IRC logs for #baserock for Tuesday, 2015-03-17

*** zoli__ has joined #baserock00:31
*** zoli__ has quit IRC00:36
*** zoli__ has joined #baserock04:27
*** rdale has joined #baserock07:20
*** rdale has quit IRC07:30
*** rdale has joined #baserock07:32
straycat04:18 [freenode] Warning: Port sent with DCC request is a lowport (0, unknown) - this isn't normal. It is possible the        |07:55
straycat          address/port is faked (or maybe someone is just trying to bypass firewall)                                          |07:55
straycat04:18 [freenode] DCC SEND from Guest22117 [ port 0]: STARTKEYLOGGER [0B bytes] requested in channel #nixos             |07:55
straycat04:23 [freenode] DCC aborted receiving file STARTKEYLOGGER from Guest2211707:55
*** a1exhughe5 has joined #baserock08:00
straycat seems useful08:03
*** gfinney has joined #baserock08:26
*** bashrc_ has joined #baserock09:08
*** mike has joined #baserock09:08
*** mike is now known as Guest1918709:09
*** jonathanmaw has joined #baserock09:12
*** mdunford has joined #baserock09:17
*** edcragg has joined #baserock09:25
*** gary_perkins has joined #baserock09:38
*** Guest19187 has quit IRC09:39
*** ssam2 has joined #baserock09:40
*** ChanServ sets mode: +v ssam209:40
*** gfinney has quit IRC09:45
*** gfinney has joined #baserock09:45
richard_mawport 0 is not usable, that's how you request an ephemeral port assignment09:49
KinnisonNot on DCC09:49
KinnisonHowever 'startkeylogger' -- pity you didn't receive that file, it could have been interesting to examine09:50
richard_mawhow does that work then? does the protocol report which port was bound back?09:50
richard_mawor does port mean something different09:51
KinnisonI think it was either a broken attacker, or an attempt to activate something which should already have  been present if the client was a trojan09:51
rjekIt'll just be a command to botnet member that is controlled over IRC09:53
rjekAnd it's broadcasting it to a channel09:53
*** XuLiu has joined #baserock09:54
*** Guest19187 has joined #baserock09:55
*** lachlanmackenzie has joined #baserock10:13
*** wschaller has joined #baserock10:19
edcraggfor work on getting baserock onto HP moonshot, we have discovered a recent change which breaks the native build stage when bootstrapping to all platforms we've tested (armv8l64 and x86_64). I'm bisecting definitions, and the change currently appears to be after 25th Feb 2015. The error involves autoconf when building bison, as follows:
richard_mawlib/ error: lib_libbison_a_SOURCES must be set with '=' before using '+='10:25
* richard_maw thinks that's a silly restriction10:25
KinnisonWell, it kinda implies a possibility of brokenness10:26
ssam2edcragg: it looks like an update to GNU Make or GNU Automake could have caused that10:27
edcraggssam2: thanks, i'll have a look10:28
ratmice__would be weird if it was make as it was documented to work10:30
rjekIs being slow to clone from for anybody else?10:32
ratmice__see "When the variable in question"... has several lorry-controller processes all eating CPU :(10:34
rjekssam2: That'd be why then :)10:34
ssam2and one Git process also eating lots of CPU and RAM10:34
pedroalvarezotoh it has a lot of cpus and ram10:34
ssam2since lorry-controller basically just runs 'lorry' and records the result, it shouldn't be using 3 CPUs constantly though10:35
Kinnisonssam2: did you blow away that huge statedb?10:35
ssam2its current one is ~250MB10:36
Kinnisonis the statedb on btrfs or some other FS?10:36
KinnisonOkay, so it's not going to be a similar issue to the journal-on-btrfs slowness10:37
ssam2I wonder if the issue is that it has 1000s of .lorry files now10:38
ssam2or 1000s of entries at least (590 files)10:39
KinnisonI'd have hoped not10:41
petefothAre we ready to change the 'Contributing' page to describe submitting patches via Gerrit rather than by git-send-email, or do we want to wait for a while?10:43
paulsherwoodgo for it10:43
ssam2yeah, we're ready I think10:43
petefothOK I'll do a shameless copy of jjardon's 'mini how-to'10:44
pedroalvarezragarding g.b.o performance: it has running various "lorry-controller-webapp" processes. Shouldn't be only one?10:44
petefothAnd I'll add something in 'News'10:45
ssam2pedroalvarez: they are run by lighttpd, which I think spawns multiple10:45
ssam2since each can handle multiple requests10:45
ssam2sorry, each handles one request, but multiple requests can run at a time10:46
ssam2although SQLite is not known for great performance with concurrent writers, so maybe that's the problem10:46
pedroalvarezI see10:47
richard_mawssam2: it's mostly used to provide synchronisation AIUI, which you explicitly want it to contend concurrent writers10:48
ssam2sorry, I don't quite understand10:50
ssam2'it' being sqlite?10:51
richard_mawthe sqlite database that lorry-controller uses, aye10:51
ssam2I guess contention for the write lock wouldn't cause high CPU usage, just slow performance10:51
ssam2unless there's some bug in how lorry-controller waits for the write to complete10:52
Kinnisonstrace should be able to hint at what it's up to, unless it's managing to chew cpu and not make syscalls10:52
rjekThen you can use ltrace :)10:53
pedroalvarezno ltrace in g.b.o :/10:54
ssam2wow, ltrace10:54
rjekssam2: Ah, new discovery?  It as as useful as it is verbose.10:55
pedroalvarezISTM that the lorry-controller-webapp processes are eating a lot of cpu10:55
pedroalvarezstrace'ing shows a lot of FUTEX_(WAIT|WAKE)_ PRIVATE calls10:57
richard_mawlock contention then10:57
*** CTtpollard has quit IRC11:03
jjardonssam2: mmm, I cant see your patches to upgrade berkely-db from the bowser, but I can from the android app.11:07
ssam2weird, I can see them in the browser11:07
ssam2can you see the baserock/local-config/lorries project in the browser?11:08
jjardonssam2: nevermind, I was in "my" tab instead "all"11:08
* jjardon goes for more coffee11:08
radiofree might be useful11:13
*** CTtpollard has joined #baserock11:17
straycatare morph's conf extensions always run with the working directory set to the path of the root repo?11:25
richard_mawstraycat: I think so11:26
pedroalvarezHi pdar, I've just seen your patch for configobj in baserock-dev11:28
pedroalvarezI assume that you need a version higher than 5?11:28
straycatrichard_maw, yep, just double checked11:28
* bashrc_ encounters gcc breakage11:30
franredpedroalvarez, we were trying to update the chunk and the repo is not longer in bazaar - so this is why we created the lorry file11:31
franredtrue, sorry11:31
pedroalvarezfranred: from the new repo I take that it is a "Python 3+ compatible port of the configobj library"11:32
franredpedroalvarez, not sure why it has that description but in the mercurial repository it has "we have moved to" the github page11:33
straycatit would be cool to get gerrit to mail out patch series11:33
franredso I imagine it is the current development repository11:33
franredso I imagine it is the current development repository^^11:34
franredpedroalvarez, ^^11:34
pdarwhat fran said11:34
pedroalvarezoh, true11:34
pedroalvarezthen, I think it makes sense11:35
pedroalvarezdoes the upgrade fix things for you?11:35
pdarI think it works without the change11:36
franredpedroalvarez, no, it is just to get an update repo and no from 2 years ago11:37
franred4 years even11:37
pedroalvarezI hope that nothing that depends on it has a upper limit of the version required11:38
franredpedroalvarez, we will be careful updating the package, don't worry11:38
jjardonssam2: I got a "Error 500 (server error)" when trying to register a new email. I just filed the issue here (with the other gerrit ones) to keep track of them :!/story/2811:44
ssam2ok. might be because it can't send emails11:50
pedroalvarezyes, that is the reason I believe.11:58
petefothI'm trying to work out how gerrit fits in with the `morph branch` workflow as described in the `Developing your change' section fot the 'Contributing' page. My suspicion is that it may not :( at least not neatly12:30
jjardonssam2: while we still dont have a lot of stuff in gerrit, do you think it would be easy to update to gerrit 2.10.x or its going to be a lot of work?12:32
ssam2jjardon: pretty easy, I think. is there a specific feature you want?12:33
petefothWill `morph branch` clone the branched project from or from
ssam2from git.baserock.org12:33
ssam2best thing for now is probably recommend `morph branch ssh:// branchname`12:34
petefothso user would need to add as a remote, presumable named 'gerrit'?12:34
ssam2or that12:34
petefothssam2: ta!12:34
ssam2i think git-review will 'just work' if you have a remote named 'gerrit'12:34
*** fay_ has quit IRC12:35
petefothssam2: I got the idea from some git-review documentation :)12:35
jjardonssam2: avatars! :)
ssam2heh, ok12:35
*** fay_ has joined #baserock12:39
pedroalvarezoh, ansible-openstack-modules doesn't have anything like neutron_manage to populate the database :/13:00
petefothDid we find a work around for the 'Gerrit expects each 'Change' to be a single commit.' problem? Do you get round it by pushung a branch which contains several commits? OR...?13:06
SotKso, ./check fails with master of morph because of a bunch of long lines :(13:13
* SotK looks forward to pre-merge ci13:14
* straycat guesses that those sorts of checks can be used to block merges with gerrit13:14
SotKhmm, should I be able to log into with a Launchpad OpenID?13:15
* petefoth checks it wasn't he who did that 13:20
petefothSotK: No you have to get an ID from http://openid.baserock.org13:20
SotKI only ask because I just did13:21
SotKI might have logged in a while ago before the restriction was in place or something I guess13:21
SotKs/logged in/signed up/13:21
* SotK links an ID from
KinnisonDo you mean delegation?13:31
SotKI clicked "Link another identity" in the identities section of my Settings on Gerrit, and logged in with my Baserock OpenID13:32
KinnisonSo multiple openids for the same gerrit account?13:32
* petefoth adds that useful information to his draft 'Contributing via gerrit' wiki edit :)13:34
*** gfinney has quit IRC13:59
*** gfinney has joined #baserock14:00
*** mdunford has quit IRC14:07
SotKIt seems I can't push to Gerrit :(14:08
* richard_maw kicks /var/run14:12
petefothSotK: what's the problem?14:12
SotKpetefoth: ^14:13
ssam2SotK: if you signed up a while back you might not be in the right group somehow..14:14
jjardonSotK: did you upload your public SSH key?14:14
ssam2although that error looks weirder than that14:14
petefothand whats is the url for origin14:14
SotK is my git config for that repo14:15
SotKjjardon: yep14:15
SotKI obtained the repo using the SSH URL14:15
SotKmy Groups are apparently "Anonymous Users" and "Registered Users"14:16
* petefoth backs away, unable to help14:17
ssam2no suspicious errors in the logs14:17
SotKmaybe I'm doing something wrong on this end14:17
ssam2SotK: I've manually added you to the Developers group, could you try to push again?14:18
SotKsame error :(14:18
ssam2right. I wonder if it's to do with you not using a openid...14:19
ssam2or not using one as your main one14:19
ssam2maybe if you remove your launchpad ID from the list of IDs and try again?14:19
SotKI'll try that14:19
SotKyep, worked this time14:19
ssam2ah, cool14:19
SotKpetefoth: probably remove that not-actually-useful information from your wiki draft :)14:20
ssam2thanks for ! bad jjardon for not running ./check before merging!14:21
*** mdunford has joined #baserock14:21
petefothSotK: because it doesn't work and causes pushes to fail? :)14:21
SotKyep :D14:21
ssam2SotK: i gave it a +214:21
petefothSotK: I'll mention that you *shouldn't* do that otherwise pushes will fail14:22
SotKpetefoth: good plan14:22
jjardonSotK: sorry and thanks for the patch14:23
SotKjjardon: no problem :)14:23
petefothssam2: would it be possible / easy / straighforward to inhibit the 'Link another identity' button (or the whole page) it if is going to break stuff?14:24
ssam2I don't know14:25
ssam2I thought that's what setting the auth.openid setting would do, but seems that it doesn't14:25
ssam2it's on my list but not very high up14:25
* petefoth adds something in storyboard 14:25
*** wschaller has quit IRC14:54
*** zoli__ has quit IRC15:06
*** zoli__ has joined #baserock15:07
SotKIs there a way to checkout a change which is on Gerrit locally?15:07
ssam2there are refs like refs/changes/01/0115:10
ssam2where 01 and 01 are the change number (not sure why it's there twice)15:10
ssam2because they're not in refs/heads/ or refs/tags you have to fetch them manually like `git fetch gerrit refs/changes/01/01`, then check out FETCH_HEAD15:10
ssam2or git-review can do it for you15:11
*** zoli__ has quit IRC15:11
* SotK looks at git review15:11
*** wschaller has joined #baserock15:12
*** edcragg has quit IRC15:20
*** flatmush has quit IRC15:20
*** sambishop has quit IRC15:20
*** paulw has quit IRC15:20
*** Guest19187 has quit IRC15:20
*** mdunford has quit IRC15:20
*** bashrc_ has quit IRC15:20
*** fay_ has quit IRC15:20
*** petefoth has quit IRC15:20
*** CTtpollard has quit IRC15:20
*** wschaller has quit IRC15:20
*** a1exhughe5 has quit IRC15:20
*** jonathanmaw has quit IRC15:20
*** wschaller_ has joined #baserock15:20
*** fay_ has joined #baserock15:20
*** ssam2 has quit IRC15:20
*** bashrc_ has joined #baserock15:20
*** petefoth has joined #baserock15:20
*** jonathanmaw has joined #baserock15:20
*** franred_ has joined #baserock15:20
*** a1exhughe5 has joined #baserock15:20
*** franred_ has quit IRC15:20
*** ssam2 has joined #baserock15:20
*** ChanServ sets mode: +v ssam215:20
*** paulw has joined #baserock15:20
*** mdunford has joined #baserock15:20
*** sambishop has joined #baserock15:20
*** edcragg has joined #baserock15:20
*** mauricemoss_ has quit IRC15:20
*** franred_ has joined #baserock15:20
*** gary_perkins has quit IRC15:20
*** franred has quit IRC15:21
*** Guest19187 has joined #baserock15:21
*** lachlanmackenzie has quit IRC15:21
*** mauricemoss_ has joined #baserock15:21
*** CTtpollard has joined #baserock15:21
*** lachlanmackenzie has joined #baserock15:21
*** flatmush has joined #baserock15:21
*** gary_perkins has joined #baserock15:22
*** CTtpollard has quit IRC15:28
*** CTtpollard has joined #baserock15:39
jmacsAm I right in saying we can't build for ARMv6 at the moment?15:41
SotKAs far as I am aware, yes15:42
jmacsWould that be a question of finding the right version of GCC, or is there something more fundamental?15:42
Kinnisonit's a whole-new-bootstrap15:43
SotKI *think* pedroalvarez was looking at ARMv6 once upon a time, but I may be mistaken.15:44
gfinneyWhen I run the morph build (on commit 5e09ce28da6bddfc75661837058743447474f185 plus a small change)15:44
gfinney'morph build systems/base-system-x86_64-generic.morph'15:44
gfinneyI get the error:15:44
gfinney"ERROR: Stratum devtools has no build dependencies for chunk vim in strata/devtools.morph"15:44
gfinneyAny idea what's happening here?15:44
ssam2too old Morph, use master15:44
jmacsActually I'm looking at ARMv5, I was just asking as ARMv6 had been discussed before15:45
pedroalvarezjmacs: yeah, I have "something", and I reworked it a couple of weeks ago, but my lack of resources and time for that made give up15:45
pedroalvarezs/made/made me/15:45
gfinneyssam2, was that an answer to my question?15:46
gfinneyor somone else?15:46
rjekgfinney: To you15:47
pedroalvarezgfinney: yes, use latest morph for that15:47
gfinneythought I had15:47
gfinneyOK thank you guys15:50
pedroalvarezjmacs: you can take a look at my armv6 attempt:
paulsherwoodbaserock on rpi? is it working?15:52
pedroalvarezpaulsherwood: has been my weekend project for two weekends (one after several months after the first one). And I failed15:58
pedroalvarezI managed to get a bootstrap system for armv615:59
pedroalvarezbut I couldn't run morph in that bootstrap system because the raspbian kernel doesn't have NAMESPACES support...16:00
pedroalvarezI tried to compile the kernel with support for that, but then it didn't recognise my HDD, where I had my baserock bootstrap rootfs16:01
pedroalvarezand then I wished I never started doing it :)16:01
jmacsJust taking a look now; it should give me some idea of what's involved16:07
* straycat may have just hurt himself with local-changes=ignore16:33
straycati've not been building what i thought i was building16:33
straycatbut deploys still happend based on what's in the local checkout16:34
*** zoli__ has joined #baserock16:34
straycatmaybe a local-changes=root-repo-ref would be prudent while we transition away from system branches >.>16:35
* straycat doesn't really get why morph deploy was applying what i had locally then16:43
straycats/ i / he / >.>16:43
straycatssam2, does the local-changes option only apply to the build command or something?16:44
ssam2should apply to 'deploy' too16:46
ssam2check the code though, deploy is a separate code path16:46
* straycat doesn't have time right now16:47
straycatwill check later16:47
SotKlooks like it applies to me16:47
straycatif it should apply then i have a feeling there's a bug16:48
straycatcan't wait till we get rid of system branches and workspaces16:48
*** a1exhughe5 has quit IRC16:51
*** wschaller___ has joined #baserock16:53
*** mdunford has quit IRC16:54
* SotK notices mason is red16:54
SotKERROR: Deployment failed as system is not yet built.16:56
*** wschaller_ has quit IRC16:57
pedroalvarezSotK: hm.. it has tried to deploy the system before building it16:59
pedroalvarezthe fact that there aren't more reds is a good signal :)17:00
*** franred_ has quit IRC17:01
* SotK wonders what took it so long to complete if it wasn't building *something*17:02
kejiahuhi guys, I am trying to cross build armv8 baserock, and I have met a cmake error while running the native-bootstrap script. It only happens when I netboot into the bootstrap rootfs and run the script, and doesn't happen if I chroot into the same rootfs from a running linux. Does anybody have any idea what may be the difference between the netboot way and chroot way? the error is thanks17:04
ssam2sorry for slight Gerrit downtime, I'm making config changes, and it's very slow to restart17:06
CTtpollardwhat db are you using with Gerrit ssam2?17:08
*** mdunford has joined #baserock17:09
*** XuLiu has quit IRC17:13
*** zoli__ has quit IRC17:17
* straycat spots 17:24
straycatssam2, cool :)17:25
CTtpollardah nice!17:25
ssam2mwilliams_ct: how are you getting on with the openid provider?17:34
*** RedHotFireBall has joined #baserock17:38
mwilliams_ctssam2: slightly on hold, hoping to get back to it tomorrow but not certain yet17:38
*** RedHotFireBall has quit IRC17:48
*** RedHotFireBall has joined #baserock17:52
ssam2I ask because I was just looking at openid myself. The 'sign in' link on now goes straight to openid.baserock.org17:55
mwilliams_cton which it works fine17:56
mwilliams_ctit could be something unique to the hosting setup rather than to the code17:56
*** bashrc_ has quit IRC17:59
*** Guest19187 has quit IRC18:01
*** RedHotFireBall has quit IRC18:02
ssam2mwilliams_ct: the 'unable to check identity' thing may be a request that storyboard isn't honouring18:03
*** jonathanmaw has quit IRC18:06
ssam2argh! gerrit seems to have created two accounts for me18:06
ssam2oh, this old chestnut: one for my openid with a trailing /, one for my openid without a trailing /18:07
ssam2i tried to fix the openid provider to strip the trailing /, but the python 'openid' module threw an exception if I did that18:07
ssam2this is really stupid18:08
ssam2I guess this won't happen now that users don't need to manually enter their OpenIDs18:09
pdarpardon me18:09
*** mdunford has quit IRC18:11
*** lachlanmackenzie has quit IRC18:12
*** rdale has quit IRC18:26
*** edcragg has quit IRC18:37
*** wschaller___ has quit IRC18:49
*** ssam2 has quit IRC18:54
SotKssam2: The Gerrit sign in change is excellent19:13
SotKoh, he's not here :)19:13
*** gary_perkins has quit IRC19:14
*** zoli__ has joined #baserock19:19
*** gfinney has quit IRC19:22
*** gfinney has joined #baserock19:32
*** pacon2 has joined #baserock19:51
*** zoli__ has quit IRC20:26
*** radiofree has quit IRC20:39
*** radiofree has joined #baserock20:39
*** zoli__ has joined #baserock20:41
*** radiofree has quit IRC20:41
*** radiofree has joined #baserock20:41
* pedroalvarez creates some space in so mason stops failing21:58
* pedroalvarez ponders incrementing size of its disk21:59
*** zoli__ has quit IRC22:23
*** edcragg has joined #baserock22:36

Generated by 2.15.3 by Marius Gedminas - find it at!