| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-01-17 | |||
| 12:17:09 | admin0 | i will try to remove the libvirt-bin package and reinstall it | |
| 12:17:19 | sean-k-mooney | its not libvirt that is the issue | |
| 12:17:31 | sean-k-mooney | it cant find the x86 qemu bindary | |
| 12:17:36 | stephenfin | sean-k-mooney: Any decisions on what we want to do about https://github.com/openstack/nova/blob/master/nova/virt/libvirt/vif.py#L299-L301 ? | |
| 12:17:42 | stephenfin | That function doesn't exist | |
| 12:17:44 | sean-k-mooney | so you need to reinstall qemu | |
| 12:18:28 | sean-k-mooney | set_vif_host_backedn_brdige_config? | |
| 12:18:34 | sean-k-mooney | stephenfin: what is the context? | |
| 12:18:43 | stephenfin | yes | |
| 12:18:53 | stephenfin | it hasn't existed since commit 92323586b5d03b31c18657ad64646a8ce3b8a742 | |
| 12:19:10 | stephenfin | ||
| 12:19:11 | stephenfin | <stephenfin> sean-k-mooney: but that function has existed since it was removed in commit 92323586b5d03b31c18657ad64646a8ce3b8a742 | |
| 12:19:11 | stephenfin | <stephenfin> sean-k-mooney: so that means that code is not being called by anything yet. In either case, want to work on the fix or will I? | |
| 12:19:11 | stephenfin | <stephenfin> sean-k-mooney: fwiw, I'm running pytype over the code base to see what it reveals. That's one of two actual issues it has found so far | |
| 12:19:45 | sean-k-mooney | oh you pinged me earlier i did not see it | |
| 12:20:20 | jawad_axd | Hi all! Question is: I want to move couple of vms from one openstack evn(with local storage) to another openstack env(with ceph backend). Both vms have 40G root disk and 100G ephemeral disk.I can not take snapshot of instances since they are using ephemeral disks. How can I take backup of these vms with ephemeral disks and boot those two instance in new environment(with ceph backend)? | |
| 12:20:53 | stephenfin | The easy fix would be to re-add the function, but if no one has hit this yet (it's been there since 18.0.0) perhaps we can remove the call | |
| 12:20:58 | sean-k-mooney | stephenfin: the only place that is called is here https://github.com/openstack/nova/blob/b44b540fc70504f3869ef23022642095de0ea99e/nova/virt/libvirt/vif.py#L611-L612 | |
| 12:20:59 | stephenfin | I call look into it if you don't have time | |
| 12:21:10 | sean-k-mooney | so it the non os-vif code path for linux bridge | |
| 12:21:10 | stephenfin | yup | |
| 12:21:22 | sean-k-mooney | i think that was there just for nova-net? | |
| 12:21:26 | admin0 | sean-k-mooney, i redid libvirt-bin , restarted libvirtd and now i see it | |
| 12:21:27 | admin0 | under hvm | |
| 12:21:36 | admin0 | trying to define the instance now | |
| 12:22:02 | admin0 | defined :) | |
| 12:22:28 | sean-k-mooney | stephenfin: i think we can remvoe that branch | |
| 12:22:32 | stephenfin | sean-k-mooney: aha, yes https://github.com/openstack/nova/blob/b44b540fc70504f3869ef23022642095de0ea99e/nova/network/os_vif_util.py#L527-L528 | |
| 12:22:35 | stephenfin | we sure can | |
| 12:22:39 | stephenfin | I'll go kill it | |
| 12:22:43 | stephenfin | thanks, sean-k-mooney | |
| 12:22:48 | sean-k-mooney | stephenfin: anything that would be taking it would be broken at the moment anyway | |
| 12:23:14 | stephenfin | but first, 🏊+🥪 | |
| 12:23:28 | admin0 | ok .. nova is taken care of .. neutron next i think: -> Cannot get interface MTU on 'brqb3371910-95': No such device | |
| 12:23:40 | yoctozepto | admin0: cool | |
| 12:23:54 | yoctozepto | so it's a bug in libvirt losing hvm | |
| 12:24:17 | admin0 | thanks yoctozepto, sean-k-mooney .. can i mention a thank you when i put put this article on my openstackfaq.com site | |
| 12:24:19 | sean-k-mooney | admin0: you might want to take this to #openstack by the way | |
| 12:24:33 | admin0 | i did started on kolla aio https://www.openstackfaq.com/openstack-kolla-all-in-one/ | |
| 12:24:45 | sean-k-mooney | admin0: sure if you want | |
| 12:25:03 | yoctozepto | admin0: /me appreciates appreciation | |
| 12:25:03 | sean-k-mooney | jawad_axd: there is noe cross cloud migration support native to nova | |
| 12:25:30 | sean-k-mooney | jawad_axd: so your best bet would be to do the backup from within the vm. | |
| 12:25:50 | sean-k-mooney | e.g. snapshot root disk | |
| 12:26:11 | jawad_axd | @sean-k-mooney : Not migration, just want to take snapshot of instance with ephemeral disk? | |
| 12:26:11 | sean-k-mooney | create a vm on the ohter cloud by downloand an uploadign the snap shot. | |
| 12:26:28 | sean-k-mooney | boot it in rescue mode and then rsync the contentes fo the ephemeral volumes | |
| 12:26:49 | sean-k-mooney | you can take a snapshot but it will only snapshot the root disk | |
| 12:26:49 | admin0 | jawad_axd, or you can just boot 2 vms, rsync the data you need | |
| 12:27:05 | sean-k-mooney | that is basicly what i was saying | |
| 12:27:36 | yoctozepto | (this is the point where one suggests using dd over socat) | |
| 12:27:47 | jawad_axd | Snapshot doest work with ephemeral disk[E.g flavor has root disk and ephemral disk], while snapshot works with just root disk when in flavor ephemeral disk=0 . | |
| 12:28:16 | jawad_axd | ah ok | |
| 12:28:22 | sean-k-mooney | really it should still work with the root disk. at least i taught it did | |
| 12:28:33 | sean-k-mooney | i just wont snapshot the ephemeral disks | |
| 12:28:58 | jawad_axd | Its not even snapshotting the root disk. | |
| 12:29:01 | sean-k-mooney | ephemeral disk are intened for fast local storage by the way. not data persitency | |
| 12:29:17 | sean-k-mooney | i see | |
| 12:29:46 | sean-k-mooney | that should be filed as a bug. thre may be a reson for that but i dont know of one off the top of my head | |
| 12:30:04 | sean-k-mooney | then what i would do is as follows | |
| 12:30:18 | sean-k-mooney | boot a second vm on the new cloud | |
| 12:30:29 | sean-k-mooney | boot both vms in rescue mode | |
| 12:30:53 | sean-k-mooney | then use dd/rsync/clonzilla to copy all disks between the two | |
| 12:30:59 | sean-k-mooney | then unrescue the new vm | |
| 12:31:48 | sean-k-mooney | if you are moving to ceph i would also recomend replaceign the epmeeral disk with cinder volumens | |
| 12:32:59 | sean-k-mooney | if your user have permission on both clouds then this is something they could do themsese if its a private cloud followign instuctions | |
| 12:33:37 | jawad_axd | Ok, I am trying to understand your instructions. | |
| 12:34:36 | sean-k-mooney | basically on the new cloud create a vm with X root disk and n cinder volumns to repalce the ephemeral disks | |
| 12:35:37 | sean-k-mooney | then use rescue mode on both vms to boot the vms with a different imave with a disk imagain tool | |
| 12:36:20 | sean-k-mooney | and then use your tool of choice to copy the data form 1 vm to the other | |
| 12:36:35 | sean-k-mooney | then unrescue to reboot from the vm root disk | |
| 12:36:48 | jawad_axd | Okay.. | |
| 12:36:57 | jawad_axd | I am going to try it.Thank you! | |
| 12:38:18 | sean-k-mooney | there are some third part disater recovery tools that claim to be able to do this for you | |
| 12:38:25 | sean-k-mooney | but i dont really recommend using them | |
| 12:38:49 | sean-k-mooney | mainly because they tend to mess with the nova db in ways we would not support | |
| 12:39:13 | jawad_axd | ok | |
| 13:50:50 | gibi | efried: left comments in the first provider config patch: https://review.opendev.org/#/c/673341 I think my comments about the schema would be interesting to you too | |
| 13:59:17 | gibi | dustinc|PTO: ^^ | |
| 14:01:41 | openstackgerrit | Vladyslav Drok proposed openstack/nova master: Make removal of host from aggregate consistent https://review.opendev.org/696392 | |
| 14:08:05 | dustinc|PTO | Thanks gibi, I read through your comments but won’t be able to follow up until Tuesday. | |
| 14:09:00 | gibi | dustinc|PTO: no worries I also need to read your other patches in the series | |
| 14:09:25 | gibi | dustinc|PTO: have a nice PTO | |
| 14:09:47 | dustinc|PTO | Thanks :) | |
| 14:31:51 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/stein: DNM: try to trace bug 1859766 https://review.opendev.org/702700 | |
| 14:31:51 | openstack | bug 1859766 in OpenStack Compute (nova) stein "functional tests intermittently fails with "ReadOnlyFieldError: Cannot modify readonly field uuid"" [Undecided,New] https://launchpad.net/bugs/1859766 - Assigned to Balazs Gibizer (balazs-gibizer) | |
| 14:43:17 | lyarwood | Did anyone get to the bottom of `mysql: command not found` on the subnodes in the LM jobs? | |
| 14:46:18 | gibi | lyarwood: based on the bug https://bugs.launchpad.net/nova/+bug/1860021 we still have the problem | |
| 14:46:18 | openstack | Launchpad bug 1860021 in devstack "nova-live-migration fails 100% with "mysql: command not found" on subnode" [Undecided,In progress] - Assigned to Radosław Piliszek (yoctozepto) | |
| 14:46:56 | lyarwood | gibi: thanks, I couldn't find that for some reason. | |
| 14:47:02 | lyarwood | and yeah glance on the subnode is the issue | |
| 14:56:03 | efried | thanks gibi | |
| 14:56:36 | lyarwood | g-api on computes is just weird tbh, now if I could just figure out where the hell that's set | |
| 15:00:56 | lyarwood | ah it is devstack-gate | |
| 15:02:40 | efried | lyarwood: If we turn that off, won't it break grenade? | |
| 15:02:53 | efried | sean-k-mooney pointed out the commit from mriedem where that was introduced. | |
| 15:05:41 | lyarwood | efried: I'd have to check that commit for context, is there a rolling g-api upgrade test in grenade? | |
| 15:06:14 | efried | I have no earthly idea. Grenade has always been a swirling mystery to me. | |
| 15:07:00 | efried | lyarwood: https://github.com/openstack/devstack-gate/commit/183a58eab185c2e6f48afb4d99417d49e88c0f5a#diff-987768d02fbf43ebb74ad9a359222903R27 | |
| 15:07:03 | lyarwood | Well that's two of us | |