| opendevreview | Merged openstack/nova master: Server Rescue leads to Server ERROR state if original image is deleted https://review.opendev.org/c/openstack/nova/+/872385 | 01:07 |
|---|---|---|
| *** bauzas_ is now known as bauzas | 03:10 | |
| opendevreview | melanie witt proposed openstack/nova master: Support snapshot with ephemeral encryption for qcow2 https://review.opendev.org/c/openstack/nova/+/870937 | 04:07 |
| opendevreview | melanie witt proposed openstack/nova master: Support cross cell resize with ephemeral encryption for qcow2 https://review.opendev.org/c/openstack/nova/+/909595 | 04:07 |
| opendevreview | melanie witt proposed openstack/nova master: libvirt: Introduce support for raw with LUKS https://review.opendev.org/c/openstack/nova/+/884313 | 04:07 |
| opendevreview | melanie witt proposed openstack/nova master: libvirt: Introduce support for rbd with LUKS https://review.opendev.org/c/openstack/nova/+/889912 | 04:07 |
| spatel | any idea what is this error in nova_libvirt | 04:26 |
| spatel | 2024-07-08 04:25:59.489+0000: 2691: error : qemuDomainObjBeginJobInternal:968 : Timed out during operation: cannot acquire state change lock (held by monitor=remoteDispatchDomainCreateWithFlags) | 04:26 |
| spatel | This is happening after ceph failure. Now ceph is health but nova doesn't like to talk | 04:26 |
| *** bauzas_ is now known as bauzas | 05:20 | |
| opendevreview | melanie witt proposed openstack/nova master: Support snapshot with ephemeral encryption for qcow2 https://review.opendev.org/c/openstack/nova/+/870937 | 07:18 |
| opendevreview | melanie witt proposed openstack/nova master: Support cross cell resize with ephemeral encryption for qcow2 https://review.opendev.org/c/openstack/nova/+/909595 | 07:18 |
| opendevreview | melanie witt proposed openstack/nova master: libvirt: Introduce support for raw with LUKS https://review.opendev.org/c/openstack/nova/+/884313 | 07:18 |
| opendevreview | melanie witt proposed openstack/nova master: libvirt: Introduce support for rbd with LUKS https://review.opendev.org/c/openstack/nova/+/889912 | 07:18 |
| *** bauzas_ is now known as bauzas | 07:32 | |
| *** bauzas_ is now known as bauzas | 08:11 | |
| sean-k-mooney | gibi: regarding the time out how would you feel if i only tested one block size for iso | 10:19 |
| sean-k-mooney | either 64 * units.Ki, or 1 * units.Mi | 10:20 |
| sean-k-mooney | both would get the entiry header section in one read | 10:20 |
| sean-k-mooney | 512 is not as bad as 17 but its proably why it takes so long | 10:21 |
| sean-k-mooney | im going to try that locally and see how much of a speed up that gives | 10:22 |
| sean-k-mooney | we can see it timed out in chunk = wrapper.read(block_size) so its definetly in that read loop and making progress it just takes too long | 10:24 |
| sean-k-mooney | 5.5-5.9 second with the optimization vs 18.9-19 seconds without | 10:30 |
| sean-k-mooney | i think that is a reasonable trade off | 10:30 |
| sean-k-mooney | none of the logic im adding for the iso inspector cares about unaliged reads, that is just generic and shared with the other parsers so im personally happy with not testing it for these formats. | 10:32 |
| sean-k-mooney | im going to commit that locally and adress the other nits but if you have concerns with that let me know | 10:33 |
| *** ykarel__ is now known as ykarel | 10:54 | |
| gibi | sean-k-mooney: I'm fine skipping the unaligned read test for iso | 11:41 |
| sean-k-mooney | ack im just adressing the other nits ill have new version up in the next 30 mins or so | 11:43 |
| opendevreview | sean mooney proposed openstack/nova master: Reproduce iso regression with deep format inspection https://review.opendev.org/c/openstack/nova/+/923507 | 12:03 |
| opendevreview | sean mooney proposed openstack/nova master: Add iso file format inspector https://review.opendev.org/c/openstack/nova/+/923533 | 12:03 |
| sean-k-mooney | gibi: dansmith: ^ when your about those are functiontally the saem just with the nits addressed | 12:04 |
| sean-k-mooney | im going to stack with those now and confirm if this correctly sovles the rebuild regression | 12:04 |
| sean-k-mooney | i.e. boot form qcow usin a KNOPPIX or similar rescue iso to repair a vm | 12:05 |
| sean-k-mooney | s/usin/then using/ | 12:06 |
| gibi | sean-k-mooney: those two patches looks good to me. | 12:10 |
| sean-k-mooney | hopefully that resolves the timeouts but we will see. the fact that it was only timing out in the coverage job ment it was borderline before | 12:11 |
| sean-k-mooney | espicaly give the spread between tox py11 at 11 mins to py312 at 22 | 12:11 |
| sean-k-mooney | where that job runs obviouly has a non trivial impact on the job time | 12:13 |
| sean-k-mooney | gibi: ill comment on the patch but i was able t boot form an cirros qcow and then rescue with an alpine iso | 12:24 |
| sean-k-mooney | unrescue reverted back to cirros as expected | 12:24 |
| *** bauzas_ is now known as bauzas | 12:27 | |
| gibi | sounds good | 12:41 |
| andrewbonney | Hi. Any reviews on https://review.opendev.org/c/openstack/nova/+/919961 would be appreciated when people have time. Thanks | 12:41 |
| opendevreview | Robert Hoffmann proposed openstack/nova master: Fix: clean up volume attachments https://review.opendev.org/c/openstack/nova/+/923646 | 13:08 |
| spatel | Any idea on nova_libvirt error : qemuDomainObjBeginJobInternal:968 : Timed out during operation: cannot acquire state change lock (held by monitor=init-instance-00000d99) | 13:37 |
| *** haleyb|out is now known as haleyb | 14:00 | |
| *** bauzas_ is now known as bauzas | 14:25 | |
| elodilles | sean-k-mooney dansmith : should we wait for any other patch to merge, or is it OK if I generate now stable release patches for nova? (so that we release the fixes for bug/2059809) | 14:43 |
| dansmith | elodilles: I think fungi would say yes, release | 14:44 |
| elodilles | ACK, then I'm preparing the patch | 14:44 |
| fungi | yeah, the sooner the better | 14:44 |
| dansmith | elodilles, fungi: bauzas is out this week, will we need an ack? | 14:45 |
| fungi | the release team is able to override/bypass that requirement, if they're willing to | 14:46 |
| dansmith | cool | 14:46 |
| elodilles | dansmith: https://review.opendev.org/c/openstack/releases/+/923656 | 14:58 |
| elodilles | Uggla auniyal : ping, as you are also release liaisons ^^^ o:) | 14:58 |
| elodilles | release jobs (validation and list-changes) logs should be available in ~10-20 mins | 15:00 |
| dansmith | elodilles: hmm, are those the right hashes for each release? | 15:00 |
| dansmith | seems like A and C are reversed, unless I'm missing something | 15:00 |
| dansmith | wait, maybe not | 15:01 |
| dansmith | okay yeah, mis-copy-paste I think, nevermind | 15:01 |
| *** bauzas_ is now known as bauzas | 15:05 | |
| mnaser | hi nova friends. https://review.opendev.org/c/openstack/nova/+/923298/2 -- can someone look at this unmaintained/zed stack, it looks good to me so far and would be nice to get going into the gate | 16:14 |
| mnaser | (CVE stuff) | 16:14 |
| dansmith | mnaser: note that nova-core doesn't even have +2 on that branch because it's unmaintained | 16:16 |
| dansmith | looks like it has the only tweak we made after the first push, so I'd assume that if it works for you it's good | 16:18 |
| mnaser | yeah I got that in | 16:20 |
| mnaser | well, elodilles did :) | 16:20 |
| JayF | mnaser: feel free to add me as reviewer for any security backports in unmaintained/ that need additional review. | 16:25 |
| JayF | mnaser: I have the one patch you've indicated opened in my window, will have a look | 16:25 |
| JayF | Speaking of; mnaser has an important fix for the Ironic driver which will need to land in master + go back to Caracal, and is probably bad enough that I'd hope we could get it in a caracal point release: https://review.opendev.org/c/openstack/nova/+/923530 -- basically Ironic+Nova integrated environments with complex networking configurations will not have their metadata generated properly | 16:30 |
| dansmith | JayF: +2d | 16:33 |
| dansmith | looks simple enough, maybe gmann or melwitt could hit it for you | 16:33 |
| JayF | dansmith: thank you very much \o/ (for this and for fixing the security issues over the last weeks) | 16:33 |
| melwitt | +Wd | 16:36 |
| JayF | awesome, I'll get a backport in once it lands | 16:37 |
| JayF | I do have an item on my todo to follow up and make sure all those tests use iter([obj1,obj2]) to catch this kinda issue | 16:38 |
| dansmith | sean-k-mooney: so the coverage job is still failing on your test_iso9660 | 17:19 |
| dansmith | I don't really know why though | 17:19 |
| dansmith | because it reads 10M 17 bytes at a time maybe? does it need to be that big? | 17:21 |
| sean-k-mooney | didnt i fixt that already https://review.opendev.org/c/openstack/nova/+/923533/8/nova/tests/unit/image/test_format_inspector.py#211 | 17:23 |
| sean-k-mooney | dansmith: https://review.opendev.org/c/openstack/nova/+/923507/3 still need a +w by the way | 17:23 |
| sean-k-mooney | that why its not currently in the ci | 17:24 |
| dansmith | oh yeah, I was looking at the wrong report | 17:24 |
| dansmith | that's why duh | 17:24 |
| dansmith | cool | 17:24 |
| sean-k-mooney | but ya the 32KB system area makes it time out with 7 or 512 bytye blocks | 17:24 |
| dansmith | I just changed to 128k locally and it works fine at that size, so no real need for it to be larger I think | 17:25 |
| sean-k-mooney | btu the 64k and 1mb block size do it in a single read | 17:25 |
| dansmith | but as long as it passes now all good, it's only temporary anyway | 17:25 |
| sean-k-mooney | well its reading 34KB total so the 64K read size is enough | 17:25 |
| dansmith | in the gate now | 17:27 |
| dansmith | totes forgot about that repro patch below :D | 17:27 |
| sean-k-mooney | we could perhaps replace the 7 bytes with 7k later to have an unaligend read test be a little fater but ya for now it should be fine | 17:27 |
| sean-k-mooney | ok those are going to take a while to go though the gate so ill likely take this time to cook dinner and ill check back after | 17:29 |
| dansmith | ++ | 17:30 |
| dansmith | hmm, might have a requirements conflict: gabbi 3.0.0 depends on jsonpath-rw-ext>=1.0.0 | 18:30 |
| dansmith | just failed in gate | 18:30 |
| dansmith | okay might've just been the requirements change that was in front of it | 18:34 |
| dansmith | that failed (with multiple issues) and the nova restart passed that point I think | 18:34 |
| frickler | that sounds like pypi is having some issues | 18:40 |
| dansmith | maybe | 18:41 |
| clarkb | those sorts of failures can occur when we get served old indexes from pypi | 18:43 |
| dansmith | okay the requirements patch wasn't changing either of those packages, and one other job on it failed in a different way (but the nova one failed the same way the nova patch behind it did) | 18:48 |
| opendevreview | Merged openstack/nova master: Reproduce iso regression with deep format inspection https://review.opendev.org/c/openstack/nova/+/923507 | 18:56 |
| opendevreview | Merged openstack/nova master: Add iso file format inspector https://review.opendev.org/c/openstack/nova/+/923533 | 20:29 |
| sean-k-mooney | sweet ^ ill start backporting those | 20:32 |
| dansmith | \o/ | 20:45 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!