[09:40:18] !log toolsbeta livehacking toolsbeta-sgebastion-04 (https://gerrit.wikimedia.org/r/c/566045 and https://gerrit.wikimedia.org/r/c/565575) and toolsbeta-proxy-1 (https://gerrit.wikimedia.org/r/c/565556) for testing T234617 [09:40:21] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Toolsbeta/SAL [09:40:21] T234617: Toolforge. introduce new domain toolforge.org - https://phabricator.wikimedia.org/T234617 [10:24:02] !log admin running `sudo systemctl restart apache2.service` in both labweb servers to try mitigating T240852 [10:24:05] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Admin/SAL [10:24:05] T240852: CloudVPS: horizon giving http/500 intermitently - https://phabricator.wikimedia.org/T240852 [12:49:36] !log toolsbeta cleanup livehackings in toolsbeta-sgebastion-04 and toolsbeta-proxy-1 [12:49:38] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Toolsbeta/SAL [17:21:26] !log tools rebooting toolschecker to recover stale nfs handle [17:21:29] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [17:43:15] !log admin remounting /mnt/nfs/dumps-labstore1007.wikimedia.org/ on all dumps-mounting projects [17:43:18] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Admin/SAL [18:30:54] !log tools depooling and rebooting tools-sgeexec-[0911,0913,0919,0921,0924,0931,0933,0935,0939,0941].tools.eqiad.wmflabs [18:30:56] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [19:17:54] !log tools depooled and rebooted tools-sgeexec-0914 because it was acting funny [19:17:56] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [19:25:02] !log tools hard rebooting tools-sgeexec-0913/14/35 because they aren't even on the network [19:25:04] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [19:46:19] Cyberpower678, (slightly dumb question, but the VPT thread is getting long) Have you tried turning reFill off and back on again? [19:46:32] Yes [19:46:47] I only know of the webservice though. [19:47:34] The documentation of the tool is less than helpful, so it leaves me effectively unaware of any backends that should be rebooted. [19:53:25] Cyberpower678, AntiComposite what is the problem? [19:54:19] !log puppet-diffs syncing puppet facts from local puppet repo [19:54:21] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Puppet-diffs/SAL [19:55:30] The https://tools.wmflabs.org/refill/ landing page wasn't loading at all for me before, it is now, so that's a start [19:56:12] https://phabricator.wikimedia.org/T213515 says that refill-api is the backend [19:56:27] that appears up too: https://tools.wmflabs.org/refill-api/ [19:57:25] Swagger is, but POSTs to https://tools.wmflabs.org/refill-api/fixWikipage 500 [20:32:08] question: for tools hosted on toolforge, if I as the tool developer don't do any data collection -- i.e. no analytics tracking scripts etc. -- is there a privacy policy to point to? i presume some information is being logged somewhere even if I haven't done it explicitly? the best I can think of is essentially saying "No additional personal data is collected per https://wikitech.wikimedia.org/wiki/Wikitech:Cloud_Services_Terms_of_use" [20:34:36] AntiComposite, yeah I don't know what's up with that one [20:34:57] app is clearly serving HTTP 500 in response to a POST to /refill-api/fixWikipage - it's not happening at the proxy or anything [20:38:12] Yeah [20:39:18] The status endpoint doesn't error out immediately, can't test it properly because there are no jobs [20:39:47] https://github.com/zhaofengli/refill/blob/115a5c42e31b65f841f30e16ffce5c81eee59036/backend/app.py#L116 is the deployed code right now [20:42:07] the thing is this app is clearly serving some endpoints just fine [20:42:10] There should be a ton of exceptions in /data/project/refill-api/uwsgi.log , I can't see it because of file permissions [20:42:54] AntiComposite, not when I try to make the request [20:43:44] oh there we go [20:43:47] just had to be patient [20:43:51] * Krenair shakes tiny fist at NFS [20:44:05] Oh hey, I'm getting a 202 now [20:44:55] AntiComposite, https://phabricator.wikimedia.org/P10239 [20:45:28] sounds like maybe it's actually supposed to be serving 4xx to this request, but there's some bug that makes it give HTTP 500 :) [20:48:50] Yeah, it's expecting api.payload to always be dict-like, but doesn't handle the case where api.payload is None [20:51:54] isaacj: that sounds fine to me. As you probably found we don't really have any strong guidance for tools about their local privacy/TOU statements [20:52:47] AntiComposite, that even is the payload attribute though? [20:52:58] what* [20:54:16] thanks bd808 : yeah, kinda wild-westy out there but also I appreciate the flexibility of the ecosystem :) just wanted to make sure i wasn't misleading users [20:55:42] Krenair, it /should/ be the data from the POST, decoded into a Python dict [20:55:56] by flask_restplus [20:57:38] AntiComposite, well, I still got HTTP 500 when sending it empty dict [20:57:57] I get a 400 [20:59:19] AntiComposite, $ curl -X POST 'https://tools.wmflabs.org/refill-api/fixWikipage' --data {} [20:59:19] {"message": "Internal Server Error"} [21:01:12] Hmm...it looks like it wants a "Content-Type: application/json" header attached [21:02:16] $ curl -X POST 'https://tools.wmflabs.org/refill-api/fixWikipage' --data {} -H "Content-Type: application/json" [21:02:16] {"message": "The browser (or proxy) sent a request that this server could not understand."} [21:02:23] that is 400 though [21:02:40] Which is the expected behavior [21:02:44] yes [21:02:46] cool [21:03:21] The lack of a header issue is an upstream problem, but it could be handled better [21:05:14] Was that the only recent traceback in the log, or were there different ones? [21:05:26] * AntiComposite still isn't sure why it stopped working and is working now [21:11:02] AntiComposite, I don't know either, while fiddling around with it I did delete the pod so maybe it getting started again helped in some way [21:11:32] Probably [21:12:58] https://www.youtube.com/watch?v=rksCTVFtjM4 [21:13:22] lol [21:41:24] * AntiComposite wonders out loud why reFill appears to be targeting python3.6 [21:53:20] AntiComposite, as opposed to 3.7/3.8/3.9? [21:54:04] I'd usually expect 3.5 or 3.7 on Toolforge [22:03:51] oic, celery didn't support 3.7 1-2 years ago [22:53:31] !log openstack deactivated image debian-9.1-stretch (deprecated 2017-11-16) [22:53:33] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Openstack/SAL