[09:47:30] !log tools run `apt-get clean` to wipe 4GB of unused .deb packages, usage on / (root) was > 90% [09:47:33] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [09:47:46] !log tools run `apt-get clean` to wipe 4GB of unused .deb packages, usage on / (root) was > 90% (on tools-sgebastion-08) [09:47:48] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools/SAL [10:39:37] !log tools.awmd-stats Deployed 426f7f6 and restarted the webservice [10:39:38] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.awmd-stats/SAL [10:48:06] !log tools.awmd-stats Tool up and running again after restart [10:48:08] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.awmd-stats/SAL [16:10:45] Lucas_WMDE: you were using mosh to connect to toolforge? [16:11:51] arturo: yes [16:12:13] could you share your setup? :-P [16:12:24] specifically, all the bastion/proxy dance [16:13:20] Lucas_WMDE: I was expecting a drop-in replacement, but apparently doesn't work that way [16:13:51] it does for me, I have nothing mosh-specific in my config… [16:13:57] I think those are the most important snippets from my SSH config: https://gist.github.com/lucaswerkmeister/53abe8d26e4f4833bedfa09c164f289a [16:15:14] and the you just `mosh server.example.com`? [16:15:21] `mosh toolforge` [16:15:49] the ~/.ssh/config.d/toolforge file maps Host toolforge to HostName login-stretch.tools.wmflabs.org [16:15:57] (which reminds me, I should remove the -stretch from there now ^^) [16:16:55] oh, you don't use a bastion to redirect the connection, you just connect to the server -_- [16:17:28] oh yeah [16:17:35] connecting via mosh to any other host can’t work, they’re firewalled [16:17:42] only the bastion is reachable [16:18:39] I guess the bastion-cloud config in that gist is actually unrelated, nevermind that [16:18:44] got confused myself :D [21:49:14] can someone with root rights on the tools grid find out what PID 7551 on tools-sgeexec-0942.tools.eqiad.wmflabs is doing? [21:49:22] it’s from my tool (wd-shex-infer), but I’m not allowed to strace it [21:51:18] zhuyifei1999_? [21:51:53] * zhuyifei1999_ looks [21:54:15] https://www.irccloud.com/pastebin/sLC5alqz/ [21:54:22] it's CPU-bound [21:54:34] lucaswerkmeister: ^ [21:54:41] hm [21:55:44] it’s running this thing https://github.com/lucaswerkmeister/RDF2Graph/blob/b4f57f57752b372b342873f97265adc30d3649fb/Makefile#L14-L19 but that’s supposed to have a `sleep 5` between iterations [21:55:59] I’ll try to stare at it a bit longer and see where it could be burning all that CPU [21:56:12] but otherwise I’d better kill that job, it’s clearly not doing anything useful [21:56:37] thanks anyways [21:57:59] https://www.irccloud.com/pastebin/22kvPX7K/ [21:58:17] if that helps [21:58:41] that looks like there should be some clone()s or fork()s in the strace output… [22:02:23] I’ve stopped the job [22:02:40] zhuyifei1999_: can you exit gdb so the zombie can be reaped? :) [22:02:58] ok [22:05:20] !log tools.wd-shex-infer killed rogue job and disabled job submission via /data/scratch/wd-shex-infer/broken-2019-05-28 [22:05:22] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wd-shex-infer/SAL [22:10:09] https://usercontent.irccloud-cdn.com/file/bT4evsns/image.png [22:11:07] wrong channel, sorry 😂 [22:11:15] :D