Monday, 2020-06-29

*** aannuusshhkkaa has joined #opendev01:09
*** rchurch has quit IRC01:43
*** ysandeep|away is now known as ysandeep04:00
*** ykarel|away is now known as ykarel04:17
*** bhagyashris is now known as bhagyashris|pto05:16
openstackgerritFederico Ressi proposed openstack/project-config master: Create a new repository for Tobiko DevStack plugin
*** dpawlik4 has joined #opendev06:20
openstackgerritMerged openstack/project-config master: Finish retirement of openstack-ux,solum-infra-guestagent
*** ysandeep is now known as ysandeep|brb06:45
*** rpittau|afk is now known as rpittau06:51
*** ysandeep|brb is now known as ysandeep07:06
*** sshnaidm is now known as sshnaidm|ruck07:24
openstackgerritFederico Ressi proposed openstack/project-config master: Create a new repository for Tobiko DevStack plugin
openstackgerritFederico Ressi proposed openstack/project-config master: Create a new repository for Tobiko DevStack plugin
avassAJaeger: is there still a probem? I would expect the 'ignore_errors' and 'when' to have opposite conditions, or check if the mirroring succeeded before running the next task in that role07:47
avassttx: ^ ?07:47
avassbut oh, the ignore_errors did't evaluate to true when it should have, taking another look07:51
avassAJaeger, ttx: the ignore_errors needs to be templated like: "{{ zuul.newrev is defined }}", only the 'when' attribute doens't need to be templated07:55
*** moppy has quit IRC08:01
openstackgerritFederico Ressi proposed openstack/project-config master: Create a new repository for Tobiko DevStack plugin
ttxavass: Thanks!08:03
*** moppy has joined #opendev08:03
ttxavass: I now think a retry is actually better, since it will avoid the workaround to make it pass tests... but good to understand why that failed to work in the first place. Ansible inconsistent use of templating is a bit confusing (at least to me as I do not practice daily)08:05
kevinzianw: there?08:05
*** ysandeep is now known as ysandeep|lunch08:17
openstackgerritFederico Ressi proposed openstack/project-config master: Create a new repository for Tobiko DevStack plugin
AJaegeravass: thanks!08:37
*** ysandeep|lunch is now known as ysandeep08:50
*** aannuusshhkkaa has quit IRC09:05
avassttx: yeah I agree retrying is probably more stable09:07
*** DSpider has joined #opendev09:11
ianwkevinz: in and out for a little09:34
ianwkevinz: on pto tomorrow but around rest of week09:34
kevinzianw: OK, could you help to lock an instance: ubuntu-bionic-arm64-large-linaro-us-0017437847?09:35
kevinzif not convenient for you, could you point anthor maintainer to help :-)09:36
kevinzI need to lock this instance and log into to debug the devstack CI job09:36
kevinzThanks ianw09:36
ianw0000000159 | openstack |     |               devstack-platform-arm64               |             .*            |     1     |                            ianw: kevinz destack debug09:41
ianwkevinz: ^ if you want to start another run, the next failure should hold09:41
kevinzianw: OK, you mean that I can just run "recheck" and the vm will be hold?09:42
kevinzianw thanks! Triggering09:46
*** hashar has joined #opendev09:50
*** tkajinam has quit IRC10:05
*** rpittau is now known as rpittau|bbl10:21
ianwkevinz: 2604:1380:4111:3e56:f816:3eff:fe57:29be ... what key do you want on it?10:27
kevinzianw: send you via private chat. thanks10:28
kevinzianw: succeed. Thanks!10:30
*** tosky has joined #opendev10:57
*** ysandeep is now known as ysandeep|brb11:02
*** dtantsur has joined #opendev11:14
dtantsurfolks, not sure how much involved in the etherpad development we are, but would it be possible to make it reconnect on network problems?11:14
*** ysandeep|brb is now known as ysandeep11:43
*** rpittau|bbl is now known as rpittau12:00
*** olaph has joined #opendev12:07
*** ianw is now known as ianw_pto12:13
*** ryohayakawa has quit IRC12:17
*** lourot has joined #opendev12:37
lourothey, I'm from Canonical / openstack-charmers, regarding , is it still true that mirroring to GitHub for OpenStack project happens for free?12:39
lourotthere are some rumors that it might not be true anymore12:39
lourotalso I can see the mirroring seems to be happening for some of ours repos but not all12:39
openstackgerritDmitry Tantsur proposed openstack/diskimage-builder master: pip-and-virtualenv: do not unconditionally enable EPEL for CentOS
openstackgerrityatin proposed openstack/diskimage-builder master: Disable all enabled epel repos in CentOS8
clarkblourot: we stopped using gerrit's replication plugin to replicate to github as it was inflexible andhard to manage. Now we run zuul jobs to replicate. ttx hasbeen managing that fir openstack projects and I would expect all of them to be replicating via that method12:52
clarkbdtantsur: I believe it detects disconnection a nd prompts you with a reconnect button but does not auto reconnect12:52
dtantsurclarkb: the problem is, it only prompts you to reconnect once you start typing something12:53
dtantsurbefore that you have no clue that it has lost connection12:53
clarkbdtantsur: that may be necessary to d etect the disconnect12:53
smcginnisclarkb: You're up early. Anyone ask about all the job failures with destroy-afs-token yet? I saw it on some periodic jobs, now it looks like it's affected a couple release jobs.12:55
lourotclarkb, thanks, the mirror of a new repo that I got created last week (via reviews to 'governance' and 'project-config') seems to remain empty. Should I ping ttx?12:56
clarkbsmcginnis: I hadn't heard, its possible that is a side effect of ze01 being in containers. smcginnis have a link to an example or three?12:56
smcginnisclarkb: Here's one of the recent ones:
smcginnis2020-06-29 09:51:31.25764412:56
clarkblourot: for new repos that may require landing code if that hasnt' happened yet. If you have landed some changes then ya I'd check with ttx12:56
clarkbsmcginnis: and ya this is early. Trying to hepl with the opendev event the next few days12:57
lourotclarkb, not landed code yet, thanks!12:57
smcginnisI'm either losing touch, or that event hasn't been publicized well. Probably the former.12:57
clarkbsmcginnis: I actually amde that comment too. ildikov suggested that I wasn't the primary audience and so missed it. Not sure12:58
smcginnisGuess I'm not either then.12:58
ttxlourot: it will be synced at the next commit12:59
smcginnislourot: You may need to add the job openstack-upload-github-mirror.12:59
ttxsmcginnis: it's in the openstack jobs template13:00
smcginnisYeah, just found
clarkbsmcginnis: that job did run on ze0113:01
clarkbI'm going to stop ze01's executor now13:02
smcginnisclarkb: So maybe some issues to work out running containerized yet?13:02
clarkbsmcginnis: yes, I think the `unlog` command is missing in the container13:02
clarkb(which is weird because in theory we have the aklog command there to get to that point in the job and I would've expected them to be installed together13:02
smcginnisclarkb: I just recently realized the even though venv is part of the standard lib in py3, but Ubuntu packages it separately. So packagers... *shrug*13:04
clarkbsmcginnis: ya ubuntu inherits that from debian.13:04
clarkb#status log Stopped zuul-executor on ze01 as the container image it is running lacks `unlog`13:06
openstackstatusclarkb: finished logging13:06
clarkbmordred: ^ fyi13:06
fungiit's not the only part of stdlib debian packages separately. in fact the majority of the stdlib is in a separate debian package form the interpreter itself13:07
*** jangutter has joined #opendev13:07
clarkbI think we are missing openafs-client package in the image13:08
clarkb(which is the ~pre image which may be problematic?)13:08
clarkb seems to confirm openafs-client is the missing pacakge13:14
fungiyeah, need that for afsd13:15
openstackgerritTristan Cacqueray proposed zuul/zuul-jobs master: ensure-dhall: update to latest version
*** Eighth_Doctor has quit IRC13:46
*** ysandeep is now known as ysandeep|afk13:50
corvusfungi, clarkb, mordred: hrm, that's unfortunate; openafs-client is the package that runs afsd; that seems weird to install in a container image13:54
fungioh, i see it's already installed/running on the host?13:55
fungiis /afs being bind-mounted in from the host?13:55
corvusfungi: yes13:55
fungioh, it's because we were calling unlog in the job13:56
corvusopenafs-client is going to bring in all the dkms stuff and try to compile a kernel module13:56
fungii should have read more closely13:56
clarkbfungi: yes we need the unlog command13:56
clarkband that is in the openafs-client package13:56
clarkbbecause its a container I doubt the container would run that daemon just because the package is isntalled13:56
clarkbbut it would have the unlog command available (not the most disk efficient situation)13:57
fungiso we don't need to start afsd in the container13:57
corvusi'm installing it in a container now to see what happens13:57
fungithat makes a lot more sense13:58
fungiwe might similarly want to be able to vos release from a job, which would also require that package13:58
corvusit does install the whole dkms build chain.  and it is building a kernel module during the install.  slowly.  i'll let you know what happens next.14:01
corvusokay, it finished.  took a long time.  it didn't try to start it or anything, so it seems like it might not be totally broken.14:05
clarkbcorvus: we might even do a post package install step to rm the kernel mod and chmod -x afsd?14:06
clarkbif we're worried about it somehow conflicting14:06
clarkbmay also cut image size to not have the kernel stuff there14:06
corvusbased on a zuul-executor image i have laying around (unsure how recent) apt-get says: After this operation, 11.9 MB of additional disk space will be used.14:06
clarkbah I wonder if they already do magic to keep apt from doing kernel stuff we don't need14:07
corvuswhat would be really cool is if there were a way to disable dkms on a host14:08
*** Eighth_Doctor has joined #opendev14:08
corvusoh interesting... i ran the install command in the zuul-executor image and it didn't build it because 'lsmod' was not found14:10
corvuslet me try this one more time with a current image14:10
corvusoh, the kernel module source is a recommended package14:14
corvusso we don't get it.  yay!14:14
corvus  libhx509-5-heimdal libkrb5-26-heimdal libwind0-heimdal openafs-client14:14
corvusAfter this operation, 10.1 MB of additional disk space will be used.14:14
corvusso still not cheap, but not that bad.14:14
clarkbya compared to the overall image size thats pretty good14:15
clarkbif we were pulling in a new kernel and modules that would be different :)14:15
corvusgetting closer to needing our "zuul-executor-heavy/light" idea :)14:15
openstackgerritMerged openstack/project-config master: Finish retirement of networking-onos
fungicorvus: clarkb: looks like if we don't install recommends only depends it won't pull in openafs-modules-dkms14:35
clarkbfungi: yup14:35
fungioh, you got there eventually too ;)14:35
fungisorry, distracted by the conference, just now catching up again14:36
*** ykarel is now known as ykarel|away14:55
*** _mlavalle_1 has joined #opendev14:55
mordredcorvus: yay openafs-client14:56
*** sgw1 has joined #opendev14:56
*** ysandeep|afk is now known as ysandeep|away14:57
*** mlavalle has quit IRC14:58
mordredinfra-root: we're closing on a house today, so I'm not really here today. and I'm gonna do some vacation days this week to, you know, play with the new house. I'll still likely poke my head in and out, but will be largely not 100% online14:58
clarkbI've been watching cacti for etherpad during the opendev event and it seems to have stabilized at a level and been happy there.15:31
clarkbits a bit different than a typical event in that we don't have a bunch of small etherpads running at the same time. Instead its a single etherpad with a ton of users. All seems fine though15:32
*** hashar has quit IRC15:36
fungiyeah, i've been impressed, it's holding up great and hardly breaknig a sweat15:38
clarkbit looks like my gitea patch is working after rechecking it over the weekend and upstream is likely to pull it in (I got a bunch of approvals)15:47
clarkb is relying on it and the workaround has been commented out15:48
clarkbno failures yet15:48
*** hashar has joined #opendev15:52
*** priteau has joined #opendev15:54
mordredclarkb: great email from wikimedia on repo-discuss about their gerrit upgrade15:56
mordredthey went all the way to 3.2 from 2.15 and were only down for an hour15:57
clarkbI saw it in my inbox but haven't read it yet15:57
mordredthere is a patch we'll want to cherry-pick into our builds15:57
mordredthat helps apparently dramatically with reindexing15:57
clarkbI'm going to take it easy this morning as I was up very very early for the opendev event, but I'll try to get ze01 restarted on the new image if no one else beats me to it15:59
clarkbI've also got a checkup this afternoon beacuse I Figure now is the best time for that vs $date_in_the_future15:59
openstackgerritElod Illes proposed zuul/zuul-jobs master: DNM: ensure-pip to upgrade pip3 if version is 8.X.X
mordredclarkb: specifically
*** _mlavalle_1 has quit IRC16:05
*** mlavalle has joined #opendev16:25
*** mlavalle has quit IRC16:30
*** shtepanie has joined #opendev16:30
*** _mlavalle_1 has joined #opendev16:36
*** priteau has quit IRC16:38
*** aannuusshhkkaa has joined #opendev16:47
*** priteau has joined #opendev16:48
clarkbI've done a docker-compose pull on ze01 and looks like it got the new image17:04
clarkbI'll start ze01's executor as soon as I'm not distracted with kids' school17:04
*** hashar has quit IRC17:12
*** priteau has quit IRC17:13
openstackgerrityatin proposed openstack/diskimage-builder master: Make ipa centos8 job non-voting
openstackgerrityatin proposed openstack/diskimage-builder master: Disable all enabled epel repos in CentOS8
clarkb#status log Restarted zuul-executor on ze01 with new docker image that includes openafs-client17:26
openstackstatusclarkb: finished logging17:26
clarkbsmcginnis: ^ as a heads up we think we've addressed the problme you discovered but do let us know if you see additional oddities17:26
smcginnisWill do, thanks clarkb17:38
*** dtantsur is now known as dtantsur|afk17:51
openstackgerritElod Illes proposed zuul/zuul-jobs master: Upgrade pip3 if version is 8.X.X
*** shtepanie has quit IRC18:40
mordredclarkb: fingers crossed19:05
*** sshnaidm|ruck is now known as sshnaidm|afk19:12
*** hashar has joined #opendev19:25
*** priteau has joined #opendev19:57
*** tobiash has quit IRC20:51
*** sshnaidm|afk has quit IRC20:51
*** tobiash has joined #opendev20:51
*** sshnaidm|afk has joined #opendev20:53
*** diablo_rojo has joined #opendev20:56
openstackgerritCole Walker proposed openstack/project-config master: Add Vault app to StarlingX
*** priteau has quit IRC21:09
*** rchurch has joined #opendev21:28
*** hashar has quit IRC21:29
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Use a temporary registry with buildx
openstackgerritJames E. Blair proposed zuul/zuul-jobs master: Use a temporary registry with buildx
ianw_ptothe openafs role went through all this with the dependencies and ordering etc.  one thing i don't like about docker is we don't reuse good things we've got outside the image builds22:24
ianw_ptospeaking of containers; i know everyone is busy but any reviews of would be welcome so i can trying bringing up the services22:30
*** tosky has quit IRC22:46
*** tkajinam has joined #opendev22:47
* clarkb is back from errands and is now trying to put together the meeting agenda22:49
clarkbI see no new complaints about ze0122:49
fungiyeah, i haven't seen any either22:54
fungiso either all's well, or nobody's around to hear that particular tree falling in our forest22:54
*** DSpider has quit IRC22:59

Generated by 2.17.2 by Marius Gedminas - find it at!