deploy Ceph on the same nodes that run oVirt? (i.e. is a 3-node oVirt + Ceph cluster possible?) 3. Is there any monitoring/management of Ceph from within oVirt? (Guessing no?) 4. Are all the normal VM features working yet, or is this planned? 5. Is making Ceph a first-class citizen (like Gluster) on oVirt on the roadmap? Thanks, Victor https://www.reddit.com/r/ovirt/comments/ci38zp/ceph_rbd_support_in_ovirt_for...
The current plan to integrate ceph is via cinderlib integration[1] (currently in tech preview mode) because we still have no packaging ready, there are some manual installation steps required, but there is no need to install and configure openstack/cinder
1. Does this require you to install OpenStack, or will a vanilla Ceph installation work? a vanilla installation will work with cinderlib
2. Is it possible to deploy Ceph on the same nodes that run oVirt? (i.e. is a 3-node oVirt + Ceph cluster possible?) I haven't tried it, but should be possible
3. Is there any monitoring/management of Ceph from within oVirt? (Guessing no?) No, cinderlib is storage agnostic
4. Are all the normal VM features working yet, or is this planned? Most features (starting/stopping/snapshots/live migration) are working, but not all are fully tested (specifically snapshots)
5. Is making Ceph a first-class citizen (like Gluster) on oVirt on the roadmap? Not at the moment, maybe once cinderlib integration matures and we have more feedback and users for the feature
[1] https://www.ovirt.org/develop/release-management/features/storage/cinderlib-... On Sun, Nov 24, 2019 at 12:41 PM <victorhooi@yahoo.com> wrote:
Hi,
I currently have a 3-node HA cluster running Proxmox (with integrated Ceph). oVirt looks pretty neat, however, and I'm excited to check it out.
One of the things I love about Proxmox is the integrated Ceph support.
I saw on the mailing lists that there is some talk of Ceph support earlier, but it was via OpenStack/Cinder. What exactly does this mean?
1. Does this require you to install OpenStack, or will a vanilla Ceph installation work? 2. Is it possible to deploy Ceph on the same nodes that run oVirt? (i.e. is a 3-node oVirt + Ceph cluster possible?) 3. Is there any monitoring/management of Ceph from within oVirt? (Guessing no?) 4. Are all the normal VM features working yet, or is this planned? 5. Is making Ceph a first-class citizen (like Gluster) on oVirt on the roadmap?
Thanks, Victor
https://www.reddit.com/r/ovirt/comments/ci38zp/ceph_rbd_support_in_ovirt_for... _______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/site/privacy-policy/ oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/QMPXXQQOMKEQJI...
Am 24.11.19 um 11:58 schrieb Benny Zlotnik:
The current plan to integrate ceph is via cinderlib integration[1] (currently in tech preview mode) because we still have no packaging ready, there are some manual installation steps required, but there is no need to install and configure openstack/cinder
1. Does this require you to install OpenStack, or will a vanilla Ceph installation work? a vanilla installation will work with cinderlib
2. Is it possible to deploy Ceph on the same nodes that run oVirt? (i.e. is a 3-node oVirt + Ceph cluster possible?) I haven't tried it, but should be possible
3. Is there any monitoring/management of Ceph from within oVirt? (Guessing no?) No, cinderlib is storage agnostic
4. Are all the normal VM features working yet, or is this planned? Most features (starting/stopping/snapshots/live migration) are working, but not all are fully tested (specifically snapshots)
5. Is making Ceph a first-class citizen (like Gluster) on oVirt on the roadmap? Not at the moment, maybe once cinderlib integration matures and we have more feedback and users for the feature
[1] https://www.ovirt.org/develop/release-management/features/storage/cinderlib-...
We are using Ceph with oVirt (via standalone Cinder) extensively in a production environment. I tested oVirt cinderlib integration in our dev environment, gave some feedback here on the list and am currently waiting for the future development. IMHO cinderlib in oVirt is currently not fit for production use, I think this matches your assessment. What can we do to help advance Ceph integration in oVirt? What are the plans for oVirt 4.4? Will standalone Cinder storage domains still be supported in oVirt 4.4? Will there be a migration scenario from standalone Cinder to cinderlib? Accidentally just yesterday I had an incident in our test environment where migration of a VM with MANAGED_BLOCK_STORAGE (=cinderlib) disks failed (causes are known and unrelated to cinderlib). Restarting the VM failed because of leftover rbdmapped devices. This is similar to the case I reported in https://bugzilla.redhat.com/show_bug.cgi?id=1697496. I don't clearly see if this fixed or not. Shall I report my recent problem? thanks Matthias
We are using Ceph with oVirt (via standalone Cinder) extensively in a production environment. I tested oVirt cinderlib integration in our dev environment, gave some feedback here on the list and am currently waiting for the future development. IMHO cinderlib in oVirt is currently not fit for production use, I think this matches your assessment.
What can we do to help advance Ceph integration in oVirt? What are the plans for oVirt 4.4? Will standalone Cinder storage domains still be supported in oVirt 4.4? Will there be a migration scenario from standalone Cinder to cinderlib? TBH, to help, bug reports are probably the most useful. Having feedback from users with "real world" setups and usage will help us improve. As stated before, our biggest issue at the moment is packaging, once it's handled it will be significantly easier to test and develop. Standalone cinder domains were never actually supported (never left tech preview), we do not have immediate plans for an upgrade path, but you can definitely submit an RFE for this
Accidentally just yesterday I had an incident in our test environment where migration of a VM with MANAGED_BLOCK_STORAGE (=cinderlib) disks failed (causes are known and unrelated to cinderlib). Restarting the VM failed because of leftover rbdmapped devices. This is similar to the case I reported in https://bugzilla.redhat.com/show_bug.cgi?id=1697496. I don't clearly see if this fixed or not. Shall I report my recent problem? We added better cleanup for migration cleanup to handle the bug. There is another issue which was not known at the time with multipath preventing unmapping rbd devices[1], and it might be what you experienced. This can be fixed manually by blacklisting rbd devices in multipath conf, but once the bug is fixed vdsm will handle configuring it.
participants (3)
-
Benny Zlotnik -
Matthias Leopold -
victorhooi＠yahoo.com