[02:12:49] KittyReaper: Yes, there is a JRE available. [02:15:07] Ah, nice [04:19:21] df [04:21:49] 0 [05:49:45] Hi everyone - I'm able to make SQL queries fine when I test my bot on the bastions, but they fail on the execnodes. https://dpaste.de/N4AuH/raw/ [05:50:22] Coren: ^ [07:57:40] hm, i'm having a weird issue [07:57:52] so i'm trying to use TUSC to authenticate users for my web tool [07:58:04] it recently moved to labs at http://tools.wmflabs.org/tusc/tusc.php [07:58:15] when i access that in my web browser, it works fine [07:58:35] but if i try accessing it from labs [07:58:39] >>> import requests [07:58:40] >>> requests.get('http://tools.wmflabs.org/tusc/tusc.php') [07:58:45] i get [07:58:46] requests.exceptions.ConnectionError: HTTPConnectionPool(host='tools.wmflabs.org', port=80): Max retries exceeded with url: /tusc/tusc.php (Caused by : [Errno 110] Connection timed out) [07:59:48] the same exact python code works fine when i run it from my laptop, so there's something that when accessing it internally from labs that borks [08:01:03] (you can also confirm this using wget if you're not a python person) [08:04:22] legoktm: this is from tools-login? [08:04:28] yes [08:04:42] and on whatever server the cgi webserver is [08:06:09] must be a firewall somewhere [08:06:19] it's not DNS [08:11:23] weird. thanks for looking [08:19:03] [bz] (8PATCH_TO_REVIEW - created by: 2Ryan Lane, priority: 4Normal - 6enhancement) [Bug 43309] Add nagios check to ensure global nfs shares are shared properly from labstore1-4 - https://bugzilla.wikimedia.org/show_bug.cgi?id=43309 [09:21:43] legoktm: doesn't it looks like python? [09:21:52] huh? [09:22:27] it seriously does look like py [09:22:34] it's not a python issue [09:29:14] legoktm: how long you ran it [09:29:26] i ran it use [09:29:31] using wget* [09:29:35] once or twice [09:29:44] I'm running on tools-dev and it goes on for seconds [09:29:59] I mean time [09:31:12] Hi, does anybody tried to access the enwiki sql database via a pbs job in the last days? Since Friday I am getting an error "Can't connect to MySQL server on 'enwiki.labsdb' (110)", but I can connect directly ("sql enwiki") or over a python script over the command line [09:31:33] legoktm: ? [09:32:00] nette: yup, Earwig reported that a bit earlier. it works on some exec nodes, but not all it seems [09:32:09] zhuyifei1999: i don't understand your question [09:32:31] nette: aye; it seems 01-04, 07, 08 are fine and 05 and 06 are broken [09:32:34] don't know exactly what's wrong [09:33:04] ahh, thanks that's good to now! Yes, my jobs were always running over 06 [09:33:22] legoktm: how many seconds passed before the exception appears? [09:33:50] zhuyifei1999: like 20-30 seconds usually, it's a timeout error [09:49:50] nice, it is working. Thanks a lot and have a nice Sunday!!! [09:53:29] legoktm: try this: import subprocess [09:53:54] subprocess.Popen("php /data/project/tusc/public_html/tusc.php", shell=True, stdout=subprocess.PIPE).stdout.read() [09:54:17] that is most definitely not the right way to do that. [09:54:45] i shouldn't be executing someone else's tools [09:54:59] plus it needs to run from a webserver, I'm going to be posting data from it [09:55:16] urllib2 tested, doesn't work [10:00:41] legoktm: is there a doc? [10:01:20] doc for what? [10:01:27] for tusc api [10:04:04] the issue isn't TUSC, it's with labs [10:15:36] I know, but how to use the tusc api? [10:17:57] read the info at the bottom: http://tools.wmflabs.org/tusc/index.html [10:23:09] petan: ping [10:28:08] gfdhd [10:40:24] petan: is there a firewall that blocks connecting http://tools.wmflabs.org from shell? [10:40:37] no [10:40:44] but nat is [10:40:52] you need to connect to some webserver instead [10:40:58] like tools-webserver-01 [10:41:03] grr nat [10:41:24] I hate nat [10:42:26] legoktm: ^^ [10:45:08] petan: is there a reason not to allow access to tools.wmflabs.org from within tools? [10:45:25] it's a limitation of nova or something like that [10:45:26] valhallasw: nat [10:45:35] it's not that it's not allowed, it's technically not easily done [10:46:33] we could eventually put tools.wmflabs.org to /etc/hosts and make it point to proxy server [10:46:37] nat does not allow connection from inside through it to inside [10:46:38] petan: and changing the iptables or /etc/hosts of the tools-* hosts so the... that, yes. [10:47:08] yes I will create a bug for that, I need to discuss that with Coren I guess [10:47:12] @notify Coren [10:47:13] This user is now online in #wikimedia-labs. I'll let you know when they show some activity (talk, etc.) [10:47:26] I will do that on -dev let's see if something break [10:47:27] legoktm: requests.get('http://tools-webserver-01/tusc/tusc.php') [10:47:59] petan: I'm on -dev, don't shut down [10:48:39] I never shut down servers no worries [10:49:08] since long time it's never really needed to reboot linux servers, even kernel updates can be done online these days [10:49:23] just old-school sysadmins do reboots :P [10:49:34] and Damianz when he moves his house [10:50:36] yes it works [10:50:52] zhuyifei1999 on -dev you should be able to directly download from tools.wmflabs.org now [10:51:38] @seen magnus [10:51:38] Steinsplitter: I have never seen magnus [10:51:43] @seen magnus___ [10:51:43] Steinsplitter: Last time I saw magnus___ they were quitting the network with reason: Ping timeout: 250 seconds N/A at 8/12/2013 3:02:55 PM (12d19h48m48s ago) [10:51:50] petan: It's actually legoktm who wants that [10:52:58] petan: works :) [10:55:29] Coren: I need to insert this to puppet, how can I do that? [11:12:55] petan: would it be possible to add a vhost for pywikipedia.org to tools.wmflabs.org/pywikibot/ ? [11:17:07] labs is slow today [11:17:54] valhallasw: yes, but that also should be consulted with Coren [11:18:05] I don't know if there isn't a legal restriction for that [11:18:51] for now I recommend you to create some redirect rule on pywikipedia.org [11:20:32] petan: ok [11:23:43] Steinsplitter: you really should try @seenrx [11:23:46] @seenrx magnus [11:23:47] petan: Last time I saw magnuse_ they were quitting the network with reason: no reason was given at 9/22/2012 6:09:38 AM (337d5h14m8s ago) (multiple results were found: magnuse, magnus__, magnus___, magnus____, magnus_) [11:24:00] @seen magnus_ [11:24:00] petan: Last time I saw magnus_ they were joining the channel, they are still in the channel #mediawiki at 8/13/2013 10:46:32 AM (12d37m27s ago) [11:24:10] you see? magnus is online [11:25:09] @seenrx [11:25:14] @seenrx test [11:25:14] Steinsplitter: Last time I saw test_ they were leaving the channel #wikipedia-zh at 8/24/2013 1:17:11 PM (22h8m3s ago) (multiple results were found: testobsessed, test123234, test, test-sz, testgnort and 38 more results) [11:25:29] @seenrx magnus [11:25:30] Steinsplitter: Last time I saw magnuse_ they were quitting the network with reason: no reason was given at 9/22/2012 6:09:38 AM (337d5h15m51s ago) (multiple results were found: magnuse, magnus__, magnus___, magnus____, magnus_) [11:44:22] @seenrx zhuyifei1999 [11:44:22] zhuyifei1999: are you really looking for yourself? [11:44:29] @seenrx zhuyifei1999_ [11:44:30] zhuyifei1999: Last time I saw zhuyifei1999_ they were quitting the network with reason: Ping timeout: 246 seconds at 8/25/2013 10:02:04 AM (1h42m25s ago) (multiple results were found: zhuyifei1999_, zhuyifei1999_zzz, zhuyifei1999__) [14:03:27] Anyone around who could complete https://wikitech.wikimedia.org/wiki/New_Project_Request/accessibility-dev maybe? [16:42:18] hey nerus [16:42:34] hey YuviPanda [16:42:39] how've you been? [16:42:43] :P [16:42:48] :D [16:42:55] had any time to work on the status stuff? [16:43:36] busy week at work. gonna work some more tonight :) [16:43:46] heh, okay! [17:49:17] @notify scfe_de [17:49:17] I'll let you know when I see scfe_de around here [17:49:19] ok [17:51:00] [bz] (8ASSIGNED - created by: 2Maarten Dammers, priority: 4Normal - 6normal) [Bug 48851] Database inserts are slow - https://bugzilla.wikimedia.org/show_bug.cgi?id=48851 [18:24:50] petan: You online? [20:39:45] that seems...silly [20:41:08] !blink [20:41:15] .. [20:41:18] legoktm: what? [20:41:22] seemed silly? [20:42:44] [03:47:27 AM] legoktm: requests.get('http://tools-webserver-01/tusc/tusc.php') [20:43:01] it works, but…i should just be able to do tools.wmflabs.org/blah [20:43:16] you should. aren't you able to? [20:44:03] nope, doesn't work [20:44:08] see the logs above [20:44:35] legoktm: ah, yes. I remember Coren telling me something about how wmflabs.org doesn't actually work *inside* labs [20:45:05] indeed. do 'curl wikistream-tools.wmflabs.org' [20:45:08] just hangs [20:45:34] yup. [20:45:40] so known issue. poke Coren [20:46:02] i did :P [20:46:33] fine! :P [20:46:40] legoktm: btw, if you are going to do the internal URL [20:46:43] use tools-webproxy [20:46:47] rather than a particular server [20:47:00] although webproxy itself will be replaced by tools-proxy at some point :P [20:47:04] er [20:47:05] so [20:47:15] tools-webproxy/tusc/tusc.php? [20:47:34] yesyup [20:47:37] for now [20:47:40] it will break in the future! [20:48:07] ok, just ping me when it does [20:48:21] legoktm: will do :) [20:48:37] will also spam all relevant lists [21:49:13] YuviPanda: It's not so much a "known issue" as it is a "limitation of the OpenStack networking layer"; you can't reach public (floating) IPs from inside. [21:49:22] ewww [21:49:30] Coren: can that be fixed at all? [21:50:03] IIRC, Ryan said upstream did that by design so no. [21:50:34] so how do we fix legoktm's use case? [21:50:44] promise that tools-webproxy will never go away? [22:16:21] YuviPanda: Or at least that the name will always point somewhere sensible. [22:16:49] hmm, so if we replace tools-webproxy with tools-proxy at some point, can we still make tools-webproxy alias to the same machine? [22:16:57] sigh, still no scfe_de [23:17:34] [bz] (8PATCH_TO_REVIEW - created by: 2Ryan Lane, priority: 4Normal - 6enhancement) [Bug 43309] Add nagios check to ensure global nfs shares are shared properly from labstore1-4 - https://bugzilla.wikimedia.org/show_bug.cgi?id=43309 [23:17:35] [bz] (8ASSIGNED - created by: 2Maarten Dammers, priority: 4Normal - 6normal) [Bug 48851] Database inserts are slow at the replicated databases - https://bugzilla.wikimedia.org/show_bug.cgi?id=48851