| imacdonn | Can anyone explain why ImageDownloadFailed uses "_msg_fmt", where all the others use "message"? https://github.com/openstack/cinder/blob/master/cinder/exception.py#L731 | 00:20 |
|---|---|---|
| imacdonn | asking because _msg_fmt doesn't seem to work for me ... wonder if I'm missing something ... | 00:22 |
| *** woojay has quit IRC | 00:25 | |
| *** Liang__ has joined #openstack-cinder | 00:25 | |
| *** Liang__ has quit IRC | 00:27 | |
| *** lixiaoy1 has joined #openstack-cinder | 00:29 | |
| *** erlon has joined #openstack-cinder | 00:33 | |
| lixiaoy1 | jungleboyj: ping | 00:35 |
| *** brinzhang has joined #openstack-cinder | 00:48 | |
| openstackgerrit | iain MacDonnell proposed openstack/cinder master: Set message property in ImageDownloadFailed https://review.openstack.org/625135 | 00:58 |
| *** jcosmao has quit IRC | 01:04 | |
| openstackgerrit | LisaLi proposed openstack/cinder-specs master: Driver reinitialization after failure https://review.openstack.org/624853 | 01:05 |
| openstackgerrit | Merged openstack/cinder master: Add secret=true to fixed_key configuration parameter https://review.openstack.org/621686 | 01:17 |
| *** psachin has joined #openstack-cinder | 01:30 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 01:39 | |
| *** Dinesh_Bhor has quit IRC | 01:43 | |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume create policy https://review.openstack.org/617542 | 01:59 |
| *** Dinesh_Bhor has joined #openstack-cinder | 02:01 | |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume create policy https://review.openstack.org/617542 | 02:03 |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume create policy https://review.openstack.org/617542 | 02:05 |
| openstackgerrit | LisaLi proposed openstack/cinder master: Driver reinitialization after failure https://review.openstack.org/618702 | 02:08 |
| *** bhagyashris has joined #openstack-cinder | 02:11 | |
| *** mhen has quit IRC | 02:26 | |
| *** sapd1_ has joined #openstack-cinder | 02:27 | |
| *** sapd1 has quit IRC | 02:29 | |
| *** mhen has joined #openstack-cinder | 02:30 | |
| openstackgerrit | Yingxin Cheng proposed openstack/os-brick master: Support RSD scenario of nvme connector https://review.openstack.org/620250 | 02:35 |
| *** dave-mccowan has joined #openstack-cinder | 02:38 | |
| *** itlinux has joined #openstack-cinder | 02:40 | |
| *** hedvig_01 has quit IRC | 02:42 | |
| *** Liang__ has joined #openstack-cinder | 02:53 | |
| *** Liang__ is now known as LiangFang | 02:56 | |
| openstackgerrit | Dave McCowan proposed openstack/cinder stable/rocky: Add secret=true to fixed_key configuration parameter https://review.openstack.org/625139 | 02:56 |
| openstackgerrit | Dave McCowan proposed openstack/cinder stable/queens: Add secret=true to fixed_key configuration parameter https://review.openstack.org/625140 | 02:57 |
| *** LiangFang has quit IRC | 03:02 | |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume update/del policy https://review.openstack.org/617543 | 03:11 |
| openstackgerrit | TommyLike proposed openstack/cinder master: Increase length limit on image metadata content https://review.openstack.org/586737 | 03:14 |
| *** Liang__ has joined #openstack-cinder | 03:23 | |
| *** Liang__ is now known as LiangFang | 03:23 | |
| *** lbragstad has joined #openstack-cinder | 03:50 | |
| *** lbragstad has quit IRC | 03:51 | |
| *** Dinesh_Bhor has quit IRC | 03:52 | |
| *** brinzh has joined #openstack-cinder | 04:02 | |
| *** dave-mccowan has quit IRC | 04:03 | |
| *** brinzhang has quit IRC | 04:06 | |
| *** udesale has joined #openstack-cinder | 04:10 | |
| *** psachin has quit IRC | 04:10 | |
| *** lbragstad has joined #openstack-cinder | 04:13 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 04:14 | |
| *** psachin has joined #openstack-cinder | 04:28 | |
| whoami-rajat | enriquetaso: Since 'backup' is a class reference, you can try the following way : | 04:37 |
| whoami-rajat | use pdb in openstack client | 04:37 |
| whoami-rajat | view the object with the following syntax: | 04:37 |
| whoami-rajat | from pprint import pprint; pprint(vars(backup)) | 04:37 |
| *** woojay has joined #openstack-cinder | 04:41 | |
| *** _alastor_ has quit IRC | 04:54 | |
| *** bhagyashris has quit IRC | 05:01 | |
| *** Dinesh_Bhor has quit IRC | 05:01 | |
| *** _alastor_ has joined #openstack-cinder | 05:02 | |
| openstackgerrit | Merged openstack/cinder master: Synchronize all LVM operations https://review.openstack.org/624809 | 05:16 |
| *** Dinesh_Bhor has joined #openstack-cinder | 05:23 | |
| *** _alastor_ has quit IRC | 05:34 | |
| *** dklyle has joined #openstack-cinder | 05:38 | |
| openstackgerrit | Merged openstack/cinder master: Set message property in ImageDownloadFailed https://review.openstack.org/625135 | 05:40 |
| openstackgerrit | iain MacDonnell proposed openstack/cinder stable/rocky: Set message property in ImageDownloadFailed https://review.openstack.org/625148 | 05:49 |
| *** lpetrut has joined #openstack-cinder | 06:13 | |
| *** bhagyashris_ has joined #openstack-cinder | 06:35 | |
| *** rcernin has quit IRC | 07:03 | |
| *** pcaruana has joined #openstack-cinder | 07:12 | |
| *** Luzi has joined #openstack-cinder | 07:13 | |
| *** erlon has quit IRC | 07:23 | |
| openstackgerrit | Ryan Liang proposed openstack/cinder stable/ocata: [Unity] Add `force detach` support https://review.openstack.org/625165 | 07:32 |
| *** dpawlik has joined #openstack-cinder | 07:38 | |
| openstackgerrit | Yong Huang proposed openstack/cinder stable/ocata: Empty option value maybe cause Unity driver failed to initialize https://review.openstack.org/625167 | 07:42 |
| *** Dinesh_Bhor has quit IRC | 07:48 | |
| *** psachin has quit IRC | 07:48 | |
| *** lixiaoy1 has quit IRC | 07:54 | |
| *** lpetrut has quit IRC | 07:58 | |
| openstackgerrit | Merged openstack/cinder master: Automate generation of api-ref samples https://review.openstack.org/619676 | 08:03 |
| openstackgerrit | Yingxin Cheng proposed openstack/cinder master: [wip] Cinder RSD Driver https://review.openstack.org/621465 | 08:06 |
| *** markvoelker has joined #openstack-cinder | 08:16 | |
| *** imacdonn has quit IRC | 08:24 | |
| *** imacdonn has joined #openstack-cinder | 08:24 | |
| *** lixiaoy1 has joined #openstack-cinder | 08:29 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 08:33 | |
| openstackgerrit | Yong Huang proposed openstack/cinder stable/pike: Update unity tests to use our test base https://review.openstack.org/625181 | 08:41 |
| *** lixiaoy1 has quit IRC | 08:55 | |
| openstackgerrit | zhufl proposed openstack/cinder master: api-ref: volume_type_access should be array instead of object https://review.openstack.org/625186 | 09:00 |
| *** brinzh has quit IRC | 09:06 | |
| *** erlon has joined #openstack-cinder | 09:13 | |
| *** arne_wiebalck has joined #openstack-cinder | 09:20 | |
| *** erlon has quit IRC | 09:22 | |
| openstackgerrit | Francois Deppierraz proposed openstack/cinder master: cinder-volume: Stop masking IOError different than ENOSPC https://review.openstack.org/612393 | 09:23 |
| *** Dinesh_Bhor has quit IRC | 09:46 | |
| *** dcdamien has joined #openstack-cinder | 09:50 | |
| *** luizbag has joined #openstack-cinder | 09:55 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of API versions api-ref samples https://review.openstack.org/620033 | 09:56 |
| *** erlon has joined #openstack-cinder | 10:21 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of volumes api-ref samples https://review.openstack.org/621478 | 10:31 |
| *** Dinesh_Bhor has joined #openstack-cinder | 10:35 | |
| *** markvoelker has quit IRC | 10:36 | |
| *** Dinesh_Bhor has quit IRC | 10:36 | |
| *** markvoelker has joined #openstack-cinder | 10:37 | |
| *** yongwc has joined #openstack-cinder | 10:38 | |
| *** e0ne has joined #openstack-cinder | 10:40 | |
| *** markvoelker has quit IRC | 10:41 | |
| openstackgerrit | Brin Zhang proposed openstack/cinder master: api-ref: add qos_specs_id to show default volume type response https://review.openstack.org/625208 | 10:54 |
| *** yongwc has quit IRC | 10:54 | |
| *** bhagyashris_ has quit IRC | 10:59 | |
| *** LiangFang has quit IRC | 11:00 | |
| *** yongwc_ has joined #openstack-cinder | 11:09 | |
| *** markvoelker has joined #openstack-cinder | 11:16 | |
| openstackgerrit | Merged openstack/cinder master: Fix A/A 'resource_backend' when scheduling volumes https://review.openstack.org/625052 | 11:18 |
| *** arne_wiebalck has quit IRC | 11:24 | |
| *** arne_wiebalck_ is now known as arne_wiebalck | 11:24 | |
| yongwc_ | Anyone would like do a review for https://review.openstack.org/#/c/562977/, which is a driver for inspur as13000 storage system, thanks. | 11:27 |
| *** gary_perkins has quit IRC | 11:37 | |
| *** Luzi has quit IRC | 11:44 | |
| *** gary_perkins has joined #openstack-cinder | 11:50 | |
| *** yongwc_ has quit IRC | 11:57 | |
| *** udesale has quit IRC | 12:11 | |
| *** tpsilva has joined #openstack-cinder | 12:14 | |
| *** pcaruana has quit IRC | 12:21 | |
| *** pcaruana has joined #openstack-cinder | 12:22 | |
| *** pcaruana is now known as pcaruana|intw| | 12:25 | |
| *** _alastor_ has joined #openstack-cinder | 12:35 | |
| *** abishop has joined #openstack-cinder | 13:04 | |
| *** kaisers_ has joined #openstack-cinder | 13:09 | |
| *** _alastor_ has quit IRC | 13:14 | |
| *** dcdamien has quit IRC | 13:18 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Allow to use _max qos option together with per_gb https://review.openstack.org/625058 | 13:18 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinderlib functional tests https://review.openstack.org/620671 | 13:45 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of API versions api-ref samples https://review.openstack.org/620033 | 13:52 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of volumes api-ref samples https://review.openstack.org/621478 | 13:53 |
| *** mriedem has joined #openstack-cinder | 13:56 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of volumes api-ref samples https://review.openstack.org/621478 | 13:56 |
| *** pcaruana|intw| has quit IRC | 14:05 | |
| dansmith | jungleboyj: smcginnis: mriedem: http://logs.openstack.org/65/621465/18/check/nova-multiattach/93859c1/logs/screen-c-vol.txt.gz?#_Dec_14_09_07_37_269750 | 14:12 |
| mriedem | sob | 14:12 |
| dansmith | waited >60s for the lock, which tells me it's not gonna help | 14:12 |
| dansmith | so when we discussed this earlier, I think it was smcginnis that said "so if lvm is taking a long time there's nothing cinder can do about it", which might be true, | 14:13 |
| dansmith | but I feel like we've got to do something here | 14:13 |
| dansmith | either change the device we're configuring lvm to use so it can be faster, | 14:13 |
| dansmith | or use the long-call rpc stuff so we just wait until it's really done.. or something. | 14:13 |
| mriedem | there could be other places in cinder that are hitting lvchange? but i don't think i saw any other obvious callers | 14:14 |
| smcginnis | I wonder if this is some kind of regression in LVM. Strange that we are hitting it more often all of a sudden. | 14:14 |
| mriedem | we did just recently upgrade tempest-full and friends to bionic, | 14:14 |
| dansmith | mriedem: this is locking all lvm ops, not just lvchange | 14:14 |
| mriedem | but the nova-multiattach job is legacy so it's using xenail | 14:14 |
| mriedem | *xenial | 14:14 |
| dansmith | smcginnis: several of these bugs have been around for a year in some form | 14:15 |
| mriedem | i support the long_rpc_timeout on attachment_update/os-initialize_connection | 14:15 |
| mriedem | that should help with a couple of the bugs | 14:15 |
| smcginnis | dansmith: Yeah, it seems like it's something that kind of comes and goes though. | 14:15 |
| dansmith | yeah, it sucks to solve it with that just from a user perspective, but at least it would make it not time out 1s before it actually completes | 14:16 |
| mriedem | right | 14:16 |
| dansmith | smcginnis: yeah, which smells so much like it's us a and not them to me, but.. I know. | 14:16 |
| dansmith | it also seems plausible that it's just an artifact of slow nodes with lots of stuff going on and a loop-mounted lvm backing store that is just getting hamstrung by terrible IO and buffer cache traffic | 14:17 |
| dansmith | like I wonder if we could get away with putting the loop on tmpfs to avoid doing as much IO | 14:18 |
| smcginnis | Hmm, took over 30 seconds just calling lvs: http://logs.openstack.org/65/621465/18/check/nova-multiattach/93859c1/logs/screen-c-vol.txt.gz?#_Dec_14_09_07_37_267680 | 14:18 |
| dansmith | smcginnis: right, see that's why I wonder if it's really block refreshes that are slowing it down.. like lvs does a scan that includes an iscsi device we're taking down | 14:19 |
| dansmith | you know, | 14:19 |
| dansmith | I wonder if we could/should be configuring lvm.conf to be super selective in what it scans | 14:19 |
| mriedem | does dstat around the time of these show anything | 14:20 |
| mriedem | ? | 14:20 |
| dansmith | mriedem: I have thought about doing that, but I have to plot a csv file I think so I haven't yet | 14:20 |
| mriedem | http://logs.openstack.org/65/621465/18/check/nova-multiattach/93859c1/logs/screen-dstat.txt.gz#_Dec_14_09_07_37_071861 | 14:20 |
| mriedem | http://logs.openstack.org/65/621465/18/check/nova-multiattach/93859c1/logs/dstat-csv_log.txt.gz | 14:20 |
| mriedem | i have a hard time reading this thing | 14:21 |
| dansmith | yeah | 14:21 |
| dansmith | ah, devstack is already setting an lvm filter | 14:21 |
| smcginnis | I thought it was, but can't find it in the collected artifacts. | 14:22 |
| dansmith | it doesn't collect it afaict, but it does log something, hang on | 14:22 |
| dansmith | oh | 14:22 |
| dansmith | http://logs.openstack.org/65/621465/18/check/nova-multiattach/93859c1/logs/devstacklog.txt.gz#_2018-12-14_08_45_31_687 | 14:22 |
| dansmith | it seems to still include everything after the loops | 14:23 |
| dansmith | no that's reject | 14:23 |
| dansmith | nevermind | 14:23 |
| dansmith | so just the loop1 and loop2 | 14:23 |
| dansmith | which should do it I think | 14:23 |
| smcginnis | Yeah, looks like that should lock it down enough. | 14:23 |
| dansmith | unless some iscsi device has loop in the name, but seems unlikely :) | 14:24 |
| smcginnis | Hah, let's hope not. | 14:24 |
| smcginnis | Man, 63.103s on an lvs just seems like something is very, very wrong. | 14:25 |
| eharney | i think if you add a number of "--debug" flags to "lvs", you'll see timing info on what it's doing in syslog | 14:25 |
| mriedem | i was wondering about that too, | 14:26 |
| smcginnis | I suppose we could add that temporarily as a debugging aid. | 14:26 |
| eharney | running "sudo lvs --debug --debug --debug --debug --debug --debug" on my machine shows it | 14:26 |
| mriedem | if there was a way to get some kind of data capture when a command takes over a given time | 14:26 |
| mriedem | but i suppose you need to know what's going on while it's running, not immediately after | 14:26 |
| dansmith | ouch, do we really need 24G of backing space for this? | 14:27 |
| *** udesale has joined #openstack-cinder | 14:27 | |
| dansmith | because with an 8G guest or whatever, we could be generating a lot of dirty pages | 14:28 |
| smcginnis | Yep :/ | 14:28 |
| smcginnis | We were running out of space on some things. | 14:28 |
| smcginnis | Which does seem ridiculous. | 14:28 |
| dansmith | so, in the past, loop had a really high overhead | 14:29 |
| dansmith | recently they added directio support | 14:29 |
| dansmith | so I wonder if we're new enough to try that and/or if we need to ask for it or get it directly | 14:30 |
| smcginnis | Oh? Any docs on that dansmith? Or something you want to try throwing a patch up for? | 14:30 |
| smcginnis | Sounds promising. | 14:30 |
| dansmith | smcginnis: gotta read first | 14:30 |
| dansmith | this multiattach job is 4.4.0.. mriedem you said that's xenial, and the oldest rght? | 14:30 |
| smcginnis | Or at least 1 idea more than the 0 I have. | 14:30 |
| smcginnis | I don't believe we are running any current jobs on anything older than Xenial. | 14:31 |
| *** pcaruana has joined #openstack-cinder | 14:31 | |
| dansmith | so, it says that it defaults on (directio) but my 4.9 machine has a couple loops that don't have directio turned on, so.. maybe worth a devstack patch to force it on and show the result and see | 14:32 |
| mriedem | master won't run anything older than xenial | 14:34 |
| mriedem | the zuulv3 jobs like tempest-full are running bionic as of this week | 14:34 |
| mriedem | nova-multiattach is not using zuulv3 so it's still xenial | 14:35 |
| mriedem | bbiab | 14:35 |
| dansmith | https://review.openstack.org/#/c/625269/ | 14:40 |
| dansmith | smcginnis: aside from figuring this specific issue out, what do you think about using the long-rpc stuff for calls from api->volume that do synchronous things that could take a while like this? | 14:41 |
| smcginnis | dansmith: I suppose that at least will help with the timeouts. Still kills me that these calls even need to wait that long, but at least right now that appears to be an external constraint that we're just going to have to deal with. | 14:42 |
| dansmith | and/or should I explain what that means? :) | 14:42 |
| dansmith | smcginnis: right, we should try to fix the acute problem for sure, | 14:43 |
| dansmith | it's just that any time you're doing something synchronous that is manipulating stuff like this you're open to it | 14:43 |
| dansmith | smcginnis: years ago some storage vendor was recommending setting their cinder rpc timeout suuuper high because they were making synchronous calls from volume to their backend which took a long time.. do you know of any such suggestion currently? | 14:44 |
| dansmith | and/or other drivers that might do such a thing? | 14:44 |
| smcginnis | We do still have some create from image calls that can take a long time. | 14:45 |
| dansmith | lvm is maybe a little special in this regard | 14:45 |
| dansmith | that the api is waiting for? | 14:45 |
| smcginnis | I can't recall now, but I think it was at the time. Not sure if we changed that. Actually... | 14:46 |
| smcginnis | I think I'm actually thinking of the boot from volume through the Nova api. | 14:46 |
| dansmith | okay, we poll in that case | 14:46 |
| smcginnis | That was the one I remember hitting the most with customers when I was closer to the front line. | 14:46 |
| dansmith | yeah for sure | 14:46 |
| *** jmlowe has quit IRC | 14:48 | |
| mriedem | that was probably before cinder did image caching? | 14:49 |
| mriedem | you had to tweak the nova configs just to keep bfv from timing out polling b/c cinder was downloading the image | 14:49 |
| smcginnis | Yeah, it was especially bad before that. | 14:49 |
| *** psachin has joined #openstack-cinder | 14:50 | |
| mriedem | eharney: i see you just recently enabled the encrypted volume tests in the ceph job | 14:53 |
| mriedem | http://logs.openstack.org/43/619143/14/check/devstack-plugin-ceph-tempest/9e84802/testr_results.html.gz | 14:53 |
| mriedem | http://logs.openstack.org/43/619143/14/check/devstack-plugin-ceph-tempest/9e84802/controller/logs/screen-c-vol.txt.gz?level=TRACE#_Dec_14_08_34_27_307760 | 14:54 |
| mriedem | Dec 14 08:34:27.307760 ubuntu-bionic-rax-iad-0001195720 cinder-volume[1362]: |__Flow 'volume_create_manager': VolumeDriverException: Volume driver reported an error: Provider plain not supported. | 14:54 |
| eharney | yeah, i need to go set the right config to turn that particular test off for that job everywhere | 14:55 |
| mriedem | how did https://review.openstack.org/#/c/625130/ pass then? | 14:55 |
| eharney | because the way i turned it off turned it off in the devstack-plugin-ceph repo and nowhere else :/ | 14:55 |
| mriedem | hmm, all the other jobs should be using that same blacklist | 14:56 |
| eharney | that was the hope, but i guess they aren't as of now | 14:57 |
| mriedem | http://git.openstack.org/cgit/openstack/devstack-plugin-ceph/tree/.zuul.yaml#n26 | 14:57 |
| mriedem | it's the same job | 14:58 |
| eharney | would zuul.project.src_dir be a different value on a different project's run? i'm not sure how that works | 14:58 |
| mriedem | hmm | 14:58 |
| eharney | http://logs.openstack.org/43/619143/14/check/devstack-plugin-ceph-tempest/9e84802/job-output.txt.gz#_2018-12-14_07_52_04_770767 checks and doesn't apply the file | 14:58 |
| *** kaisers_ is now known as kaisers_away | 14:58 | |
| mriedem | revert here for the time being https://review.openstack.org/#/c/625277/ | 15:00 |
| *** mchlumsky has joined #openstack-cinder | 15:00 | |
| mriedem | yup the non-devstack-plugin-ceph runs are definitely not using the blacklist | 15:01 |
| mriedem | http://logs.openstack.org/43/619143/14/check/devstack-plugin-ceph-tempest/9e84802/job-output.txt.gz#_2018-12-14_07_52_08_319347 | 15:01 |
| mriedem | gmann: ^ | 15:01 |
| mriedem | oh actually https://review.openstack.org/#/c/570015/ is probably the problem | 15:03 |
| *** zul has joined #openstack-cinder | 15:04 | |
| *** jmlowe has joined #openstack-cinder | 15:04 | |
| eharney | i think i just need to change it to a zuul.projects.['something'] reference | 15:05 |
| *** markvoelker has quit IRC | 15:06 | |
| mriedem | why did that need to change at all? | 15:07 |
| mriedem | there was no mention in the comments or commit message | 15:07 |
| mriedem | i posted a revert for it here https://review.openstack.org/#/c/625279/ but apparently it's already in merge conflict | 15:08 |
| eharney | the previous path coded in zuul.yaml was not resulting in the blacklist file loading | 15:09 |
| mriedem | hmm, gonna need gmann or AJaeger to help then probably, | 15:10 |
| mriedem | i've posted reverts of both changes for now if we want to push those through until it's sorted out and we can show it working on both devstack-plugin-ceph changes and non-dpc changes | 15:11 |
| mriedem | we can test that using a depends-on from cinder to the dpc change | 15:11 |
| dansmith | smcginnis: we are not currently using directio: https://pastebin.com/raw/Hktap6xd | 15:16 |
| smcginnis | dansmith: Good check. So we need to add something to devstack to set that? | 15:17 |
| dansmith | smcginnis: that patch is up and I'm waiting for it to hit the same spot to see if that 0 flips to a 1 | 15:17 |
| dansmith | smcginnis: boom: https://pastebin.com/raw/7cw7rd7G | 15:18 |
| openstackgerrit | Francois Deppierraz proposed openstack/cinder master: cinder-volume: Stop masking IOError different than ENOSPC https://review.openstack.org/612393 | 15:19 |
| smcginnis | Nice! | 15:19 |
| smcginnis | dansmith: Got a link? | 15:19 |
| dansmith | smcginnis: https://review.openstack.org/#/c/625269/1 | 15:19 |
| dansmith | I'll modify the swift one as well and submit without the debug patch underneath | 15:20 |
| dansmith | mriedem: ^ | 15:20 |
| smcginnis | dansmith: Thanks for jumping on that. I think it will be good to just merge it and see if we see a difference. | 15:20 |
| dansmith | yeah | 15:21 |
| *** psachin has quit IRC | 15:21 | |
| *** gnufied has joined #openstack-cinder | 15:21 | |
| dansmith | actually, I'm not sure where the swift one is | 15:21 |
| smcginnis | Maybe less of a concern though. Maybe | 15:21 |
| mriedem | does that need to be conditional on the filesystem? we have a check for that in nova https://github.com/openstack/nova/blob/5ca357c100b15341001bff1bba663c36bb13d5a8/nova/privsep/utils.py#L35 | 15:22 |
| dansmith | ah, swift loop mounts | 15:22 |
| dansmith | so yeah I guess I'll punt on that for now | 15:23 |
| *** dpawlik has quit IRC | 15:24 | |
| dansmith | aw man, xenial doesn't have it in losetup even though the kernel should be new enough | 15:26 |
| dansmith | so the bionic jobs can do it but the xenial ones won't | 15:27 |
| *** psachin has joined #openstack-cinder | 15:28 | |
| dansmith | mriedem: sorry I missed that the above comment was related until I just saw it | 15:33 |
| dansmith | on the patch | 15:34 |
| *** dpawlik has joined #openstack-cinder | 15:48 | |
| *** dpawlik has quit IRC | 15:52 | |
| dansmith | mriedem: so if we do this directio thing on bionic, we can track and see if it helped, and if so justify moving the other jobs to bionic for most things? | 15:59 |
| dansmith | at least, jobs that run volume tests? | 15:59 |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - changing from 8.4 to 9.0 REST endpoints https://review.openstack.org/625297 | 16:05 |
| *** dklyle has quit IRC | 16:07 | |
| mriedem | heh all jobs run volume tests :) | 16:07 |
| *** dklyle has joined #openstack-cinder | 16:07 | |
| mriedem | gmann is on a crusade to move all legacy jobs over to zuulv3, which would get them to bionic | 16:07 |
| mriedem | so i think that will happen with time, | 16:07 |
| mriedem | i'm working on trying to kill the nova-multiattach job and fold it into tempest-full and tempest-slow, which i thnk would also resolve that one, | 16:08 |
| mriedem | but that's hung up on other stupid stuff not working | 16:08 |
| mriedem | iow, i don't think we need justification to move jobs to bionic, just people power | 16:08 |
| mriedem | soylant green gives me people power | 16:09 |
| dansmith | no, I meant "provide another reason to do it" | 16:09 |
| dansmith | my tweaked devstack patch passes xenial and bionic now | 16:09 |
| mriedem | an incentive if you will | 16:10 |
| dansmith | yes, incentive is what I meant | 16:10 |
| *** dklyle has quit IRC | 16:13 | |
| *** pcaruana has quit IRC | 16:20 | |
| *** e0ne has quit IRC | 16:24 | |
| *** dklyle has joined #openstack-cinder | 16:24 | |
| *** jmlowe has quit IRC | 16:27 | |
| *** kaisers_away is now known as kaisers_ | 16:37 | |
| *** dklyle has quit IRC | 16:41 | |
| openstackgerrit | Eric Harney proposed openstack/python-cinderclient master: Re-enable shell UUID completion cache https://review.openstack.org/610619 | 16:43 |
| openstackgerrit | Eric Harney proposed openstack/python-cinderclient master: Fix doc build error https://review.openstack.org/611708 | 16:44 |
| openstackgerrit | Eric Harney proposed openstack/cinder master: DNM: Test Ceph job https://review.openstack.org/625308 | 16:50 |
| *** kaisers_ is now known as kaisers_away | 16:52 | |
| *** kaisers_away is now known as kaisers_ | 17:00 | |
| *** kaisers_ is now known as kaisers_away | 17:00 | |
| *** kaisers_away is now known as kaisers_ | 17:01 | |
| *** udesale has quit IRC | 17:07 | |
| *** mriedem is now known as mriedem_lunch | 17:10 | |
| *** erlon_ has joined #openstack-cinder | 17:10 | |
| *** dklyle has joined #openstack-cinder | 17:11 | |
| *** e0ne has joined #openstack-cinder | 17:13 | |
| *** erlon has quit IRC | 17:14 | |
| *** luizbag has quit IRC | 17:15 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Allow to use _max qos option together with per_gb https://review.openstack.org/625058 | 17:17 |
| noonedeadpunk | Folks, can someone give a review for https://review.openstack.org/#/c/624970/1 ? | 17:24 |
| eharney | noonedeadpunk: is there a unit test that could hit this case? | 17:27 |
| noonedeadpunk | eharney: existing ones are not hitting this exact thing, but I may re-work them a bit | 17:29 |
| *** markvoelker has joined #openstack-cinder | 17:31 | |
| noonedeadpunk | this one is supposed to check this thing, but numbers are luckely correct for not facing it https://github.com/openstack/cinder/blob/master/cinder/tests/unit/volume/test_connection.py#L301 | 17:31 |
| *** markvoelker has quit IRC | 17:35 | |
| *** psachin has quit IRC | 17:39 | |
| *** e0ne has quit IRC | 17:41 | |
| *** e0ne has joined #openstack-cinder | 17:43 | |
| *** dklyle has quit IRC | 17:51 | |
| *** kaisers_ is now known as kaisers_away | 17:51 | |
| openstackgerrit | Merged openstack/python-cinderclient master: Fix doc build error https://review.openstack.org/611708 | 17:56 |
| *** dklyle has joined #openstack-cinder | 18:00 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Allow to use _max qos option together with per_gb https://review.openstack.org/625058 | 18:01 |
| *** kaisers_away is now known as kaisers_ | 18:06 | |
| *** kaisers_ is now known as kaisers_away | 18:07 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Fix _per_gb_min usage with _per_gb https://review.openstack.org/624970 | 18:09 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Fix _per_gb_min usage with _per_gb https://review.openstack.org/624970 | 18:11 |
| *** itlinux has quit IRC | 18:13 | |
| *** itlinux has joined #openstack-cinder | 18:13 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Fix _per_gb_min usage with _per_gb https://review.openstack.org/624970 | 18:18 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Fix _per_gb_min usage with _per_gb https://review.openstack.org/624970 | 18:19 |
| noonedeadpunk | sorry for this - mixed up with commits at the end of the day((( | 18:20 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Allow to use _max qos option together with per_gb https://review.openstack.org/625058 | 18:21 |
| *** dklyle has quit IRC | 18:25 | |
| *** alkhodos has quit IRC | 18:36 | |
| *** kaisers_away is now known as kaisers_ | 18:37 | |
| *** kaisers_ has quit IRC | 18:42 | |
| *** N3l1x has joined #openstack-cinder | 18:42 | |
| *** N3l1x has quit IRC | 18:42 | |
| *** N3l1x has joined #openstack-cinder | 18:43 | |
| *** dklyle has joined #openstack-cinder | 18:46 | |
| *** mriedem_lunch is now known as mriedem | 18:49 | |
| *** mriedem has quit IRC | 18:52 | |
| *** dklyle has quit IRC | 18:54 | |
| *** mriedem has joined #openstack-cinder | 18:56 | |
| *** dklyle has joined #openstack-cinder | 19:11 | |
| *** e0ne has quit IRC | 19:13 | |
| *** dklyle has quit IRC | 19:21 | |
| *** jmlowe has joined #openstack-cinder | 19:32 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/cinder master: Allow to use _max qos option together with per_gb https://review.openstack.org/625058 | 19:35 |
| *** dklyle has joined #openstack-cinder | 19:42 | |
| *** dklyle has quit IRC | 19:55 | |
| *** e0ne has joined #openstack-cinder | 20:11 | |
| *** e0ne has quit IRC | 20:11 | |
| *** noonedeadpunk has quit IRC | 20:15 | |
| *** mchlumsky has quit IRC | 21:01 | |
| *** tpsilva has quit IRC | 21:23 | |
| *** dklyle has joined #openstack-cinder | 21:27 | |
| *** abishop has quit IRC | 21:30 | |
| *** dklyle has quit IRC | 21:32 | |
| *** dklyle has joined #openstack-cinder | 21:33 | |
| *** woojay has quit IRC | 21:35 | |
| *** mchlumsky has joined #openstack-cinder | 21:35 | |
| *** woojay has joined #openstack-cinder | 21:37 | |
| *** N3l1x has quit IRC | 21:38 | |
| *** mchlumsky has quit IRC | 21:53 | |
| *** dklyle has quit IRC | 21:54 | |
| *** dklyle has joined #openstack-cinder | 21:56 | |
| *** mriedem has quit IRC | 22:01 | |
| *** dklyle has quit IRC | 22:02 | |
| *** dklyle has joined #openstack-cinder | 22:13 | |
| *** dklyle has quit IRC | 22:18 | |
| *** e0ne has joined #openstack-cinder | 22:23 | |
| *** yongwc has joined #openstack-cinder | 22:36 | |
| *** dklyle has joined #openstack-cinder | 22:37 | |
| *** e0ne has quit IRC | 22:40 | |
| *** dklyle has quit IRC | 22:41 | |
| *** dklyle has joined #openstack-cinder | 22:42 | |
| *** dklyle has quit IRC | 22:57 | |
| *** lbragstad has quit IRC | 23:08 | |
| openstackgerrit | Angela Smith proposed openstack/cinder master: REST connector for Brocade zone driver https://review.openstack.org/617100 | 23:08 |
| *** lbragstad has joined #openstack-cinder | 23:16 | |
| *** rosmaita has left #openstack-cinder | 23:20 | |
| openstackgerrit | Merged openstack/cinder stable/rocky: VMAX Driver - VMAX OS Upgrade Bug https://review.openstack.org/624450 | 23:22 |
| openstackgerrit | Merged openstack/cinder stable/rocky: Ensure image utils don't block greenthreads https://review.openstack.org/624497 | 23:22 |
| *** lbragstad has quit IRC | 23:22 | |
| *** yongwc has quit IRC | 23:34 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!