IRC logs for #baserock for Friday, 2014-10-31

*** rdale_ [] has joined #baserock00:14
*** flatmush [] has quit [Read error: Connection reset by peer]00:14
*** rdale [] has quit [Read error: Connection reset by peer]00:14
*** violeta_ [] has quit [Read error: Connection reset by peer]00:14
*** violeta_ [] has joined #baserock00:14
*** flatmush [] has joined #baserock00:14
*** rdale_ [] has quit [Ping timeout: 260 seconds]00:20
*** violeta_ [] has quit [Ping timeout: 258 seconds]00:21
*** flatmush [] has quit [Ping timeout: 272 seconds]00:21
*** rdale [] has joined #baserock00:30
*** violeta_ [] has joined #baserock00:32
*** flatmush [] has joined #baserock00:32
*** genii [~quassel@ubuntu/member/genii] has quit [Write error: Connection reset by peer]01:43
*** violeta_ [] has quit [Read error: Connection reset by peer]01:48
*** rdale_ [] has joined #baserock01:48
*** rdale [] has quit [Read error: Connection reset by peer]01:48
*** violeta_ [] has joined #baserock01:48
*** flatmush [] has quit [Read error: Connection reset by peer]01:49
*** flatmush [] has joined #baserock01:50
*** thecorconian [4b1b2c1f@gateway/web/freenode/ip.] has joined #baserock03:40
*** VVIBIN [~VVIBIN@] has joined #baserock04:20
*** aananth [~caananth@] has joined #baserock05:24
*** aananth [~caananth@] has quit [Remote host closed the connection]07:07
*** aananth [~caananth@] has joined #baserock07:18
*** rdale_ [] has quit [Read error: Connection reset by peer]07:35
*** violeta_ [] has quit [Read error: Connection reset by peer]07:35
*** flatmush [] has quit [Read error: Connection reset by peer]07:35
*** rdale [] has joined #baserock07:35
*** violeta_ [] has joined #baserock07:35
*** flatmush [] has joined #baserock07:36
*** ssam2 [] has joined #baserock08:06
Mode #baserock +v ssam2 by ChanServ08:06
aananthssam2: I sent you an email regarding my trove deployment. Kindly reply.08:16
ssam2Aananth: I shall do now08:16
ssam2aananth: I can reply here, in fact08:18
ssam2if your trove is running correctly then visiting it in a web browser, you'll see a list of the repos you have indexed08:19
paulsherwoodaananth: did you manage to create your own trove? :)08:19
ssam2at the bottom of that page is a link that says 'view lorry controller status' which should show info on where it is mirroring from and the progress of the mirroring08:19
ssam2if there seem to be problems, the first place to check for errors is to run: `journalctl --full`08:20
aananthOk sam. NTPD is failing...08:22
ssam2That would suggest the Trove can't connect the network08:23
ssam2I guess probably a proxy issue08:23
aananthYeah. Madhu sent me some link regarding this yesterday, let me fix that.08:23
ssam2actually, NTP goes on port 123 by default08:24
ssam2so it could be being blocked by your firewall08:24
ssam2that said, as long as you're sure the time is set correctly you can disable ntpd.service and ntpd-boot.service08:25
ssam2Baserock has `date` for checking and setting time08:25
*** petefoth [] has joined #baserock08:28
*** petefoth [] has quit [Remote host closed the connection]08:57
*** petefoth [] has joined #baserock08:57
*** mariaderidder [] has joined #baserock08:58
* pedroalvarez wonders if anybody has ever looked into upgrading busybox08:59
pedroalvarezs/into/at/ :P09:00
jjardonI think we are using the latest already09:00
pedroalvarezjjardon: not sure, glibc requieres a patch that is IIRC in the latest version09:00
jjardonIIRC Kinnison updated it09:01
jjardonOk, i was wrong: we are using 1.21 but 1.22.1 is available09:06
pedroalvarezyeah, I was revisiting the patches needed to move to glibc, and I wondered if it was better to upgrade busybox rather than patching  it09:09
pedroalvarezbut busybox chunk has a lot of changes from us09:10
pedroalvarezalthough I can always merge the latest tag09:10
*** dutch [] has joined #baserock09:15
pedroalvarezwhen moving to glibc I found a problem with lsof. It was complaining because it couldn't find "rpc/rpc.h". Upgrading it to the latest version solved the issue but I wonder if I'm missing something important09:17
pedroalvarezAlso, in they say that the latest version depends on libtirpc-0.2.5, and I managed to get lsof building and running without doing anything else than upgrading to the latest version.09:19
pedroalvarezbtw, to upgrade to the latest version I need to do this first:
*** tiagogomes [] has joined #baserock09:23
jjardonYeah, rpc are not installed by default in modern versions Og glibc anymore; you can use --enable-obsolete-rpc if you needed09:26
jjardonAbout busybox; don't know the reasons of the patches buy my opinion is that the less amount of patches we carry, the better09:27
*** jonathanmaw [] has joined #baserock09:29
*** ssam2 [] has quit [Ping timeout: 264 seconds]09:32
pedroalvarezjjardon: wrt rpc - aha, that makes sense. I wonder why is working without installing the ti-rpc in the system09:32
*** ssam2 [] has joined #baserock09:50
Mode #baserock +v ssam2 by ChanServ09:50
*** franred [] has joined #baserock09:58
*** aananth [~caananth@] has quit [Ping timeout: 258 seconds]10:02
*** dutch [] has quit [Ping timeout: 264 seconds]10:07
*** dutch [] has joined #baserock10:13
persiare: busybox: have those patches been sent upstream?  Were any merged?10:13
pedroalvarezpersia: as far as I could see, they were systemd units10:15
persiaHmm.  Those probably don't belong upstream for either busybox or systemd.  I wonder where they should belong.10:16
*** dutch [] has quit [Client Quit]10:16
*** dutch [] has joined #baserock10:17
*** rdale_ [] has joined #baserock10:20
*** rdale [] has quit [Read error: Connection reset by peer]10:20
*** aananth [~caananth@] has joined #baserock10:20
*** dutch [] has quit [Ping timeout: 255 seconds]10:24
*** dutch [] has joined #baserock10:29
*** aananth [~caananth@] has quit [Read error: Connection reset by peer]10:38
*** aananth [~caananth@] has joined #baserock10:50
jjardonpedroalvarez: note that libtirpc is already in stratum/nfs10:50
pedroalvarezjjardon: I know, but that is not present in the devel system10:51
ssam2it could be, I think10:52
jjardonssam2: thanks for the coreutils review! seems tcpwrapper are not really needed anymore: so I will try to get rid of them10:52
ssam2cool, yeah I don't think we actually use it in Trove / distbuild for anything10:52
ssam2it was just a required build dep for some NFS component at the time10:52
pedroalvarezssam2: what do you mean exactly with "it could be"10:54
* pedroalvarez is confused10:54
ssam2you could add it there, if you need10:54
*** thecorconian [4b1b2c1f@gateway/web/freenode/ip.] has quit [Quit: Page closed]11:04
*** thecorconian1 [] has joined #baserock11:10
*** thecorconian1 [] has quit []11:10
*** petefotheringham [] has quit [Quit: leaving]11:11
*** petefotheringham [] has joined #baserock11:11
petefotheringham is now known as petefoth111:12
jjardonssam2: build system built fine with coreutils + patch in baserock/jjardon/tcpwrappers :)11:21
jjardonssam2: Do you want me to test any other system?11:21
ssam2should be OK11:22
ssam2I might have a go at deploying a build system and checking NFS still works11:22
*** tiagogomes [] has quit [Quit: Leaving]11:35
*** tiagogomes [] has joined #baserock11:35
pedroalvarezha! I just figured out that I *can't* put a init script in /root11:39
pedroalvarezbecause /root is a btrfs subvolume!!!11:40
pedroalvarez /me sighs11:40
* pedroalvarez too11:40
pedroalvarezjjardon: are you updating patch?11:49
pedroalvarezupgrading even11:49
jjardonpedroalvarez: in my coreutils branch, yes11:50
pedroalvarezjjardon: there was already a branch for that (baserock/v2.7.1). I was using that in my glibc branch11:51
jjardonpedroalvarez: yeah, Im using that one as well11:52
aananthssam2: I could see trove server from any browser on any machines with in our local network. But it says "no repositories found"11:57
aananthpaul: Sorry I missed your question. The answer is "yes". But I need to fix the ntpd issue and make my trove server to fetch git repos from baserock git.12:01
pedroalvarezaananth: at the bottom of the page there is a link which says "View Lorry Controller Status"12:03
pedroalvarezaananth: if you click in that link, you go to the lorry controller status page12:04
pedroalvarezthat page has to be similar to this one:
aananthOk pedroalvarez. Let me check and come back.12:06
aananthpedroalvarez: Please find the text pasted in
aananthCould you explain me what that means?12:09
ssam2Hi aananth12:09
ssam2I see that it has '' in the list of configured troves12:09
ssam2actually, I know what the problem is. Sorry, there's a step I failed to tell you about!12:10
ssam2your Trove's 'lorry' key needs an account to be created on 12:10
ssam2if you can look inside the Trove for /etc/trove/ and send it to me, I can create an account12:11
pedroalvarezssam2: are your sure? 12:11
Kinnisonssam2: can they not use http?12:11
aananthOh! I will send that key now.12:11
ssam2Kinnison: perhaps, how can we change the Trove to use http ?12:11
Kinnisonin the lorry controller config12:11
Kinnisonin the trove stanza12:11
pedroalvarezIs httpby default, no? 12:11
KinnisonI think it's "protocol": "http"12:11
Kinnisonpedroalvarez: I don't recall12:12
ssam2it should be, but clearly it isn't12:12
ssam2aananth: OK, so rather than sending the key you should be able to edit a config file12:12
pedroalvarezMaybe the ls is taking time 12:12
ssam2when I deployed the ct-training trove I had to add a key to the upstream trove12:12
ssam2I forgot that with HTTP you don't need to do that12:13
ssam2but clearly it's not the default12:13
ssam2aananth: if you do 'git clone ssh://git@TROVE_HOST/TROVE_ID/local-config/lorries.git' you'll get a local clone of the lorries configuration repo12:13
aananthssam2: I sent the public key to your email id.12:14
ssam2ok, thanks. I'll add the key so that you can get up and running12:14
aananthssam2: What to do after cloning lorries.git? 12:15
ssam2seems I was wrong about the default value for 'protocol'12:16
ssam2it is 'http' in the Trove I created on Tuesday12:16
ssam2aananth: if you cloned lorries.git successfully, can you paste the 'lorry-controller.conf' file somewhere so we can see it?12:17
ssam2that's the file that configures how the mirroring is done12:18
ssam2to clone the lorries.git repo you might need to add yourself to the correct group first so you can access it12:19
ssam2if you do 'ssh git@TROVE group adduser local-config-writers USERNAME' it'll give you read and write permissions for the lorries repo12:19
ssam2Trove has fine-grained access-control, as we said :)12:20
ssam2Kinnison, pedroalvarez: I've deleted the key I added for the ct-training-trove from to see if I can see mirror-over-HTTP working in action12:22
pedroalvarezssam2: the baserock-clone trove doesn't have the keys in git.baserock.org12:23
ssam2I guess I was impatient and lacking in knowledge12:23
pedroalvarezmaybe something has changed12:23
ssam2certainly, aananth is not seeing any repos yet :(12:23
pedroalvarezhow can we do a 'ls' in gitano using http?12:24
Kinnisonthere's a gitano-command.cgi which straycat wrote \o/12:25
aananthssam2: I think I found an issue which could be a reason.12:25
aananthI need to install corkscrew and configure git, ssh to work behind the proxy in trove server.12:25
aananthI have not done this step.12:25
aananthLet me try that.12:26
KinnisonOh dear12:27
Kinnisontroves can be configured with proxy access details12:27
Kinnisonbut I don't think they can corkscrew per-se12:27
pedroalvarezKinnison: thanks for the hint :)
ssam2aananth: it should be cloning over HTTP, so it shouldn't need corkscrew I think12:28
ssam2it should just need access on port 8012:28
ssam2we have set up a Trove behind a firewall that prevented internet traffic on all ports except 8012:28
pedroalvarezssam2: indeed, and if that was the issue, at least the trove has to manage to get the remote repositories before cloning12:30
pedroalvarezthat's why I was checking that 'ls' works in g.b.o12:30
aananthssam2, I just tried cloning my project in github, inside the trove server :), but I get this error "Couldn't resolve host ''". I used git clone
ssam2in my trove, I see a lot of this in the journal:12:33
ssam2aananth: that's a good idea as a way of testing the connection12:34
ssam2aananth: seems that something is wrong with the networking in the Trove12:34
ssam2does 'ping' work ?12:34
ssam2if so, it has network but not DNS, I think12:34
aananthIt works, here is the output: 64 bytes from seq=1 ttl=40 time=274.663 ms12:35
ssam2ok, so network works but not DNS12:35
ssam2what do you see in the file /etc/resolv.conf ?12:35
aananthI get the one in the link:
aananthLet me try adding to the resolv.conf and try.12:37
ssam2good idea12:37
pedroalvarezis not resolv.conf going to be regenerated?12:37
aananthOnly on next boot, right?12:38
ssam2yes. it's good as a test, to see if the problem is that none of the DNS servers currently listed are working12:38
*** VVIBIN [~VVIBIN@] has quit [Ping timeout: 264 seconds]12:39
aananthssam2: PING is with name seems to be disabled in our company. Even that did not work on my PC. 12:45
jjardonssam2: any idea why nfsv4 support got disabled?12:45
ssam2jjardon: I think because at the time it did't work as well as nfsv3, and disabling it caused nfsv3 to be the default12:45
aananthThen, I used wget, without any proxy setting (http_proxy) inside trove server. It did not work or got "bad address" error.12:45
ssam2jjardon: I don't if nfsv4 has improved, but I have an impression from a couple of years ago that it's quite bussy12:46
aananthThen, I configured http_proxy environment variable and did a wget, it worked.12:46
ssam2aananth: aha !12:46
aananthsounds to me like a proxy issue.12:46
ssam2jjardon: also, the policy for Baserock is that you need +2 before merging, not +112:46
ssam2bear that in mind in future at least :)12:47
ssam2aananth: yes12:47
ssam2aananth: did you manage to clone local-config/lorries.git from your trove?12:47
aananthNo, that again did not work12:48
ssam2was it a permissions error or a networking error?12:48
aananthssh: Could not resolve hostname TROVE_HOST: Name or service not known. 12:48
abdul_ssam2: getting error while building wandboard image in our target "ERROR: In staging area /src/tmp/staging/tmpNbVXt7: running command 'sh -c make' failed."12:49
ssam2aananth: replace TROVE_HOST with the actual IP or hostname of your trove12:49
aananthsorry :)12:49
ssam2abdul_: ok, does it give you the build output too? if so, paste that to and I'll see if I can see the problem12:49
aananthssam2: I cloned it into my PC, but got errors. See
ssam2OK, that means you don't have the right SSH key available on your PC12:58
ssam2aananth: you can run `ssh-add --l` to see what keys you have currently available on a machine12:58
ssam2sorry, `ssh-add -l`12:58
ssam2and to add a new one to your SSH agent so that it can be used, you need `ssh-add /path/to/key`12:59
*** dutch [] has quit [Quit: Quit]12:59
ssam2since you created the Trove, you created the 'trove-admin' key12:59
ssam2that's the one you need now to authenticate yourself13:00
ssam2abdul_: that problem is very special, I've never seen it before!13:00
ssam2abdul_: is this building on NFS by any chance ?13:00
abdul_ssam2: yes13:01
aananthOk I understand why I am not able to clone the git from my trove server. Is there any relation with the trove server is unable to fetch repos?13:01
ssam2aananth: no, I think that the Trove server can't fetch repos because you need to add the proxy configuration to 'lorry-controller.conf'13:01
ssam2you can find that file in local-config/lorries.git, once you can get at local-config/lorries.git :-)13:02
ssam2abdul_: we don't really test building on NFS, so there could be some problem in the Makefiles for XML-Parser that don't normally see13:02
aananthssam2: Thanks I understand.13:03
abdul_ssam2: okay. but in our hardware we dont have enough space to do baserock build. so opted for NFS13:04
ssam2abdul_: OK. if you want to try and investigate the failure, you can do 'chroot /src/tmp/failed/tmpNbVXt7'13:04
ssam2(i worked that path out from the last line of the log file you sent)13:05
ssam2then you'll be in the build environment Morph is using for that chunk13:05
ssam2there'll be a directory called something like / where the source for XML-Parser is13:05
ssam2you can try running 'make' manually etc. and see if you can debug it13:05
ssam2I hope to get a chance later to try building on NFS myself13:06
*** thecorconian [] has joined #baserock13:07
abdul_ssam2: okay i will try and let you know13:08
ssam2alternately, if you can find an external hard disk I think the problem will go away13:11
abdul_ssam2: make is working when i do it manually13:11
abdul_ssam2: how can go back and give morph build again13:12
pedroalvarezabdul_: type `exit` to exit from the chroot13:13
abdul_ssam2: thanks sam13:14
radiofreeit looks like the build for that returns make: *** [Makefile] Error 1 as part of the build13:14
radiofreeyou have to run make twice, morph is treating that as a build error?13:14
ssam2radiofree: perhaps, yeah13:16
ssam2radiofree: it may be a timestamps issue13:16
radiofree==> Your Makefile has been rebuilt. <==\n==> Please rerun the make command.  <==\nfalse\nmake: *** [Makefile] Error 113:16
radiofreehowever in the manual build just running make again after that works13:16
ssam2radiofree: makes sense13:17
ssam2radiofree: but we always build XML-Parser and we don't normally have this problem13:17
ssam2it might be a good enough workaround to change the build instructions to 'make || make' though :)13:17
aananthssam2: I have sent you an email asking clarification w.r.t key generation. Kindly reply.13:21
ssam2aananth: done13:29
ssam2aananth: this is also documented here:
abdul_again same problem in xml-parser13:30
ssam2although it's really lacking quite a lot of info13:30
ssam2abdul_: OK13:30
ssam2abdul_: you could try radiofree's suggestion13:30
ssam2currently XML-Parser is using the default build commands13:30
ssam2you could add a chunk morphology for it which overrides them13:31
ssam2so instead of running `make`, it runs `make || make`13:31
ssam2so that if `make` fails the first time, it'll run it again instead of raising an error13:31
ssam2this is a bit of a hack :) but it's worth a try13:31
ssam2do you feel confident to add a chunk morph or do you want me to guide you through it ?13:31
abdul_okay i will give a try..interesting !!!13:33
* paulsherwood has missed backscroll... is this a new/different version of XML-Parser?13:33
*** tiagogomes [] has quit [Ping timeout: 264 seconds]13:33
ssam2paulsherwood: no, abdul_ is using NFS for /src, so he can build it on a small embedded device without an external HD13:33
ssam2so it's probably some kind of timestamps / timing issue13:34
* pedroalvarez ponders enabling "--enable-obsolete-rpc" in glibc after realising that libtirpc doesn't build with glibc13:34
aananthssam2: thanks. In case of different trove-admin, shouldn't I regenerate admin's public and private key again?13:38
ssam2aananth: you mean, if you wanted to make someone else the admin of the Trove ?13:39
ssam2aananth: i'd suggest just giving them the public and private trove-admin keys13:39
ssam2it'd be possible to generate new keys and replace them, using 'ssh git@.. sshkey del ..'  and 'ssh git@... sshkey add ..'13:40
aananthNo, I used "trove-admin"  in ssh-keygen command instead of "caananth" which is my ID. Hence I need to regenerate the keys. Not required?13:40
aananthI got it.13:40
ssam2aananth: think of the trove-admin as a separate, special role13:41
aananthThanks ssam2. I understood your email better now. Sorry, its friday evening :) and my mind is already tired with full weeks work load :). Let me do that and let you know.13:43
ssam2aananth: trove is complex, I think you're doing really well to have one up and running already13:44
ssam2we'll still be here on Monday, you should make sure to enjoy your Friday evening !13:45
jjardonpedroalvarez: enable it , seems its needed to rebuild glibc13:47
*** tiagogomes [] has joined #baserock13:48
jjardonAlso Arch and fedora enable it as well so seems to be the correct thing to do13:51
pedroalvarezjjardon: hm... i shall give  up13:58
jjardonpedroalvarez: never! ;) what is the problem now?13:59
pedroalvarezjjardon: you told me to give up enabling "--enable-obsolete-rpc"14:00
pedroalvarezI was trying to figure out how others build ti-rpc14:00
pedroalvarezusing patches from fedora, looking at the patches of debian, etc14:00
jjardonpedroalvarez: oh yeah, I though you were giving up on building glibc14:01
paulsherwoodssam2: max-jobs?14:02
ssam2paulsherwood: for xml-parser? I don't think that's the problem, but it might help14:03
jjardon pedroalvarez :) Also you should finish soon so you can take a look to my coreutils branch: I need it for systemd 217 ;)14:03
pedroalvarezafter this I'm going to update gcc14:06
paulsherwoodthat will be a big rebuild :)14:08
pedroalvarezslightly bigger than the one caused by moving to glibc14:10
pedroalvarezand I think that these upgrades are important14:13
jjardonpedroalvarez: they are. Even more for arm14:36
pedroalvarezjjardon: so is your coreutils branch ready?14:39
jjardonpedroalvarez: it is14:40
jjardonI build and run a build and a genivi-baseline systems with it14:41
*** madhu [~madhu@] has quit [Quit: Leaving]14:41
jjardons/I build/I built/14:42
paulsherwoodvioleta_: did you or anyone manage to establish whether tegra can run in big endian mode?14:58
*** dutch [] has joined #baserock15:01
paulsherwoodridgerunner: ^^ ?15:01
*** dutch [] has quit [Client Quit]15:05
rjekpaulsherwood: The ARM part running in BE is part of the standard that they must implement.15:05
rjekpaulsherwood: My understanding is that the problem is if the drives can cope or if they need patching to be endian-agnostic.15:06
rjek(ie, if they need to byte-swap data written to peripherals)15:06
paulsherwoodright thanks15:06
* rjek is not an authoritative source, however.15:07
*** dutch [] has joined #baserock15:08
petefothdo we need to publicise the project policies that have been discussed on the ML? The one I recall is ‘who gets committer access?” Maybe there were others? Should they be on the wiki?15:21
ssam2Publicise meaning 'put them on the wiki' ?15:22
pedroalvarezI assume so15:22
ssam2I think that's good, if they can be summarised sufficiently15:22
ssam2petefoth: the training this week highlighted a couple of places the docs could do with improving, are you interested in me listing them?15:22
persiaDidn't we already do that?15:23
pedroalvarezyeah, I think we did15:23
pedroalvarezand really well summarised15:25
persiaIt helps that the policy changes went through the code-review process on the patches, so lots of dicussion to get the wording right.15:26
* persia wonders a bit why the large # got approved, but isn't concerned enough to submit a policy patch candidate.15:27
ssam2I thought we were strongly pro-#15:27
persiaApparently so.  Change was merged by paulsherwood, presumably after review.  I just had failed to notice I was asserting a pro-# position when I gave my +1 on policy.15:28
pedroalvarezI'll remove it15:28
petefothpersia: we don’t do patch reviews for wiki changes I’ll get rid of the rogue #15:28
persiapetefoth: We *do* do patch reviews for wiki changes to the policy page (see the top of the policy page)15:29
persiaAs far as I'm concerned, anything else is fair game.15:29
pedroalvarezwow, I missed that bit15:29
persia(and honestly, it's probably safe to do IRC review for the # change, as with any trivial change)15:29
petefothpersia: OK I missed it too (though being pedantic it only says that changes to ‘ploicies’ are subject to review, not the plicy page :)15:30
persiaheh, you're right.  That's my fault for writing unclear policy :)15:30
persia(and so you can work around it within the text of the policy if you prefer)15:31
petefothpersia: I’ll change that. Oh, wait…..15:31
pedroalvarezI'd put that *important* bit in a highlighted note15:31
petefothI have made some changes to the policies wiki page Please will someone review the changes?15:34
petefothAny objections if I make the Policy page a bit more visible (by putting a link to it in the ‘Community’ section of the sidebar?15:35
ssam2sounds good to me15:36
pedroalvarezgo ahead15:36
* pedroalvarez drops the patch to create the highlighted note15:37
pedroalvarezI should learn to type faster15:37
petefothpedroalvarez: it’s hard to beat ‘edit in place’ for speed :)15:37
paulsherwoodreview of my change to the policy page happened on irc iirc15:38
paulsherwood(ie i screwed up the implementation of a discussed change)15:38
pedroalvarezpaulsherwood: so do you want a jetson running in big endiian?15:40
paulsherwoodyes. i even asked folks to do it, ages ago :)15:41
pedroalvarezpaulsherwood: did anybody say "this is impossible" or "it's going to take a while" ?15:41
paulsherwoodyes, someone says it's impossible. 15:42
paulsherwoodhence, now we *have* to try :)15:42
pedroalvarezaha! :)15:42
pedroalvarezI used to have some hardware running baserock in big endian.. now I don't know15:44
radiofreeben boiled it down for me at the pub, essentially it's i) build a big endian kernel ii) boot a big endian rootfs with it15:44
radiofreeit's that easy, if you're ben15:45
pedroalvarezyeah! easy like that :)15:46
pedroalvarezbut if someone thinks that it's impossible,then maybe there are more implications15:47
radiofreeoh yeah there was a big "if the hardware supports it" bit15:47
persiapaulsherwood: I thought it was in the thread about maintainers, where I proposed a ludicrously complicated scheme, and ssam2 suggested the model we use now.15:50
paulsherwoodpersia: you may be right. my memory is soso16:16
aananthssam2: where can I read proxy settings inside trove vm?17:24
ssam2aananth: what do you mean?17:24
ssam2the proxy settings that are set in lorry-controller.conf in the local-config/lorries.git repo are just for the importer process17:25
aananthssam2: I configured the git proxy and ssh proxy  on the vm hosting trove server. I need to find where the http proxy settings are stored.17:25
ssam2I'm a little confused, sorry17:26
aananthno problem :)17:26
ssam2what do you mean when you say 'git proxy' ? What did you need to do on the VM host to set one up?17:27
aananthssam2: we create a file where we call or exec corkscrew to tunnel through our proxy server for every git command. 17:29
aananthSimilarly there is a setup for ssh17:29
ssam2ah, OK17:30
aananthI can send a pdf file explaining the details.17:30
ssam2do all your connections to the Trove have to go via your proxy ?17:30
ssam2I would have thought that the proxy would only be between your company, and the internet17:30
ssam2so as long as the Trove is inside your company network, there'd be no need for a proxy17:30
ssam2and all you'd need would be to set the appropriate HTTP proxy settings in lorry-controller.conf in local-config/lorries.git, for the mirroring of git.baserock.org17:31
ssam2I don't think I explained this clearly before, I apologise for any confusion17:32
*** jonathanmaw [] has quit [Quit: Leaving]17:32
*** dutch [] has quit [Quit: Quit]17:34
aananthOk. I am just trying different options.17:34
aananthssam2: I deployed another trove vm on same machine, I forgot to add proxy.conf when I created it. Now I am adding it, is this path OK "/home/lorry/confgit/proxy.conf"17:40
ssam2aananth: no, it needs to go in the lorry-controller.conf file in the local-config/lorries.git repo17:40
ssam2it's described here:
ssam2aananth: hmm, I wonder if adding it in the confgit folder directly will be OK17:41
ssam2it's best to add the proxy.conf file to the lorries.git repo, commit it, and push17:41
ssam2but actually, /home/lorry/confgit IS that repo17:42
ssam2so if you want to just add it in there from the commandline, I guess you can :)17:42
ssam2be sure to commit, though, so other users can see it if they clone the lorries.git repo17:42
ssam2sorry, I hadn't realised that it could be done in that way17:43
aananthsam: I committed it. Should I re-boot or the server will pick the configs?17:46
ssam2i'm not sure.17:46
ssam2may as well reboot :)17:47
aananthssam: Any problem you see in this "Oct 31 23:20:28 vtsc-trove sshd[2305]: reverse mapping checking getaddrinfo for vtsc-trove [::1] failed - POSSIBLE BREAK-IN ATTEMPT!"17:54
ssam2I see that in my Trove too17:54
ssam2nothing to worry about I think17:54
aananthOk. How about this? "Oct 31 23:19:27 vtsc-trove lighttpd[194]: 2014-10-31 23:19:27: (mod_fastcgi.c.2673) FastCGI-stderr: GitanoCommandFailure: Failed to run "ls" on Gitano"17:56
ssam2hmm, that does sound bad17:56
ssam2I'm afraid this is quite outside my area of expertise17:57
ssam2I've not worked on Trove much, so it may be better to wait until next week to look at this problem when the Trove experts are back in the office17:57
ssam2could you paste the lorry-controller.conf file ?17:57
aananthI will send more lines from log around that point.17:57
ssam2more lines of log would be good, too17:58
*** mariaderidder [] has quit [Quit: Ex-Chat]17:58
aananthssam2: Log is available in
ssam2aananth: ok, it's getting 'Name or service not known' still18:00
aananthssam2: Have a look at my lorry-controller.conf Sounds like I missed to type https username password.18:02
ssam2the 'protocol' field is correct18:06
ssam2did you do 'git add' of the proxy.conf and 'git commit' ?18:07
aananthssam2: I think adding from commandline within VM is not accepted. 18:08
ssam2it could be a permissions thing18:09
aananthI was able to type committ message and commit it, but now they dont exit.18:09
ssam2check all the files are still owned by 'lorry' user and 'lorry' group18:09
ssam2I'm going to have to go now, sorry18:10
ssam2I'll be back Monday though, I'll be happy to continue looking through this with you18:10
ssam2and hopefully I can work on making this process easier in future!18:10
aananthOk Sam.18:11
*** rdale_ [] has quit [Ping timeout: 272 seconds]18:11
*** ssam2 [] has quit [Quit: Leaving]18:12
*** aananth [~caananth@] has quit [Quit: Leaving]18:12
*** tiagogomes [] has quit [Ping timeout: 240 seconds]18:20
*** franred [] has quit [Quit: Leaving]18:43
*** abdul__ [uid49809@gateway/web/] has quit [Ping timeout: 272 seconds]19:12
*** abdul__ [uid49809@gateway/web/] has joined #baserock19:12
jjardonmmm, seems nobody is building GNOME in a wayland-only system22:53
* jjardon cleaning x11 deps from several modules22:53

Generated by 2.14.0 by Marius Gedminas - find it at!