[00:08] *** Start has quit IRC (Excess Flood) [00:08] *** Start has joined #internetarchive.bak [00:08] *** svchfoo1 sets mode: +o Start [00:20] *** Start has quit IRC (Read error: Connection reset by peer) [00:20] *** Start_ has joined #internetarchive.bak [01:34] *** zottelbey has quit IRC (Remote host closed the connection) [01:36] *** VADemon_ has quit IRC (Read error: Connection reset by peer) [01:43] *** niyaje4 has joined #internetarchive.bak [02:41] *** Start_ is now known as Start [02:41] *** Start has quit IRC (Disconnected.) [02:41] *** Start has joined #internetarchive.bak [02:41] *** svchfoo3 sets mode: +o Start [03:28] *** niyaje4 has quit IRC (Ping timeout: 600 seconds) [04:54] We just had some discussion about this. [04:54] (Even brewster's seeing it.) [04:54] We pull from all over the infrastructure, so this is something worth noticing. [04:55] When closure returns, I'd like to know if we need to get more clients. Because we seem to be stalled on downloads. [04:55] Could be slowness, but that time graph is pretty damning. [04:55] They really should be climbing, unless we're intentionally waiting at a certain point for number of clients. [04:56] where's the IA network weathermap again? [04:56] https://monitor.archive.org/weathermap/weathermap.html [04:56] ah ha [04:57] wow, that's some outgoing [05:08] *** niyaje4 has joined #internetarchive.bak [05:09] Yeah, we don't fuck around. [08:59] *** niyaje4 has quit IRC (Ping timeout: 600 seconds) [13:07] *** zottelbey has joined #internetarchive.bak [13:58] SketchCow: the graph is also giving us the number of files, rather than the number of bytes [13:59] the large vertical jumps could be people not reporting in regularly [14:06] ive started download of shard2 now [14:42] reports are hourly right? [14:52] yes [14:53] when you run iabak we start a background process that does git annex sync once per hour [14:53] but it's not really very robust; if it dies, or is killed, or the computer is rebooted, or any number of other things, then it won't be restarted until you run iabak again [14:54] it really should be a cron job [14:54] but we haven't gotten around to it yet [15:01] are cronjobs available on *bsd/osx? [15:01] yep [15:03] even with non-root-access? [15:03] yep [15:06] every user has their own crontab [15:14] what happens if cronjobs are not available? [15:17] then we're in windows [15:17] and everything is a lot more work [15:39] db48x: my custom LFS-OS does not have cronjob installed [15:57] *** kyan has quit IRC (Quit: Leaving) [16:31] it'll be cron and crontab [16:31] cron is the daemon which runs the jobs [16:31] crontab is a helper program which lets you edit the job specifications (also called a crontab, or cron job table) correctly [16:32] the alternative it to use systemd timers [16:32] they're actually nicer in some ways [16:32] but not as widely available [16:55] so, I've added a feature that will let us have a leaderboard [16:55] :D [16:57] what it provides is a list of uuids of git-annex repos, and the size of the data they're storing [17:12] midas: your iabak seems to be running an old version of git-annex. could you please update it [17:13] also if someone here has username jdamery at home, you too [17:13] also Sean Palmer, and someone running it as root [17:15] as root> tut tut [17:15] meh if it's a vm who cares [17:16] yipdw: defense in depth [17:16] hopefully you're using selinux as well [17:16] maybe but I find that "fuck it, it works" overrules that every time [17:16] it's not me btw [17:16] and giving iabak it's own security label [17:16] which is more pain than running it as its own user [17:16] and making sure the policy prevents it from doing crazy things [17:17] like connecting to any ports except ssh and http [17:17] https, that is [17:18] it's probably also worth noting that a large percentage of the web runs on systems that you can log into with ec2-user@ and have passwordless sudo to root :P [17:18] so far, no major disasters that we know of [17:18] but digression etc [17:19] install-git-annex seems to think I have the latest version -- 20150409 [17:21] chr [17:21] oops [17:57] *** Start has quit IRC (Read error: Connection reset by peer) [17:57] *** Start has joined #internetarchive.bak [17:57] *** svchfoo3 sets mode: +o Start [18:23] http://iabak.archiveteam.org/stats/SHARD1.leaderboard [18:23] oh, look, IA is leading [18:24] heh [18:26] missing argument to sendStat [18:27] missing argument to sendStat [18:27] seen at the end of a shardstats-all run [18:27] closure: seems that my pubkey is missing for shard2 [18:27] closure: run it with -x? [18:28] 1517859797619: e3faaa97-6ec6-4dc3-85e3-0295dfcae38b -- [18:29] has no user/host in shard1? [18:29] Aand enter! [18:34] db48x: ok, it's because ALL.size is empty [18:35] shouldn't be [18:35] ah, I broke it [18:36] I changed the units from tb to bytes [18:37] that should be fine [18:37] broke the parser, I'll fix it up [18:37] oh, the regex [18:38] what's with all the places it uses su? [18:39] closure: no need to convert it to terabytes [18:39] those files are used on the html page.. [18:39] just use bytes throughout, and we'll change the graph [18:40] it su's because the SHARD users own their repos [18:43] testing changes to this script keeps getting harder and harder [18:43] why does git annex info . need to su? [18:44] because it might merge git-annex branch changes and write files [18:46] SI units?! [18:46] it alreasy was [18:46] * db48x gasps [18:47] is nothing sacred any more? [18:48] http://source.git-annex.branchable.com/?p=source.git;a=blob;f=Utility/DataUnits.hs;h=6e40932ef5a8801479277bf469d9032e43a0b6ee;hb=HEAD [18:49] heh [18:49] How awkward would it be to change the uid I run iabak under? :-) [18:49] Senji: chown -R [18:50] It won't care that the username doesn't match the one in the repo location any more? [18:50] ssh might need you to say yes to a host key again [18:50] no [18:50] lol @ line 102 [18:50] Cool. [18:51] db48x: I'm bitter, I once bought an ISDN line and didn't understand their absurd units [18:52] ah [18:54] grrr: ./shardstats: line 70: syntax error near unexpected token `fi' [19:16] hrm [19:17] putting units in the leaderboard instead of bytes is problematic [19:17] oh, that's just me [19:18] where is shardstats getting an upgraded git-annex from? [19:19] I temporarily threw it in /usr/bin [19:20] ah [19:20] must have done that after I tried it [19:46] testing these three lines took way longer than writing them [19:47] ok, lunch [19:49] *** SN4T14_ has joined #internetarchive.bak [19:53] *** SN4T14__ has quit IRC (Ping timeout: 306 seconds) [19:58] ht damn [20:38] *** zottelbey has quit IRC (Remote host closed the connection) [20:43] I seem to be getting some 403 Forbidden errors from the web [20:49] yeah, those happenn [20:49] http://iabak.archiveteam.org:8080/render/?width=836&height=522&_salt=1428871679.369&target=iabak.shardstats.leaderboard.*.all&title=leaderboard&lineMode=connected&hideLegend=true [20:54] I prefer http://iabak.archiveteam.org:8080/render/?width=1060&height=717&_salt=1428872058.728&target=keepLastValue%28highestAverage%28iabak.shardstats.leaderboard.*.shard2%2C10%29%29&from=-2hours [20:55] any way to label the lines on the y axis instead of in the legend? [21:02] I don't believe so [21:03] heh: http://iabak.archiveteam.org:8080/render/?width=1060&height=717&_salt=1428872569.057&from=-2hours&target=legendValue%28keepLastValue%28highestAverage%28iabak.shardstats.leaderboard.*.all%2C10%29%29%2C%22last%22%29 [22:39] oops, forgot to pull that leaderboard stat update [22:39] oh, it's in. I wonder why no lines have moved yet [22:40] whats the leaderboard url? [22:41] http://iabak.archiveteam.org:8080/render/?width=1060&height=717&_salt=1428872058.728&target=keepLastValue%28highestAverage%28iabak.shardstats.leaderboard.*.shard2%2C10%29%29&from=-2hours and http://iabak.archiveteam.org/stats/ALL.leaderboard [23:02] *** kyan has joined #internetarchive.bak [23:20] *** Start-mob has joined #internetarchive.bak [23:24] *** Start-mob has quit IRC (Remote host closed the connection) [23:34] *** balrog has quit IRC (Quit: Bye)