| openstackgerrit | Merged openstack/project-config master: Remove Searchlight projects from infra https://review.opendev.org/c/openstack/project-config/+/764535 | 00:37 |
|---|---|---|
| openstackgerrit | Merged openstack/project-config master: Remove Qinling projects from infra https://review.opendev.org/c/openstack/project-config/+/764536 | 00:39 |
| openstackgerrit | Merged openstack/project-config master: Announce openstack/ossa in #openstack-security https://review.opendev.org/c/openstack/project-config/+/765389 | 00:42 |
| *** tosky has quit IRC | 00:54 | |
| *** openstackgerrit has quit IRC | 00:58 | |
| *** stevebaker has joined #opendev | 01:05 | |
| *** openstackgerrit has joined #opendev | 01:52 | |
| openstackgerrit | Ian Wienand proposed opendev/gerritlib master: docker-compose: update header syntax https://review.opendev.org/c/opendev/gerritlib/+/765898 | 01:52 |
| openstackgerrit | Ian Wienand proposed opendev/system-config master: WIP: initalize gerrit in testing https://review.opendev.org/c/opendev/system-config/+/765224 | 02:05 |
| *** zer0c00l has quit IRC | 02:08 | |
| *** paladox has quit IRC | 02:10 | |
| *** paladox has joined #opendev | 02:15 | |
| *** ysandeep|away is now known as ysandeep | 02:54 | |
| *** stevebaker has quit IRC | 03:41 | |
| *** hamalq_ has quit IRC | 03:45 | |
| *** marios has joined #opendev | 06:17 | |
| *** marios is now known as marios|rover | 06:31 | |
| *** elod_pto has joined #opendev | 06:31 | |
| *** elod_pto has quit IRC | 06:38 | |
| *** hamalq has joined #opendev | 06:44 | |
| *** elod_pto has joined #opendev | 06:45 | |
| *** sboyron has joined #opendev | 07:51 | |
| *** lpetrut has joined #opendev | 07:55 | |
| *** hashar has joined #opendev | 08:07 | |
| *** slaweq has joined #opendev | 08:08 | |
| *** rpittau|afk is now known as rpittau | 08:10 | |
| yoctozepto | infra-root review.opendev.org looks down :/ | 08:11 |
| gibi | looks down to me too | 08:13 |
| yoctozepto | and that's it for the productive morning plan :D | 08:13 |
| ianw | hrm, the host is up and the container is running | 08:13 |
| ianw | it's not responding to ssh | 08:15 |
| marios|rover | yoctozepto: just came here to say same | 08:16 |
| *** andrewbonney has joined #opendev | 08:16 | |
| yoctozepto | marios|rover: only sad reactions! | 08:22 |
| marios|rover | ;) | 08:23 |
| slaweq | hi, I just came here to say the same but I see that You are already aware :) | 08:27 |
| ianw | #status log gerrit restarted after a short outage | 08:27 |
| openstackstatus | ianw: finished logging | 08:27 |
| ianw | it should be coming back up now | 08:27 |
| marios|rover | thanks ianw | 08:27 |
| marios|rover | ianw++ | 08:27 |
| ianw | we'll have to dig further to the exact issue | 08:27 |
| slaweq | ianw++ it works :) | 08:28 |
| zbr | thanks ianw | 08:28 |
| zbr | any idea what happened? OOM? | 08:29 |
| ianw | memory usage did seem to spike a couple of hours ago according to cacti | 08:31 |
| ianw | we'll have to pull the logs and correlate to have more of an idea | 08:31 |
| yoctozepto | thanks ianw | 08:32 |
| ianw | heh well restarting it isn't great. but i'm out of time for today; i have logged some details for other roots in the usual places | 08:33 |
| zbr | lets hope it survives for the next ~6h until someone else comes online. | 08:35 |
| zbr | i observed that my session was reset by the restart, is not gerrit able to persist sessions on disk too? so a restart would not force everyone to login again? | 08:35 |
| zbr | nevermind, i found why it does this, is not configured correctly,I will make a CR to address that. | 08:37 |
| yoctozepto | zbr: oddly, it has not logged me out | 08:45 |
| yoctozepto | but yeah, I noticed it logs me out more frequently than the old one | 08:45 |
| openstackgerrit | Sorin Sbârnea proposed opendev/system-config master: Avoid logging out users on restart https://review.opendev.org/c/opendev/system-config/+/765921 | 08:48 |
| zbr | i also observed something weird, based on docs the memory caches are measures in bytes, and the values that I seen there seam far too small for our deployment. | 08:48 |
| zbr | false alarm, memoryLimit stats is measured in entries not bytes. Apparently google developers did not learn to properly name variables :D | 08:52 |
| zbr | https://gerrit-review.googlesource.com/Documentation/config-gerrit.html#cache.name.memoryLimit paragraph is super confusing | 08:53 |
| zbr | even the documentation is misleading. it stated that diskLimit is default 128 MiB except 4 specific caches. But when you read web_sessions, you discover that this defaults to 0. | 08:58 |
| yoctozepto | zbr: well, it did *not* log me out | 08:58 |
| yoctozepto | zbr: so I guess there is *some* cache | 08:58 |
| zbr | devil may be in the implementation, when using external authentication it is possible to restore session from remote. | 09:00 |
| zbr | last time i accessed gerrit was ~12h ago, this morning I had to login again. Based on the fact that I do not dicover this each morning, I have a high degree of confidence that these two are related. | 09:01 |
| zbr | let me survey my collegues.... | 09:01 |
| *** hamalq has quit IRC | 09:10 | |
| *** hamalq has joined #opendev | 09:14 | |
| yoctozepto | zbr: well, let's try the config change and see if it helps; should not hurt :D | 09:15 |
| yoctozepto | I did not notice the auth redirects so unsure what worked | 09:15 |
| *** hamalq_ has joined #opendev | 09:16 | |
| zbr | that is my view too, but lets see what those with extra experience think. sometimes docs lie. | 09:16 |
| *** hamalq has quit IRC | 09:18 | |
| *** hamalq_ has quit IRC | 09:20 | |
| *** DSpider has joined #opendev | 09:52 | |
| *** hamalq has joined #opendev | 10:00 | |
| *** hamalq has quit IRC | 10:04 | |
| openstackgerrit | daniel.pawlik proposed openstack/diskimage-builder master: Remove centos-repos package for Centos 8.3 https://review.opendev.org/c/openstack/diskimage-builder/+/765963 | 10:34 |
| openstackgerrit | daniel.pawlik proposed openstack/diskimage-builder master: Remove centos-repos package for Centos 8.3 https://review.opendev.org/c/openstack/diskimage-builder/+/765963 | 10:38 |
| yoctozepto | interesting case -> recheck only run against arm64 pipeline https://review.opendev.org/c/openstack/kolla/+/765860 | 10:41 |
| ianw | yoctozepto: possibly a message got lost during that gerrit downtime? | 10:56 |
| *** dtantsur|afk is now known as dtantsur | 11:00 | |
| yoctozepto | ianw: that could be it; I thought it would queue it | 11:07 |
| *** dviroel has joined #opendev | 11:10 | |
| gibi | hi! I saw plenty of "openstack-tox-py38 https://zuul.opendev.org/t/openstack/build/85b94d33f4774bebb61b0a4f7f5ae5e0 : ERROR Failed to update project None in 7m 25s" Zuul results, is it also due to the gerrit outage happen couple hours ago? Example: https://review.opendev.org/c/openstack/nova/+/757109 | 11:12 |
| yoctozepto | failing to do nothing should be a success! | 11:14 |
| gibi | especially during the incoming holiday season | 11:16 |
| yoctozepto | ++ | 11:26 |
| *** hashar is now known as hasharLunch | 11:26 | |
| slaweq | hi infa-team | 11:44 |
| slaweq | can You take a look at https://bugs.launchpad.net/neutron/+bug/1907242 and check if that can maybe be another pypi cache issue? | 11:44 |
| openstack | Launchpad bug 1907242 in neutron "ERROR: No matching distribution found for wrapt==1.11.* in the CI jobs" [Critical,Confirmed] | 11:44 |
| slaweq | because it happens a lot in the neutron jobs since yesterda | 11:44 |
| slaweq | *yesterda | 11:45 |
| slaweq | *yesterday :) | 11:45 |
| *** hamalq has joined #opendev | 12:01 | |
| *** hamalq has quit IRC | 12:05 | |
| *** fressi has joined #opendev | 12:25 | |
| *** hasharLunch is now known as hashar | 12:37 | |
| *** dtantsur has quit IRC | 12:50 | |
| *** dtantsur has joined #opendev | 13:06 | |
| *** owalsh has quit IRC | 13:20 | |
| *** fressi has quit IRC | 13:26 | |
| *** fressi has joined #opendev | 13:41 | |
| *** tosky has joined #opendev | 13:44 | |
| *** hamalq has joined #opendev | 14:02 | |
| *** slaweq has quit IRC | 14:03 | |
| *** slaweq has joined #opendev | 14:04 | |
| *** hamalq has quit IRC | 14:07 | |
| frickler | slaweq: I see the same issue locally, so not an issue on our side | 14:09 |
| slaweq | frickler: hmm, I didn't saw it on my vm locally | 14:09 |
| slaweq | interesting | 14:09 |
| frickler | slaweq: if you try to reproduce, make sure to upgrade pip in your venv before running tox | 14:10 |
| slaweq | frickler: ok, maybe that is the case indeed, I will check later | 14:10 |
| slaweq | thx | 14:10 |
| frickler | so "python3 -m venv .tox/shared/; .tox/shared/bin/pip install -U pip; .tox/shared/bin/pip install -chttps://releases.openstack.org/constraints/upper/master -r/opt/stack/neutron/requirements.txt -r/opt/stack/neutron/test-requirements.txt" shows the error for me | 14:10 |
| *** haleyb has joined #opendev | 14:13 | |
| *** bcafarel has joined #opendev | 14:13 | |
| *** hamalq has joined #opendev | 14:17 | |
| *** hamalq has quit IRC | 14:22 | |
| *** tkajinam has quit IRC | 14:47 | |
| *** tkajinam has joined #opendev | 14:47 | |
| *** marios|rover is now known as marios|call | 14:57 | |
| *** marios|call is now known as marios|rover | 15:06 | |
| *** ysandeep is now known as ysandeep|away | 15:16 | |
| *** tkajinam has quit IRC | 15:18 | |
| *** hashar has quit IRC | 15:31 | |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Reduce gerrit heap limit to 44g https://review.opendev.org/c/opendev/system-config/+/766020 | 15:34 |
| yoctozepto | can we start rebuilding centos as part of opendev? 😂 | 15:41 |
| clarkb | yoctozepto: I won't :P, in theory the system allows for it, but building a linux distro is a massive undertaking and i already have way too much on my plate | 15:42 |
| yoctozepto | clarkb: I wonder if we can do a subset though | 15:42 |
| yoctozepto | 100% not interested in the graphics bloat | 15:43 |
| yoctozepto | and otherwise centos is much smaller than debian | 15:43 |
| yoctozepto | but I feel you | 15:43 |
| yoctozepto | oh what a pandora box there | 15:43 |
| clarkb | ya but its also not a super valuable task for our projects, there are many distro options including stable alternatives | 15:43 |
| clarkb | if centos isn't fitting the bill anymore use another | 15:43 |
| yoctozepto | but rhel is good, and the sources are lying around :D | 15:44 |
| yoctozepto | as for me, I will probably go with stream for some uses and ubuntu (or debian) for others | 15:45 |
| *** tosky has quit IRC | 15:45 | |
| yoctozepto | unless someone rebuilds elsewhere | 15:45 |
| dtantsur | aha, so the folks are already discussing it. great! | 15:51 |
| dtantsur | I guess the infra folks are just as surprised as I am, but I really wonder what will happen with nodesets for CentOS | 15:52 |
| dtantsur | (as a sub-question, what will happen with DIB support) | 15:52 |
| clarkb | dtantsur: we'll delete centos-8 when it eols (roughyl) and you'll have the option to use stream (stream is already available) | 15:53 |
| clarkb | prior to eol I think everyone should be shifting to stream now | 15:53 |
| dtantsur | clarkb: even for stable branches? | 15:53 |
| clarkb | dtantsur: maybe? the way we handle fedora on stable branches is to just drop those jobs when fedora eols | 15:54 |
| dtantsur | I think you have fedora-latest as an alias? | 15:54 |
| clarkb | (we actually try to switch to fedora latest first iirc and if that breaks drop them, so ya switching to stream, if that beraks drop it) | 15:54 |
| clarkb | dtantsur: yes | 15:54 |
| clarkb | essentially I would stop using centos 8 for anything new right now. FIgure out what using stream looks like for new stuff then decide how to handle old stuff based on that info | 15:55 |
| dtantsur | I assume, since we have nodesets for stream, DIB supports it? | 15:55 |
| clarkb | yes dib supports stream | 15:55 |
| dtantsur | mmmok, thank you. the ironic community has something to think about. | 15:55 |
| dtantsur | how is the nodeset called? centos-stream? | 15:56 |
| clarkb | dtantsur: fwiw this is based on how we handle other EOLs but I don't expect we'll diverge much for centos 8 | 15:56 |
| dtantsur | sure, we just did not expect centos 8 EOL so soon :) | 15:57 |
| clarkb | the only real difference here is this eol is unexpected rather than previously planned for | 15:57 |
| clarkb | ya that | 15:57 |
| clarkb | but we don't control that either :/ | 15:57 |
| dtantsur | I feel a bit bad since I advocated for relying on CentOS as a reliable foundation for years... and now this..... | 15:57 |
| yoctozepto | clarkb, dtantsur: just enter #centos-devel and welcome in hell :-) | 15:57 |
| dtantsur | mm, thank you but no thank you | 15:58 |
| yoctozepto | dtantsur: you are welcome nonetheless ;D | 15:58 |
| dtantsur | I feel really, really bad for people working on centos now.. they'll get a shitstorm they don't deserve | 15:58 |
| yoctozepto | yes, they are now | 15:58 |
| yoctozepto | well, they actually promised the eol of centos8 in 2029 | 15:59 |
| yoctozepto | and now it's in 2024 | 15:59 |
| yoctozepto | due to stream | 15:59 |
| clarkb | I guess scientific linux is still alive, I thoughti t was also effectively eol but I probably just confused CERN moving to centos with that | 15:59 |
| dtantsur | "we will shift our investments to CentOS Stream exclusively on December 31, 2021" (c) Red Hat | 15:59 |
| clarkb | yoctozepto: ^ I expect if you want to help a rebuild that is a good bet | 15:59 |
| yoctozepto | dtantsur: for stream in CI you can checkout my change to devstack | 15:59 |
| dtantsur | yoctozepto: 2021, not 2024. 2024 is for CentOS 7. | 15:59 |
| yoctozepto | dtantsur: 2024 for centos 8 stream | 16:00 |
| *** hashar has joined #opendev | 16:00 | |
| yoctozepto | 2021 for centos 8 releases | 16:00 |
| yoctozepto | still 5 years too early from the planned 2029 | 16:00 |
| yoctozepto | clarkb: hmm, I thought sl was not going to produce 8 | 16:01 |
| yoctozepto | and yes, it does not produce it | 16:01 |
| yoctozepto | at least not ey | 16:01 |
| clarkb | yoctozepto: I imagine that things have signfiicantly changed today :) but ya they may still avoid it | 16:01 |
| yoctozepto | yet | 16:01 |
| dtantsur | scientific linux was abandoned in favour of centos | 16:01 |
| yoctozepto | clarkb: yeas, it is one crazy day | 16:01 |
| dtantsur | now that centos is a different thing.. who knows | 16:02 |
| yoctozepto | indeed | 16:02 |
| dtantsur | clarkb: do you plan on making the nodeset versioned? i.e. centos-8-stream vs future centos-9-stream? | 16:07 |
| clarkb | dtantsur: I think it already is centos-8-stream | 16:08 |
| clarkb | so ya I would expect a centos-9-stream once it arrives | 16:08 |
| dtantsur | ah, I should have verified my assumptions before asking :) | 16:08 |
| fungi | we'll presumably follow suit with what rh does, so if they have separate 8 and 9 streams we'd presumably carry both until they eol | 16:08 |
| dtantsur | sorry, a crazy day. started with all the lower-constrains madness.... | 16:08 |
| clarkb | fungi: ya looks like they will do separate streams beacuse each stream is the feeder for the corresponding rhel release | 16:08 |
| dtantsur | this ^^ | 16:08 |
| yoctozepto | dtantsur: it was lower-constraints and centos 8.3 today for me and now this, crazy day :D | 16:10 |
| dtantsur | ah, yeah, 8.3 as well | 16:11 |
| yoctozepto | centos 8.3 brought repo rename to match stream | 16:11 |
| yoctozepto | I SEE WHAT YOU DID THERE | 16:11 |
| yoctozepto | dtantsur: the job I mentioned is https://review.opendev.org/c/openstack/devstack/+/759122 | 16:12 |
| dtantsur | thanks! so, it's green without too much effort? this is a good sign! | 16:15 |
| clarkb | I imagine the delta is minimal right now since centos 8 just did an update too? | 16:15 |
| dtantsur | yep, 8.3 was out today | 16:15 |
| *** hamalq has joined #opendev | 16:18 | |
| *** hamalq has quit IRC | 16:23 | |
| *** fressi has quit IRC | 16:29 | |
| openstackgerrit | Matt McEuen proposed openstack/project-config master: New Project Request: airship/vino https://review.opendev.org/c/openstack/project-config/+/763889 | 16:35 |
| openstackgerrit | Matt McEuen proposed openstack/project-config master: New Project Request: airship/sip https://review.opendev.org/c/openstack/project-config/+/763888 | 16:35 |
| *** kopecmartin has quit IRC | 16:43 | |
| yoctozepto | clarkb: yeah, the tribute has been paid | 16:45 |
| yoctozepto | dtantsur: yes, it's green; it's not that very different | 16:45 |
| *** kopecmartin has joined #opendev | 16:48 | |
| *** lpetrut has quit IRC | 16:48 | |
| *** hamalq has joined #opendev | 16:56 | |
| *** hamalq has quit IRC | 17:01 | |
| *** hamalq has joined #opendev | 17:06 | |
| zbr | clarkb: fungi: this morning gerrit was restarted and forced me to login again. that make me google a bit and check the config. I ended up proposing: https://review.opendev.org/c/opendev/system-config/+/765921 | 17:12 |
| zbr | does it makes sense? | 17:12 |
| zbr | config states that disk persistency is not enabled by default for web-session objects. | 17:13 |
| zbr | and recommends enabling it | 17:13 |
| *** owalsh has joined #opendev | 17:13 | |
| zbr | i am not yet sure if the value is correct, the docs are very confusing regarding the meaning of limit, varies a lot between different fields. | 17:14 |
| clarkb | zbr: can you link to that because I'm raeding the docs and it says it is | 17:14 |
| fungi | show-caches is showing it disk-backed on our deployment | 17:14 |
| clarkb | under cache.<name>.diskLimit it talks about defaulting to 128MB except for specific cases. web sessions is not a specific case | 17:14 |
| clarkb | https://gerrit-review.googlesource.com/Documentation/config-gerrit.html#cache that is the documentation I am lokoing at | 17:15 |
| fungi | Name |Entries | AvgGet |Hit Ratio| | 17:15 |
| fungi | | Mem Disk Space| |Mem Disk| | 17:15 |
| zbr | the page is linked from CR, search for: "cache "web_sessions" -- yep not anchors | 17:15 |
| fungi | D web_sessions | 382 285323 128.02m| | 86% 1%| | 17:15 |
| yoctozepto | fungi: that would align with my observation of not being logged out | 17:15 |
| *** marios|rover is now known as marios|out | 17:15 | |
| zbr | If no disk cache is configured (or cache.web_sessions.diskLimit is set to 0) a server restart will force all users | 17:15 |
| zbr | that is copy/paste from docs | 17:16 |
| zbr | most caches have defaults, that one is zero, but is not listed among those.... | 17:16 |
| fungi | zbr: fair warning, we're discovering lots of places where gerrit's docs are not entirely correct too, so have been making sure to check any assumptions before believing them | 17:16 |
| clarkb | zbr: the docs don't say its default is zero? | 17:16 |
| clarkb | they say if it is zerop | 17:16 |
| clarkb | and where they talk about defaults they don't mention web sessions as being special impying it has a 128mb limit | 17:17 |
| clarkb | which lines up with what we see from show-caches | 17:17 |
| clarkb | I don't think that configuration will change anything for us | 17:17 |
| zbr | tbh, these docs are a total mess, i spent an hour this morning trying to understand how these limits work. different sections have contradicting statements. | 17:17 |
| clarkb | zbr: agreed, I'm still waiting on a response to my question if specific values need to go in jgit.cofnig or not because the docs don't make that clear | 17:18 |
| clarkb | zbr: I think I've decided we'll just set them in jgit.config and gerrit.config until we understand it better | 17:18 |
| zbr | for this particular option I think is clear that is inside gerrit.config | 17:19 |
| clarkb | zbr: yes for this one its fine. But per our cache overservations the default seems to be 128MB on disk | 17:19 |
| clarkb | which should be plenty. I don't think we need your change based on fungi's metric gathering | 17:19 |
| zbr | https://review.opendev.org/c/opendev/system-config/+/765921/ -- is the review. | 17:19 |
| clarkb | (I don't know why you were forced to log back in again, I was too) | 17:20 |
| clarkb | s/overservations/observations/ I can't type | 17:21 |
| zbr | yeah, that was weird because I asked others, at least two people did not had to login again. | 17:21 |
| zbr | maybe sometimes the re-auth with ubuntu happens in the background? | 17:21 |
| clarkb | it shouldn't, ubuntu one should always ask you to re acknowledge the log in | 17:22 |
| fungi | openid doesn't have a "background" | 17:22 |
| fungi | you literally have to bounce through http redirects to the identity provider and back | 17:22 |
| zbr | do we had only ~380 users in the last 7 days? | 17:24 |
| clarkb | no thats the memory column | 17:24 |
| clarkb | its ~380 cached entries since the restart | 17:25 |
| clarkb | which seems in the realm of posibility (or the memory cache is quite small because the disk cache is large?) | 17:25 |
| *** tosky has joined #opendev | 17:26 | |
| fungi | also cached sessions may not be equal to users, you could have multiple sessions for a single user on different devices | 17:26 |
| zbr | i guess you already double checked all memoryLimit values? as the object count vs bytes can easily lead to big mistakes. | 17:28 |
| clarkb | zbr: default should be 1024 web sessions according to the docs for memoryLimit | 17:30 |
| zbr | probably for accounts* we do want to be sure we have 100% memory hits, as the username lookup is likely to search among them. | 17:30 |
| *** marios|out has quit IRC | 17:30 | |
| clarkb | its harder to confirm that without diggin in the source, or waiting for the show-caches command to show you've hit a round number limit like that | 17:31 |
| zbr | any change we could log all caches daily so we can look for trends? | 17:32 |
| clarkb | maybe? It is an administrator command iirc so may take some figuring out | 17:32 |
| zbr | i wonder if i may have being logged out due to 1024 limit, it would explain why only some users may have. | 17:33 |
| zbr | a cron would do | 17:33 |
| clarkb | but you'd still expect to end up in the disk cache since it is lru? | 17:33 |
| clarkb | but maybe your session was old enough to have rolled over in the lru or something | 17:34 |
| zbr | that is what I was wondering, maybe mine was old, now the question is how do you age a session? do they age from last usage or original creation time? | 17:34 |
| clarkb | you know as much as I do :) probably need to read the docs then cross check against the source code | 17:35 |
| zbr | because I am sure the previous page load was about 12h before that. | 17:35 |
| zbr | clarkb: i kinda feel sorry for you having to tune these, tuning live systems is not really a joy. | 17:37 |
| clarkb | ya, also its really difficult to get good data out of the test system (I've tried a few times to reproduce and then tune and see if we can monitor differences, but without real world activity its just not the same) | 17:37 |
| zbr | maybe lucam can have a look at our cache config and tell use if we made a mistake? somehow I believe that gerrithub.io may have more users. | 17:39 |
| clarkb | I think we have more changes than they do (but not sure about users) | 17:41 |
| clarkb | and yes I've been pinging upstream periodically for clarification, though still waiting on that last question I asked last week | 17:41 |
| *** rpittau is now known as rpittau|afk | 17:43 | |
| *** cloudnull has quit IRC | 17:47 | |
| *** cloudnull has joined #opendev | 17:47 | |
| clarkb | I need to step out for a bit then prepare for the meeting, back for that. | 17:48 |
| fungi | in the past 35 minutes the web_sessions cache entries have grown from 382 to 397 mem, 285323 to 285329 disk, and 128.02m to 128.03m space. cache hit ratios have remained the same (within the degree of precision reported anyway) | 17:51 |
| fungi | zbr: also keep in mind that the maxage for web_sessions may be how long a session is considered valid from the time of creation, not since time of last update? | 17:52 |
| zbr | anyone against making that 30days? | 17:54 |
| clarkb | the 7 days we've set seems reasonable | 17:55 |
| clarkb | login in every monday | 17:55 |
| clarkb | (or whatever your schedule is, and now really need to eat breakfast and catch up on my morning after the board meeting) | 17:55 |
| fungi | needing to re-log-in takes me a moment because i have to dig out my 2fa key and plug it in and retrieve a code, but still not so long that doing that once a week is a problem for me | 17:56 |
| zbr | my impression is that asks more often than this but i started to log the timestamps, anyway that is low-prio kind of issue, we have far bigger ones to address | 17:56 |
| fungi | yeah, i mean if you can confirm that we're logging sessions out way faster than weekly, that's probably good to dig into and figure out why | 17:58 |
| zbr | sure, i logged the approx time and remember to see. | 17:59 |
| zbr | without written records I do not trust my own memory or "feelings" | 17:59 |
| fungi | same here. it seems believable to me that i log into the gerrit webui roughly once a week, but i really can't say for sure | 18:00 |
| fungi | especially since i do most reviewing in gertty instead | 18:00 |
| fungi | and administration via the ssh command-line interface it provides | 18:01 |
| zbr | fungi: clarkb can we do the WIP experiment now? https://review.opendev.org/c/openstack/project-config/+/765821 -- is marked as WIP, add workflow to it and lets see if it goes in. | 18:01 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/project-config master: Revive os_monasca https://review.opendev.org/c/openstack/project-config/+/765799 | 18:01 |
| fungi | sure, taking a look | 18:01 |
| zbr | even if it goes in, there is no damage done. but we will know the answer to the the yesterday question. | 18:01 |
| *** hashar is now known as hasharDinner | 18:02 | |
| zbr | based on that we can decide if we need to implement changes or only document it. | 18:02 |
| *** dtantsur is now known as dtantsur|afk | 18:05 | |
| fungi | zbr: so worth noting, zuul enqueued it into the gate and reported verified +2 but did not submit it to merge | 18:25 |
| fungi | i think that means it's not taking wip status into account for enqueuing decisions | 18:25 |
| fungi | we might want to discourage use of the wip state in gerrit until we get it solved, in projects like tripleo that's likely to just cause more unnecessary delays and gate resets as people accidentally approve wip changes and then those make it all the way to the top of the gate and pass all their testing and then get their submit call rejected by gerrit in the end | 18:32 |
| fungi | (and so everything queued behind them has to be retested) | 18:33 |
| fungi | i'll add this to our tracking etherpad | 18:35 |
| zbr | fungi: i will try to make a PR tomorrow to make zuul avoid that, if I figure out how to do it. | 18:42 |
| fungi | zbr: ooh, thanks! corvus just mentioned a good path forward in #zuul if you're looking for where to start | 18:42 |
| *** andrewbonney has quit IRC | 18:51 | |
| zbr | I wonder if we can change the default sort-order on dashboard to consider Review-Priority flags. | 18:55 |
| zbr | now it only looks for Update, but it should be double key: RP first, update second. | 18:55 |
| clarkb | zbr: I think someone put a similar thought on the post upgrade notes etherpad and i suggseted an upstream bug | 19:00 |
| clarkb | as I expect that whatever ordering they are doing is somewhat intended | 19:00 |
| clarkb | (it almost looks like its based on the hierarchy of the config where the category comes from) | 19:00 |
| *** stevebaker has joined #opendev | 20:07 | |
| *** hasharDinner has quit IRC | 21:34 | |
| *** sboyron has quit IRC | 22:20 | |
| zigo | Hi guys! First, thanks for the new Gerrit. Now, I just want to know: do you intend at some point to have the gitweb link working again? I very much liked it, and miss it. | 22:23 |
| fungi | zigo: yep, we're collecting a list of stuff which needs solving post-upgrade at https://etherpad.opendev.org/p/gerrit-3.2-post-upgrade-notes and that's line 39 at the moment i think | 22:26 |
| zigo | Ah, thanks, good to know it's in the TODO. | 22:26 |
| fungi | yeah, the more immediate concerns have been related to stability and performance tuning under load | 22:27 |
| corvus | clarkb: sn8 test in 2m https://www.youtube.com/watch?v=nf83yzzme2I | 22:27 |
| corvus | this one looks like a spaceship and not a water tower or grain silo | 22:28 |
| openstackgerrit | Merged opendev/system-config master: Reduce gerrit heap limit to 44g https://review.opendev.org/c/opendev/system-config/+/766020 | 22:29 |
| zigo | fungi: In fact, what I really would appreciate, would be a direct link to the actual .diff file (and *not* a .diff.base64 or .diff.zip ...). | 22:29 |
| zigo | Because I very often just need to download the patch with wget -O debian/patch/bla.patch | 22:30 |
| zigo | :P | 22:30 |
| zigo | I don't know the internals of gerrit, but if it's simple to do, please do it. | 22:30 |
| fungi | seems like we can programmatically assign hyperlinks for stuff like the old gitweb linking, in which case it may be possible to link to the individual raw change diffs on gitea | 22:35 |
| fungi | but first we need to focus on getting everything working again | 22:35 |
| clarkb | corvus: did I miss it? /me about to eat some lunch | 22:35 |
| corvus | clarkb: raptor abort at t-1 | 22:36 |
| corvus | i'm guessing that means try again tomorrow | 22:37 |
| clarkb | oh too bad for them I guess | 22:37 |
| corvus | yep; livestream says done for the day | 22:39 |
| mordred | fungi: I have an old probably bitrotted change for updating the links to link to gitea instead of gitweb | 22:58 |
| mordred | fungi: I don't know if it will provide any use of not | 22:58 |
| mordred | fungi: https://review.opendev.org/c/opendev/system-config/+/723526 | 22:59 |
| openstackgerrit | Monty Taylor proposed opendev/system-config master: Use gitea for gerrit gitweb links https://review.opendev.org/c/opendev/system-config/+/723526 | 23:01 |
| *** tkajinam has joined #opendev | 23:01 | |
| fungi | mordred: oh, yep, i recall seeing that one in there though i assumed it may no longer be relevant for polygerrit (even the built-in gitweb linking seems to be broken in ours) | 23:11 |
| fungi | but definitely worth trying out | 23:11 |
| mordred | yeah - I have no idea of the relevance, but figured if that's an area you're thinking about - maybe that patch points in a direction that might shorten any work you consider doing. or maybe not :) | 23:11 |
| fungi | absolutely, thanks for writing it and for the reminder it's there! | 23:14 |
| clarkb | re gitweb ya I think we should ocnsider turning it off then I think the gitiles will default in | 23:52 |
| clarkb | then for gitea we can figure it out form there? | 23:52 |
| fungi | wfm | 23:56 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!