| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-05-04 | |||
| 11:27:00 | sean-k-mooney | dpawlik: you will need to update the docs for the new image metadata item but i belive that would be in glance's docs not novas | |
| 11:28:39 | dpawlik | sean-k-mooney: hey. So better when I add here DocImpact and ask glance team to add it in its docs | |
| 11:30:01 | sean-k-mooney | dpawlik: https://wiki.openstack.org/wiki/Documentation/DocImpact | |
| 11:30:38 | dpawlik | sean-k-mooney: thx | |
| 11:44:23 | openstackgerrit | Daniel Pawlik proposed openstack/nova master: Added kvm_hidden feature https://review.openstack.org/459753 | |
| 11:45:07 | sean-k-mooney | sdague: hi, 2 quick quetstions regarding the new systemd workflow for devstack will the screen workflow be tested and maintined going forword? also will devstack now work across reboots again? | |
| 11:49:44 | sdague | sean-k-mooney: screen goes away in Queens | |
| 11:50:03 | sdague | http://lists.openstack.org/pipermail/openstack-dev/2017-May/116301.html | |
| 11:50:15 | sdague | I'm not sure what works across reboots means | |
| 11:50:37 | sdague | mdbooth: that's entirely possible... | |
| 11:50:51 | mdbooth | I don't think the journal itself drops messages | |
| 11:51:02 | sdague | mdbooth: we did disable it now, so we're out of that woods for the moment | |
| 11:51:09 | mdbooth | Yeah, makes sense | |
| 11:51:20 | mdbooth | One to investigate though, for sure | |
| 11:53:37 | sean-k-mooney | sdague: ok i might need to make changes to our devstack plugins/CIs. | |
| 11:54:11 | sean-k-mooney | i think they will work but they may not | |
| 11:54:32 | sdague | sean-k-mooney: yeh, that thread has some common issues people hit | |
| 11:56:21 | sean-k-mooney | sdague: i guess il have to give it a fair shot but honestly the fact that the process ran in screen sessions was one of the things i liked about devstack as it made it easy to wrap your head around when you were new to openstack | |
| 11:57:06 | kashyap | sean-k-mooney: It isn't any more? I haven't run a DevStack run in a couple of weeks | |
| 11:57:12 | sdague | sean-k-mooney: ok, an equally large number of people have stated that screen was the most confusing thing ever | |
| 11:57:24 | sdague | but, regardless, there are other reasons we need this change | |
| 11:57:35 | kashyap | I've seen the systemd / 'screen' related change whoosh by. /me looks closer... | |
| 11:57:50 | kashyap | Hmm, it's this patch: "make USE_SCREEN=False imply USE_SYSTEMD=True" | |
| 11:58:50 | sean-k-mooney | yes i swapped the flag back and it still failed to stack so my local failure wasnt related to the systemd change so im going to start with a clean environment again | |
| 11:59:14 | sean-k-mooney | kashyap: ^ | |
| 11:59:31 | kashyap | sean-k-mooney: What is the implication of that patch? Still reading this - https://review.openstack.org/#/c/460062/ | |
| 11:59:36 | sean-k-mooney | sdague: screen not being maintianed for one | |
| 12:00:17 | sean-k-mooney | kashyap: the openstack service deployed by devstack are now executed via systemd instead of in screen with there logs sent to the journal | |
| 12:01:11 | kashyap | sean-k-mooney: Hmm. I like systemd journal logs (for its nice 'query-ability'), but I miss firing up a quick `screen -r` to flip through all the services | |
| 12:02:23 | sean-k-mooney | kashyap: that and quickly running a service in a debugger or viewing the live logs though journalctl supports that too. | |
| 12:02:53 | kashyap | Hmm. What is the rationale for the "feature"? | |
| 12:03:02 | kashyap | I'd agree with mtreinish there, where he says: | |
| 12:03:08 | kashyap | "I think expanding this comment a bit to say the intent here would have been nice, since I got confused by this at first." | |
| 12:03:40 | sean-k-mooney | kashyap: make test closer to production | |
| 12:04:05 | sdague | kashyap: all sorts of stuff on the mailing list over the last month, I'd rather not just reiterate all that here | |
| 12:04:25 | kashyap | Then we can _really_ claim: "If it works on DevStack, it's all kosher" :P | |
| 12:04:39 | sean-k-mooney | kashyap: haha sure... | |
| 12:04:46 | kashyap | sdague: Yeah, sorry, I'd go catch up the thread. I did see the mail, but spent all of 2 seconds on it. | |
| 12:06:12 | sean-k-mooney | sdague: if devstack works across reboots again this would be cool sideffect but it might cause some people to swap to other tools for dev. | |
| 12:07:13 | openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova master: Use check_string_length from oslo_utils https://review.openstack.org/286390 | |
| 12:07:33 | kashyap | sean-k-mooney: But _does_ it work cope with reboots with this patch? If so, that'd be nice | |
| 12:07:35 | sean-k-mooney | sdague: anyway thanks for trying to keep it working for plugins. at least 1 of your plugins seeam to be working fine with the change ill check the otheres when i get time or delegate to the rest of my team :) | |
| 12:07:56 | sean-k-mooney | kashyap: dont know havent tried it | |
| 12:08:37 | sean-k-mooney | i miss rejoin stack for that abiltiy it was really useful for demoing stuff but it also had lots of issues if you didnt know how it worked | |
| 12:10:37 | openstackgerrit | Matthew Booth proposed openstack/nova master: WIP: Automatically revert resize which fails on destination https://review.openstack.org/462521 | |
| 12:15:26 | sdague | sean-k-mooney: you are going to need to explain what your reboot workflow is, because I only reboot my devstack environments to get rid of cruft. :) So I honestly am not sure what "works with reboots" means | |
| 12:17:06 | sean-k-mooney | sdague: before rejoin_stack.sh was removed you just had to start ovs,mysql,apache and rabbit then run rejoin_stack.sh and your devstack enviorment came back up | |
| 12:17:44 | sdague | sudo systemctl start devstack@* post boot might do that | |
| 12:17:44 | sean-k-mooney | sdague: this was useful for demos as you could prepera an enviorment and snapshot it in a vm and give it to someone to try | |
| 12:18:20 | sean-k-mooney | sdague: yep you are enabling the service by default i think so it might just work aslong as you dont run unstack.sh | |
| 12:18:28 | sean-k-mooney | sdague: or clean.sh i guess | |
| 12:18:30 | sdague | yep | |
| 12:18:52 | sean-k-mooney | anyway it might just be a nice sideffect to get that fucntionality back | |
| 12:19:28 | sean-k-mooney | i used to jsut run in offline mode instad but that ment you still had to recreated the vms/netwoks/flavour ... | |
| 12:19:41 | sean-k-mooney | thats not hard jsut extra steps | |
| 12:21:40 | alex_xu_ | johnthetubaguy: sdague sfinucan appreciate if you can help this two easy api-ref changes https://review.openstack.org/461998 https://review.openstack.org/458738 | |
| 12:22:20 | sdague | sean-k-mooney: yeh, let me know how it works. If we need some tweaks to make that use can solid for you, that would be cool | |
| 12:22:41 | sdague | there was so much hacktastic around getting screen to be reliable, I look forward to that code being deleted with fire | |
| 12:26:27 | sshnaidm | has anybody seen nova errors like: AttributeError: 'Instance' object has no attribute '_save_tags' ? | |
| 12:33:20 | johnthetubaguy | sshnaidm: I think I saw a bug and patch up for that somewhere | |
| 12:34:30 | johnthetubaguy | sshnaidm: I think gibi has this patch up: https://review.openstack.org/#/c/450265/ | |
| 12:34:57 | johnthetubaguy | dansmith: I wasn't sure about the object change in this one actually https://review.openstack.org/#/c/450265/ | |
| 12:36:48 | sshnaidm | johnthetubaguy, thanks, so as I understand it's not critical, right? | |
| 12:37:38 | johnthetubaguy | sshnaidm: I believe so | |
| 12:47:07 | sou | Hey, I am stuck with an issue for past 2 days | |
| 12:47:15 | sou | I am trying a nova rebuild | |
| 12:47:20 | sou | and it fails all the time | |
| 12:47:44 | sou | with error: ERROR oslo_messaging.rpc.dispatcher [req-b5730d47-3c34-4949-9836-01139238f829 8a205b2aa31c4c1185a417c6e99e852c fcb05e141ede4164ab6db9ad0724a08e - - -] Exception during message handling: id | |
| 12:48:24 | sou | nothing else in logs | |
| 12:48:45 | sou | the exception is generated from /usr/lib/python2.7/site-packages/glanceclient/openstack/common/apiclient/base.py | |
| 12:48:52 | sou | line 494, in __getattr__ | |
| 12:49:05 | sou | any one else has faced similar issue? | |
| 12:49:16 | sou | I am using Kilo version | |
| 12:49:43 | sou | compute nodes uses KVM and images are lvm | |
| 12:52:21 | johnthetubaguy | sou: please try #openstack, see channel topic | |
| 12:53:31 | sou | Thanks @johnthetubadguy | |
| 13:26:24 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for server-tags instead of stevedore https://review.openstack.org/461645 | |
| 13:26:24 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for os-interface endpoint instead of stevedore https://review.openstack.org/459979 | |
| 13:26:25 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for server-migrations endpoint instead of stevedore https://review.openstack.org/461646 | |
| 13:30:09 | abhi89 | johnthetubaguy: can we back port this fix https://review.openstack.org/#/c/460305/ to stable/ocata.. we were looking to have this change in this release in our product.. | |
| 13:30:47 | johnthetubaguy | abhi89: I think that should be OK | |
| 13:31:07 | dansmith | johnthetubaguy: yes, as you said, and as convention in the code, the empty save handler is the right approach | |
| 13:33:01 | abhi89 | johnthetubaguy: grt.. i will cherry pick the fix to ocata.. | |
| 13:33:24 | gibi | mriedem: hi! I could not attend on the nova meeting today, I left small notes on the agenda | |
| 13:33:32 | mriedem | gibi: ok thanks | |
| 13:45:34 | mriedem | gibi: do we need to call out https://review.openstack.org/#/c/401992/ when it's approved? | |
| 13:46:54 | mriedem | melwitt: sdague: dansmith: looks like the use_journal=True thing was causing the cells job failures http://status.openstack.org/elastic-recheck/#1580728 | |
| 13:47:03 | mriedem | we should still get https://review.openstack.org/#/c/461912 in though | |
| 13:48:44 | dansmith | mriedem: okay, I've yet to see my devstack patch fail with journald remoed | |
| 13:48:54 | dansmith | well, fail for a related reason anyway | |
| 13:49:13 | abhi89 | johnthetubaguy: please review cherry picked fix https://review.openstack.org/#/c/462561/ for stable/ocata | |
| 13:56:10 | mriedem | nova meeting in 4 minutes in #openstack-meeting | |
| 14:32:48 | mtreinish | kashyap: fwiw, I was talking about the grenade line there. It's not 100% obvious that it's there for making sure systemd isn't used on grenade | |
| 14:33:05 | mtreinish | I had to think it through out loud to understand the logic in that if | |
| 14:33:07 | kashyap | mtreinish: Noted. :-) | |
| 14:33:40 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform aggregate.updateprop notification https://review.openstack.org/462576 | |
| 14:34:56 | nammo | Hi How can I remove a RetryFilter? | |
| 14:35:12 | nammo | I removed it from nova.conf but it still exists | |
| 14:37:49 | bauzas | manjeets__: pull it out from enabled_filters | |
| 14:38:05 | bauzas | oops | |