[00:15:45] Coren: I'm nearly positive labstore3/4 already have NFS rules set up [00:15:49] they were used for gluster [00:16:06] Ryan_Lane: From ae0-105.cr1-sdtpa.wikimedia.org (10.4.16.252) icmp_seq=1 Packet filtered [00:16:33] Can't ping, and tcpdump don't see the packets arriving. [00:16:48] can't ping from in labs? [00:17:02] Nope. [00:17:19] icmp isn't allowed [00:17:20] NFS works [00:17:38] tcpdump calls you mistaken. :-) [00:17:54] from labs: showmount -e labstore1.pmtpa.wmnet [00:18:14] no clue why it isn't working for labstore4 [00:18:22] Perhaps that's just labstore*1*? [00:18:37] nope [00:18:40] works on 2 as well [00:18:46] and the filters are the same for all 4 [00:18:49] is portmap running? [00:18:55] Well, labstore3 times out. [00:19:07] hm [00:19:08] weird [00:19:21] portmap works, actually. [00:19:24] oh [00:19:30] you need to set the ports [00:19:38] Oh! You're not using the defaults! [00:19:42] nfs doesn't work through filters or firewalls by default [00:20:00] It does if you use the standard ports. :-) [00:20:00] you have to set the services to run on specific ports [00:20:05] Which ones do you guys use? [00:20:16] no it doesn't, because portmap will assign most services random ips [00:20:17] err [00:20:18] ports [00:20:41] I know, but there are usual ports to set them to. :-) Nevermind, just tell me the ones you guys use so I can close the ticket. [00:20:53] no clue [00:21:01] check the config of labstore1/2 ;) [00:21:03] it's in puppet [00:21:34] I see 'em. [00:21:48] 32769-32767 [00:21:51] tsk, tsk. [00:22:41] I took this from ubuntu docs, btw [00:22:51] to make sure I was doing it in ubuntu's standard way [00:22:58] there are no "correct" standard set of ports for this [00:23:33] No, but whichever you choose should be >=49152 [00:24:03] Well, in theory anyways. :-) [00:24:23] why? [00:24:31] to be outside of well known ports? [00:25:16] I don't think I've seen any nfs docs that use ports that high :) [00:29:50] Heh. RFC 6056 commands; though I agree I've yet to see anything besides X and IRC obey. :-) [00:30:09] (They used to be much higher than 6k once upon a time) [00:31:05] heh [00:31:17] for the most part well known ports is ignored :) [00:40:28] !log rebooting labstore3 -> silly lockd module [00:40:28] rebooting is not a valid project. [00:45:24] :D [00:53:41] Ryan_Lane: Almost, but not quite. Apparently, even the normally fixed mountd port is changed, and that's not in puppet. [00:53:51] o.O [00:54:01] I think it's gluster doing it, isn't it? [00:54:01] that's due to gluster, then [00:54:17] * Coren tries to find where /that/ is configured [00:54:25] heh [00:54:36] well, ideally we'll set it to that now and change it to the default [00:57:12] Oh, man, you guys keep reusing port 32768 for all sorts of different things! [00:57:54] what other things are we using it for? [00:58:06] nginx, for one [01:02:28] Actually, now that I look at gluster, it uses /its/ defaults, which aren't the same ones you use with your normal nfs servers either. :-) [01:02:43] 100005 3 tcp 38465 mountd [01:02:43] 100005 1 tcp 38466 mountd [01:02:43] 100003 3 tcp 38467 nfs [01:02:43] 100024 1 tcp 55659 status [01:02:43] 100024 1 udp 44153 status [01:02:43] 100021 4 tcp 38468 nlockmgr [01:02:43] 100021 1 udp 750 nlockmgr [01:03:06] nginx uses it? for what? [01:03:22] for sending logs maybe? [01:03:49] Oh, nevermind. I misread: [01:03:50] $nginx_worker_connections = '32768' [01:04:06] 32k connections, not port 32k. :-) [01:17:11] Change on 12mediawiki a page Wikimedia Labs/Account creation improvement project was modified, changed by Ryan lane link https://www.mediawiki.org/w/index.php?diff=672841 edit summary: [+126] /* OpenID as a provider */ [01:18:24] Odd. [01:18:49] Change on 12mediawiki a page Wikimedia Labs/Account creation improvement project was modified, changed by Ryan lane link https://www.mediawiki.org/w/index.php?diff=672842 edit summary: [+60] /* OpenID as a provider */ [01:23:29] * Coren tries to understand why the [bleep] tools-login still tries to mount via port 2049 [01:23:46] Hm. Portmapping cache? [01:35:47] maybe? [01:36:16] where are you trying to mount it? [01:36:37] when you want to override /data/project in your project you can use local autofs overrides [01:36:48] in auto.master/auto.data [01:37:43] we'll need that to be puppetized [01:38:11] then we can do one project at a time by modifying instances' puppet config [01:43:23] Yeah, right now I'm just testing through /mnt [01:43:32] And... I haz a sukses! [01:43:48] labstore3.pmtpa.wmnet:/srv/tools 37T 33M 37T 1% /mnt [01:45:24] Heh. I win. [01:45:31] projectstorage.pmtpa.wmnet:/tools-project 38083885056 10103715072 27980169984 27% /data/project [01:45:31] labstore3.pmtpa.wmnet:/srv/tools 39056055296 33792 39056021504 1% /mnt [01:54:09] how does it have more space than gluster? [02:00:32] Coren: you may want to change the cluster name for this [02:00:36] for labstore3/4 [02:00:51] maybe labsnfs for the cluster, and a new multicast group for it? [02:01:02] it's in the gluster's ganglia cluser now [02:01:06] *cluster [02:01:13] Ah, good point. [02:01:14] err [02:01:22] it's in gluster's ganglia cluster now [02:01:34] * Coren tries to find that. [02:01:44] heh [02:01:58] decent amount of system cpu use: https://ganglia.wikimedia.org/latest/graph_all_periods.php?h=labstore3.pmtpa.wmnet&m=cpu_report&r=hour&s=by%20name&hc=4&mc=2&st=1365818411&g=cpu_report&z=large&c=Glusterfs%20cluster%20pmtpa [02:02:10] Different layout, remember? I don't have an OS disk on the shelves. [02:02:31] It's using CPU because it's rebuilding the raid now. [02:02:35] ahhhh. ok [02:07:16] * Coren starts setting up the timetravel now. [02:07:30] ok. off for the night [02:07:31] * Ryan_Lane waves [02:09:37] * Coren waves. [02:52:02] Coren- ... So if something on tools edits while logged out, it will be coming from an IP address like '10.64.0.126'? [02:57:31] anomie: yes [02:58:06] legoktm- Any idea what the full possible range is? [02:58:26] not sure, all ive seen is https://www.wikidata.org/wiki/Special:Contributions/10.64.0.127 [03:14:16] Ryan_Lane: can has transwiki? or Coren ? [03:14:45] .. transwiki? [03:14:48] 12 18:26:05 < jeremyb_> want to transwiki https://en.wikipedia.org/wiki/Module:String ? [03:15:05] To wikitech? [03:15:10] yes please :) [03:15:18] * Coren tries to figure out the right userright. [03:15:25] sysop should be plenty [03:15:32] [[special:import]] [03:15:38] idk if it has import sources set up [03:16:30] Wha' be yer wikitech username again? [03:17:05] Nevermind, found you. :-) [03:17:22] You should have import now. Dunno about sources. [03:17:25] hah [03:17:36] i thought you'd just do it for me [03:17:38] * jeremyb_ tests [03:17:59] There's a happy fun userright just for that jazz on wikitech: contentadmin. [03:18:09] orly [03:18:43] gah, /me stabs the Echo that don't work with monobook [03:19:01] ~= enwp admin; doesn't have openstack magic or editinterface. [03:19:24] enwiki admin does have editinterface though [03:19:27] afaik [03:19:33] Which is why ~= and not == :-) [03:22:06] heh [03:22:28] ohhh. maybe transwiki can't do contenthandler? [03:23:48] huh. but it saved... [03:23:52] let's see what happens [03:23:55] > Import failed: Can't save non-default content model with $wgContentHandlerUseDB disabled: model is wikitext , default for Module:String/enwiki/Module:String is Scribunto [03:24:18] (that was on mediawikiwiki. which is the only source enabled on wikitech. now on to wikitech) [03:24:48] huh, wikitech did it without error [03:25:13] but probably MW version is wikitech < enwiki < mediawikiwiki [03:26:10] wooot, and it works [03:34:51] hrmmmmm. can't quite figure out how to use replace [03:35:05] anomie seems not too idle [03:35:55] jeremyb_- ? [03:36:19] anomie: how do i use replace from [[Module:String]] ? [03:36:56] anomie: https://wikitech.wikimedia.org/w/index.php?title=Sandbox&oldid=66494 [03:37:41] hrmmm, maybe i need source= instead of s=. but i've no idea why [03:37:56] examples would be lovely :) [03:38:09] Two errors there: the parameter is named "source", not "s". And lua uses % rather than \ for its patterns, so it should be "pattern=%d" not "pattern=\d" [03:38:18] ewwww [03:38:29] Manual: https://www.mediawiki.org/wiki/Extension:Scribunto/Lua_reference_manual [03:38:45] wooo, that looks maybe helpful [03:39:09] i think replace is missing from that last link [03:39:40] or there's multiple things named string [03:40:42] Coren: This is odd. It looks like some packages are only installed on tools-exec-01 and not tools-exec-02. For example, /usr/share/perl5/JSON.pm seems to only be on -01. [03:41:05] That's annoyingly possible. [03:41:06] jeremyb_- Module:String is not the same thing as the string library in Lua. [03:41:13] anomie: The first few installs were manual [03:41:19] * Coren fixes that [03:42:44] woooo, i think that worked. https://wikitech.wikimedia.org/w/index.php?title=Template:InstanceStatus&diff=prev&oldid=66495 [03:43:31] anomie: fix't [03:44:49] anomie: can i make it return the string unchanged if there's no match? [03:45:08] jeremyb_- That should be possible [04:07:01] Coren- I made a table of which packages are installed on one or two but not all of tools-login, tools-exec-01, and tools-exec-02. http://tools.wmflabs.org/anomiebot/packages.html [04:18:00] anomie: You do hard work. :-) The discrepancy between -login and the run environment is normal though. [04:18:14] I tried to create account at https://wikitech.wikimedia.org/w/index.php?title=Special:UserLogin&action=submitlogin&type=signup&returnto=Help%253AGetting+Started but got LOGIN ERROR: There was either an authentication database error or you are not allowed to update your external account. OKAY, a note at page bottom says, 'Note: The error message "There was either an authentication database error or you are not allowed to update y [04:18:14] our external account." generally indicates an invalid shell account name was used (see bug 16524).' SO, are 'roger' 'rogerhc' and 'rogervoy' all invalid shell account names? Where did I go wrong? Thanks! [04:38:46] roger_: Give me a second while I check. [04:40:32] Hi Coren, I have just now also posted this question at http://www.mediawiki.org/wiki/Talk:Wikimedia_Labs#Create_account_26228 but I will wait here to see what you can figure out... [04:42:22] I do not have shell access nor obviously any shell account name. Is that a prerequisite? [04:43:42] No, it works the other way around; you pick the shell account name when you create your wikitech account. [04:45:03] Okay, that's what I was trying to do. I'll wait here to see what you can figure out. [04:52:17] As far as I can tell, there's nothing wrong with any of those account names; I'm afraid I'm not yet familiar enough with OpenStack to figure out what breaks in your case. I'll tell Ryan, who is the guru. Sorry about the trouble. [04:57:15] Coren: Thanks. If convenient, Ryan can email me at rogerchrisman@gmail.com or reply at http://www.mediawiki.org/wiki/Talk:Wikimedia_Labs#Create_account_26228 [04:57:42] He's most likely to contact you through your userpage. [05:03:04] Coren: there's the spam blacklist? or abuse filter or something? [05:04:44] https://wikitech.wikimedia.org/wiki/MediaWiki:Titleblacklist [05:05:12] roger_: which is your first choice of those names? [05:05:17] Yeah, not applicable. [05:05:58] jeremyb_: rogerhc is my first choice [05:06:15] * Coren needs sleep. [05:06:17] * Coren waves [05:06:21] bye :) [05:06:29] i do too though! [05:06:48] wtf, i was granted skipcaptcha i thought? [05:07:28] roger_: and wiki username? also rogerhc ? [05:07:41] yes [05:08:28] Coren|Sleep: could it be opendj is still broke? [05:08:44] Ryan_Lane: ^ [05:09:03] i see no creations in ~11 hours [05:11:29] roger_: good luck [05:11:31] * jeremyb_ sleeps [06:57:41] @seen wm-bot [06:57:41] liangent: I am right here [06:58:11] petan: why doesn't it work in #mwbot ? [07:02:07] @seenrx [Bb]awolff.* [07:02:07] liangent: Last time I saw bawolff they were quitting the network with reason: Quit: g2g at 4/10/2013 11:46:07 PM (2.07:15:59.4378260 ago) (multiple results were found: bawolff_away, bawolff_) [07:20:41] liangent try @seen-on [07:50:34] petan: ah and it seems working in privmsgs too [07:50:59] rx doesnt [07:51:04] just seen works [07:51:10] to prevent some regex ddos :D [08:08:51] petan: you already have some timeouts for regexes right? [08:09:31] yes [08:09:40] 1 second or that [08:09:58] they are being processed in separate thread anyway [08:09:58] but it still eats lot of cpu [08:11:16] anyway someone can invite wm-bot to their own channel, then perform dos there, if it's doable in privmsgs [08:27:38] @notifyrx [08:43:53] liangent ok but that someone would need to have a wikimedia cloak [09:42:46] Hi, everyone [09:43:13] I'm moving my bot to wikimedia labs. [09:44:58] From the documentation, I should place the code at /data/project/"mybotname", but when I mkdir that directory, it shows "Permission denied". What should I do? [09:47:22] hey! [09:47:43] I think it automatically makes the home directories on the hour, so you might need to wait a bit longer [09:48:24] Okay. Thanks for your reply :) [10:24:08] Nullzero no it doesn t [10:24:15] Nullzero make it in userdata folder [10:24:18] !botsdocs [10:24:18] https://wikitech.wikimedia.org/wiki/Nova_Resource:Bots/Documentation [10:24:55] updated [10:25:21] oh? [10:25:35] you need to use /data/project/userdata [10:26:09] that´s a+rwx [10:35:55] Then, what about creatting my folder in public_html [10:45:06] Webserver is availible at /data/project/public_html/username, the path can't change to anywhere else. It's essential to have my folder there. [12:00:35] Nullzero yes I believe these folders are created automatically if not... I can create one for you [12:04:41] Well, then, please do so. Thank you :) [12:12:38] Nullzero done [12:13:47] And for bot, it should be in userdata, right? [12:17:27] your bot can be either in your home or in userdata [12:17:34] that is up to you... [12:17:59] if you want to make it public and accessbile to others then you should use userdata [12:18:32] you can also use tools project as bots project will become a staging area for tools in future [12:18:46] basically, it is a good idea to have accounts on both projects [12:19:16] Got it now. Thanks a lot ;) [12:31:07] [bz] (NEW - created by: Chris McMahon, priority: High - normal) [Bug 46166] Add automated browser tests to beta labs (tracking) - https://bugzilla.wikimedia.org/show_bug.cgi?id=46166 [12:36:14] [bz] (NEW - created by: Željko Filipin, priority: Unprioritized - normal) [Bug 47194] GettingStarted extension not at en.wikipedia.beta.wmflabs.org - https://bugzilla.wikimedia.org/show_bug.cgi?id=47194 [17:22:24] [bz] (NEW - created by: Željko Filipin, priority: Unprioritized - normal) [Bug 47203] Math extension broken at en.wikipedia.beta.wmflabs.org - https://bugzilla.wikimedia.org/show_bug.cgi?id=47203 [17:32:43] [bz] (NEW - created by: Željko Filipin, priority: Unprioritized - normal) [Bug 47205] Sandbox not at en.wikipedia.beta.wmflabs.org - https://bugzilla.wikimedia.org/show_bug.cgi?id=47205 [17:32:44] [bz] (NEW - created by: Željko Filipin, priority: Unprioritized - normal) [Bug 47203] Math extension broken at beta.wmflabs.org - https://bugzilla.wikimedia.org/show_bug.cgi?id=47203 [17:53:25] hmm [17:53:28] I can't create an account [17:53:31] on labs / wikitech [17:53:32] help? [17:56:50] anyone? [17:56:53] I'm getting 'There was either an authentication database error or you are not allowed to update your external account.' [17:57:28] for shell name suchabot [18:07:02] Coren: petan ^ ? [19:00:06] YuviPanda: There's a bug atm in openstack that will probably need Ryan to fix. But a better question is why such an account name? Don't you already have an account? [19:00:23] Coren: mediawiki.org/wiki/User:Yuvipanda/G2G [19:00:30] Coren: I'm working on a GitHub Sync bot [19:00:40] I already have github.com/SuchABot [19:01:02] You /really/ don't want a new wikitech account for this! [19:01:30] okay [19:01:32] what do I want? [19:01:37] That's what tool accounts are for. :-) [19:01:49] hmm, that just whooshed me by :) [19:01:51] explain? [19:02:00] I have not been reading your emails to wikitech-l :( [19:02:05] should I read them now? [19:02:05] !toolsdocs [19:02:05] http://www.mediawiki.org/wiki/Wikimedia_Labs/Tool_Labs/Help [19:02:17] Coren: it needs to push to Gerrit, btw. [19:02:21] * YuviPanda reads [19:02:27] Check the "tool account" section. [19:02:36] reading [19:03:27] Coren: sure, but it is a Gerrit bot and needs a Gerrit account [19:03:33] Coren: that page tells me nothing of Gerrit [19:03:49] this is not a wiki bot... [19:04:01] Gerrit credentials, otoh, are a different matter. Your use case is attpical enough (there was no provision for bots pushing to gerrit) that it'll need special magic. [19:04:26] It's the same idea, you have a tool that needs to be maintainable by more than one person. [19:05:09] Coren: sure, I'll move it to labs when it's doing things. But even then it'll need a Gerrit account (+ private keys) [19:05:12] A shared shell account is a big no-no for a number of reasons. That said, we can do it too. :-) [19:05:49] Coren: I'm writing / testing it now, and am forced to test it via my Gerrit account [19:05:53] resulting in things like https://gerrit.wikimedia.org/r/#/c/59058/ [19:05:58] nogood! [19:06:31] Coren: and why was I getting an error for that shell name? [19:06:34] bot is blacklisted? [19:06:51] Yes, the gerrit account will have to be dealt with, but Monday. We need to figure out the right way to do it without having a big hole like a shared wikitech account. [19:07:18] so for now I'll have to do it on mine? [19:07:24] Not as far as I know, there's a problem with openstack right now that fails every name. Unrelated bug. [19:07:29] oh [19:07:36] so nobody can create a gerrit account now? [19:07:37] o_O [19:07:38] O_o [19:08:30] Yeah, major boo boo, but beyond my current meager openstack extension knowledge to fix. [19:08:35] :'( [19:08:36] ok [19:08:40] * Coren is waiting for Ryan. [19:08:43] how long has this been the case, btw? [19:09:05] Last night (my time) is the first I heard of it. [19:09:24] ah [19:09:25] ok [19:09:33] so, for now I"ll just use my Gerrit account itself [19:09:38] and figure this one out later [19:13:12] And, to make matters more fun, I've been banging my head against a problem that, it turns out, cannot be solved. *grumble* [19:16:00] Coren: a problem that cannot be solved by Coren the Amazing? [19:16:02] *GASP* [19:16:13] are you trying to be a travelling salesperson? [19:16:21] Hardware issue. :-) [19:16:25] ah [19:34:40] andrewbogott_afk: I think the uid/gid generator is broken in OSM :( [19:38:08] Ryan_Lane: 'morning. [19:38:16] howdy [19:38:42] Ryan_Lane: productive :) [19:38:57] Ryan_Lane: Except for Gerrit account creation being broken :) [19:39:05] yeah. looking into this [19:39:08] has to do with a code change [19:39:08] Ryan_Lane: wheee :) [19:39:38] I've a somewhat working GitHub -> Gerrit thing that also handles people pushing new commits to the Pull Request [19:39:44] cool [19:39:51] be back in a few. picking up food [19:39:54] :D [20:00:08] ah ha [20:00:10] search limit [20:00:14] it wasn't applied on virt0 [20:00:26] fixed [20:00:42] A semirelated problem of YuviPanda. [20:01:20] He needs gerrit credentials for his tool. I felt that a new wikitech account was... Nit the right way to go about it. [20:01:44] that's the only way to do it [20:01:45] (unmaintainable without shared creds when maintainers change, etc) [20:02:00] Is it? Eew! [20:02:03] how else? [20:02:09] gerrit uses ldap auth [20:02:28] and we have a shared system [20:02:34] how is it unmaintainable? [20:02:58] and yeah, it needs shared creds [20:03:03] Well, it means you're stuck sharing password for one, and won't really work with two factor. [20:03:04] so do bot accounts on wikipedia ;) [20:03:16] we don't expect everyone to use two factor [20:03:21] only people who want to and admins [20:03:30] admins are required [20:03:38] Hm. Still uglyhackyuk territory. [20:04:39] But unless you have a brilliant idea for gerrit-only accounts, we be stuck with it. [20:05:10] And yeah, bot accounts on wikipedia are also teh suck. [20:05:24] For much the same reasons, really. [20:07:07] this is a common authentication problem [20:07:14] and there's not a really great solution [20:07:48] oauth usually helps, but you also don't want things tied to a real user either [20:08:28] * YuviPanda looks [20:08:55] Ryan_Lane: Coren so, I should just register the Gerrit account? [20:09:00] once it is back, that is [20:09:20] Yeah, it's ugly but it'll work. [20:09:46] * Ryan_Lane doesn't think it's ugly [20:10:00] But please write the tool with a real tool account and from your own rather than actually use that one for anything but the gerrit creds. [20:10:07] it would be nice if the auth had roles and we could limit its access [20:10:12] whee created! [20:10:36] Coren: sure, I don't expect to login to that at all [20:10:44] if we could disable it we should [20:11:02] you need to occasionally [20:11:11] to change ssh keys or change the email address [20:11:12] etc [20:12:03] yeah, monthly ssh keys cycling or something [20:12:27] no need for that :D [20:12:31] only if the key is compromised [20:12:58] Coren: this tool requires a webserver, so I think I'll get it's own instance / project? [20:13:04] whee! https://gerrit.wikimedia.org/r/#/c/59058/ :) [20:13:15] hmm, ANSI color codes [20:13:16] * YuviPanda fixes [20:13:30] Ryan_Lane: should this go into a new project? or is there a project for Gerrit stuff already? [20:13:43] eh? [20:13:54] oh you mean in gerrit [20:13:54] no ide [20:13:54] *idea [20:13:54] oh no [20:13:55] not gerrit [20:13:57] I mean in Labs [20:14:18] it can probably exist in tool labs [20:14:23] this requires a web server so that GitHub can notify me whenever a new pull request is there. [20:14:29] * YuviPanda reads Tool labs things again [20:16:52] Coren: what is 'A web URI mapped to its ~/public_html: http://tools.wmflabs.org/toolname ' running? [20:16:54] apache? [20:17:03] oh [20:17:04] read [20:17:05] nvm [20:17:07] * YuviPanda reas [20:18:52] Coren: this needs a python webserver. Will I *have* to use CGI? [20:28:09] Require a python webserver? You mean you have a python app that listens to port 80? [20:28:48] Coren: WSGI? [20:31:38] Coren: a python webapp, that is [20:33:05] Huh. Interesting. Well, I haven't yet looked into that use case since you're the first, but there's no reason we couldn't support WSGI. Not sure if I can use mod_wsgi with the current setup, but worst case is the CGI wrapper middleware. Either way, doesn't change your application. [20:33:40] Coren: hmm, mod_wsgi would be nice, I guess. But I don't know how easy it is to setup in a way that works the same way as PHP [20:33:46] so you can just pop a file in... [20:34:14] Coren: so, i can haz tools lab account? [20:34:37] Coren: shell is yuvipanda [20:34:39] Yep. What name do you want for it? [20:34:42] wiki Yuvipanda [20:34:47] Coren: suchabot? [20:35:02] Coren: or would that be too confusing? [20:35:12] since there's also going to be a ldap user called suchabot? [20:35:18] Isn't that the name you gave the gerrit account? i'm thinking that'd be confusing [20:35:28] yeah [20:35:32] suchabot-serveR? [20:35:34] err [20:35:36] something like that? [20:35:47] Suchaserver? [20:35:48] sure [20:36:11] Gimme a min. [20:36:15] whee [20:39:22] YuviPanda: What's your wikilabs user? [20:39:33] Coren: Yuvipanda [20:40:24] YuviPanda: All done. [20:41:32] whee Coren :) [20:42:40] Coren [20:42:51] Yes? [20:44:30] When execute the bot, in error.log appears outo of memory [20:44:42] *out [20:46:20] The default allowance is fairly conservative. You might want to try it with -mem 1g at first, and use qstat to see what its actual requirement is. [20:47:18] What language is it written in? [20:47:35] perl [20:47:52] It works [20:48:17] Perl is relatively frugal, so unless you're working with huge datasets, you shouldn't be too far off. [20:48:38] Use qstat -j job_number [20:48:48]