| *** DSpider has quit IRC | 00:20 | |
| *** keonwoolee has quit IRC | 00:50 | |
| *** Goneri has quit IRC | 01:31 | |
| *** qchris has quit IRC | 06:21 | |
| *** qchris has joined #opendev | 06:33 | |
| *** hillpd has quit IRC | 06:48 | |
| *** jrosser has quit IRC | 06:48 | |
| *** donnyd has quit IRC | 06:48 | |
| *** dviroel has quit IRC | 06:48 | |
| *** ajya|afk has quit IRC | 06:48 | |
| *** odyssey4me has quit IRC | 06:48 | |
| *** jrosser has joined #opendev | 07:25 | |
| *** donnyd has joined #opendev | 07:25 | |
| *** hillpd has joined #opendev | 07:25 | |
| *** dviroel has joined #opendev | 07:25 | |
| *** ajya|afk has joined #opendev | 07:25 | |
| *** odyssey4me has joined #opendev | 07:25 | |
| *** moppy has quit IRC | 08:01 | |
| *** moppy has joined #opendev | 08:02 | |
| *** DSpider has joined #opendev | 08:31 | |
| *** tosky has joined #opendev | 08:56 | |
| openstackgerrit | Sorin Sbarnea (zbr) proposed opendev/system-config master: Make install-from-bindep allow custom workdir https://review.opendev.org/751608 | 09:34 |
|---|---|---|
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Fix python-stow-versions https://review.opendev.org/751610 | 10:47 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed zuul/zuul-jobs master: Add support to use stow for ensure-python https://review.opendev.org/751611 | 11:31 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Run autoremove on post-install step https://review.opendev.org/751614 | 12:13 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Run autoremove on post-install step https://review.opendev.org/751614 | 12:14 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/diskimage-builder master: Run autoremove on post-install step https://review.opendev.org/751614 | 12:14 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed zuul/zuul-jobs master: Add support to use stow for ensure-python https://review.opendev.org/751611 | 12:33 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed zuul/zuul-jobs master: Add support to use stow for ensure-python https://review.opendev.org/751611 | 12:50 |
| *** user_19173783170 has joined #opendev | 14:48 | |
| user_19173783170 | hello, is someone online? | 14:48 |
| user_19173783170 | i want to ask what the differences between individuals and companies when i register my opendev account | 14:49 |
| fungi | user_19173783170: are you asking about contributor license agreements? or can you be more specific about where you are "registering your opendev account?" | 15:39 |
| fungi | user_19173783170: are you talking about this step? (signing up for an ubuntuone openid) https://docs.opendev.org/opendev/infra-manual/latest/gettingstarted.html#get-a-single-sign-on-openid | 15:43 |
| *** tosky has quit IRC | 15:45 | |
| fungi | user_19173783170: it's been a while since i set up an ubuntu one account for myself, but it doesn't look like they ask you to specify whether you're making an individual or company account, they just ask for an e-mail address | 15:45 |
| user_19173783170 | but if i submmit my code, how to use company's name? | 15:51 |
| fungi | user_19173783170: you could contribute with an e-mail address which has your company's domain on it | 15:51 |
| user_19173783170 | i see some info at https://www.openstack.org/join/ | 15:53 |
| user_19173783170 | What is Fundation Member Account used for? | 15:54 |
| fungi | user_19173783170: yes, opendev and the openstack foundation are separate, but many people who contribute to projects in opendev also set up an openstack foundation account | 15:54 |
| fungi | opendev doesn't use that for anything directly, but you can declare an affiliation in your profile there which we'll include in some demographic analysis we perform | 15:55 |
| fungi | main reasons to create an osf account are if you want to participate in the governance of the osf (vote in elections for or run for a seat on the osf board of directors) or any projects represented by the osf (vote in leadership elections, run for elected office) both of which require becoming a member of the foundation. you can also use an osf account without being a member if you want to coordinate your | 15:59 |
| fungi | schedule for their conferences or submit proposed talks for one of their conferences | 15:59 |
| fungi | but also if you want to declare company affiliations for your personal account, you can also do that in your osf account profile | 15:59 |
| fungi | opendev and the osf don't have "company accounts" for authentication, and sharing a single account among multiple people is a terrible idea whether they work at the same company or not | 16:00 |
| user_19173783170 | thanks a lot | 16:06 |
| user_19173783170 | i read for some time because my english is poor | 16:07 |
| fungi | user_19173783170: no need to apologize, english is not the first language for most people here | 16:08 |
| user_19173783170 | it's very kind of you ! | 16:10 |
| user_19173783170 | qqwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww | 16:10 |
| *** user_19173783170 has quit IRC | 16:11 | |
| *** tosky has joined #opendev | 16:48 | |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Run multiple zuul-web processes https://review.opendev.org/751426 | 19:26 |
| clarkb | SO_REUSEADDR is likely more complicated and needs more help from cherrypy so doing the brute force option instead | 19:28 |
| clarkb | (also SO_REUSEPORT is now a thing and I don't understand the difference) | 19:28 |
| fungi | heh, and apparently on stackoverflow someone has asked exactly that: https://stackoverflow.com/questions/14388706/how-do-so-reuseaddr-and-so-reuseport-differ | 20:02 |
| openstackgerrit | Clark Boylan proposed opendev/system-config master: Run multiple zuul-web processes https://review.opendev.org/751426 | 20:23 |
| clarkb | we have been using the config item for the port :) | 20:23 |
| clarkb | I think that means I can clean up the config a bit since port=%(ZUUL_WEB_PORT)s will be ignored if not in the zuul web context | 20:24 |
| clarkb | I thought it would be checked via something like voluptuous but the ini config isn't | 20:24 |
| clarkb | anyway lets see if that works and worry about cleanup later | 20:24 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed zuul/zuul-jobs master: Add support to use stow for ensure-python https://review.opendev.org/751611 | 20:45 |
| clarkb | yup that latest patchset works according to testinfra | 20:59 |
| clarkb | I'm not sure if that load balanced setup will work if we add auth to zuul-web but we don't do auth now so should be fine | 21:00 |
| fungi | if we hash by source address then that would probably still work | 21:02 |
| fungi | albeit not with as even a distribution across the backend listeners | 21:02 |
| clarkb | ya I figured we can sort that out if/when we add auth | 21:02 |
| clarkb | for now the simple setup should be fine | 21:02 |
| fungi | right, more pointing out we have options when we decide it's desirable | 21:03 |
| clarkb | reading that stackoverflow response it seems SO_REUSEPORT (with SO_REUSEADDR) is what we'd need to do what the earlier patchset attempted | 21:07 |
| clarkb | also zuul web is much happier today but I can't seem to induce a cached status on the opendev vhost | 21:13 |
| clarkb | something must still be going on there beyond just the slowness coupled with short ttl | 21:14 |
| fungi | yeah, unsurprising it's performing better, zuul-web is using ~10% of a vcpu according to top | 21:15 |
| fungi | as opposed to 100%+ | 21:15 |
| clarkb | I'm also tempted to restart zuul-web to pick up https://review.opendev.org/#/c/751425/ but I'm not sure I'll have any backup if it gets sad | 21:28 |
| fungi | i'm around | 21:28 |
| fungi | mowing the lawn before it gets too wet again, but around | 21:28 |
| clarkb | fungi: cool should I go ahead and do a zuul-web restart then? | 21:28 |
| fungi | go for it | 21:28 |
| fungi | happy to have the excuse to take a break if something breaks ;) | 21:29 |
| clarkb | its been restarted | 21:29 |
| clarkb | pulling up ff dev tools to see if status stops being loaded when the tab is inactive | 21:30 |
| clarkb | yup it seems to have corrected that issue | 21:31 |
| clarkb | I can see status get loaded on an interval then I switch tabs and it doesn't load again until I open that tab up again | 21:31 |
| clarkb | previously it just kept loading status | 21:31 |
| fungi | good, maybe that'll be enough to smooth things out this week | 21:33 |
| fungi | not that we should stop trying to work out the caching issue | 21:33 |
| clarkb | anything that helps is a goo dthing :) including bug fixes, caching, and more zuul-webs :) | 21:33 |
| fungi | but i have a feeling caching for that has been broken for a while, maybe even forever | 21:33 |
| clarkb | ya | 21:34 |
| clarkb | re Mowing, I need to mow and pick up apples (the wind storm early this week basically dropped the entire tree's worth of apples) but air outside is still bad | 21:35 |
| clarkb | we got two apple crisps out of it this year so I won't be too sad about it | 21:35 |
| clarkb | fungi: is Paulette going to give you trouble? | 21:36 |
| fungi | it's way, way, waaay too soon to guess. current nws/noaa/nhc projections have it running scared | 21:43 |
| clarkb | the internet wants everyone to know its 16e8 day (unix epoch time) | 21:43 |
| clarkb | er that was yseterday! | 21:43 |
| * fungi runs `date +%s` | 21:44 | |
| fungi | indeed, 1600033438 now | 21:44 |
| fungi | this time yesterday was 1599947126 | 21:45 |
| clarkb | would've been during sleep time for me I guess | 21:46 |
| clarkb | maybe not for you? | 21:46 |
| fungi | no idea | 21:46 |
| ianw | hey, sorry running late today | 22:36 |
| ianw | looking back at the zuul-web stuff which i seem to have missed | 22:36 |
| ianw | #status rebooted regionone.linaro-us mirror as it had gone into shutdown | 22:40 |
| openstackstatus | ianw: unknown command | 22:40 |
| ianw | #status log rebooted regionone.linaro-us mirror as it had gone into shutdown | 22:40 |
| openstackstatus | ianw: finished logging | 22:40 |
| fungi | it's certainly an intriguing situation. if you can work out why apache just doesn't cache the multi-tenant status api responses, that would be a breakthrough | 22:40 |
| fungi | (the regular expressions and locationmatch attempt are a red herring, even just specifying the openstack tenant path with cacheenable specifically so no wildcarding whatsoever, it still doesn't seem to even try) | 22:42 |
| ianw | clarkb: so nb03 is also shutoff in linaro, i'm presuming you didn't do that | 22:42 |
| clarkb | ianw: I did not | 22:43 |
| clarkb | its a bionic node to match the other buildres but the old one was xenial so maybe a kernel thing | 22:43 |
| ianw | yeah, the mirror is a focal node, so it seems unlikely to be both having issues at once | 22:44 |
| ianw | #status started nb03.opendev.org which had gone into shutdown | 22:47 |
| openstackstatus | ianw: unknown command | 22:47 |
| ianw | #status log started nb03.opendev.org which had gone into shutdown | 22:47 |
| openstackstatus | ianw: finished logging | 22:47 |
| fungi | yeah, i definitely did not shut off that server | 22:47 |
| fungi | maybe we need netconsole set up streaming from both instances | 22:47 |
| ianw | kevinz: ^ i think it might be useful for us to look through the logs and see what might have happened | 22:47 |
| ianw | fungi: yeah, i can setup that on a bridge screen | 22:48 |
| fungi | that could get us some useful details the next time it happens, thanks! | 22:52 |
| *** tkajinam has joined #opendev | 22:56 | |
| *** tosky has quit IRC | 23:14 | |
| ianw | fungi: the response from zuul has "Cache-Control: public, max-age=1" which seems to me to be saying the json file is valid only for one second. did we already consider that angle? | 23:28 |
| clarkb | yes, according to corvus the 1 second ttl is intentional but we should cache it for that long | 23:33 |
| ianw | [2020-09-13 23:32:58.490] "GET /api/status HTTP/1.1" 200 29653 cache hit "https://zuul.openstack.org/status" "Mozilla/5.0 (X11; Fedora; Linux x86_64; rv:79.0) Gecko/20100101 Firefox/79.0" | 23:33 |
| clarkb | over the weekend we may not get enough requests to hit that though I tried to induce it with wget | 23:33 |
| clarkb | ya thats the opendtack vhost which does work | 23:33 |
| ianw | ahh, right, so that's what the multi-tenant responses bit means | 23:34 |
| ianw | I wonder if this has to do with "ServerName zuul.openstack.org" | 23:51 |
| ianw | 200 98463 cache miss: attempting entity save "https://zuul.openstack.org/status" | 23:52 |
| ianw | the only one it seems to say that for is "zuul.openstack.org/status" | 23:52 |
| ianw | oh i see, zuul.opendev.org is in 000-default | 23:57 |
| ianw | ok, it's kind of confusing at first because they both log to the same file | 23:59 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!