[00:34] The wordpress example seems _actually_ broken [00:44] <_mup_> juju/trunk-merge r342 committed by kapil.thangavelu@canonical.com [00:44] <_mup_> merge trunk [00:44] <_mup_> juju/ftests r11 committed by gustavo@niemeyer.net [00:44] <_mup_> Reduced wget retries. [00:52] Oh dear.. [00:52] Ben didn't revert the changes in the example formulas [00:52] :-( [00:53] niemeyer, i thought i saw the change go through [00:53] hazmat: He changed to using hostname.. [00:54] niemeyer, which is better than hardcoding to ec2 md server [00:54] is hostname not routable? [00:54] one c2 [00:55] hazmat: I don't know, but the ftests are not passing anymore.. I'll give this a shot [01:33] <_mup_> Bug #862680 was filed: EC2 api call TerminateInstances returns the wrong response < https://launchpad.net/bugs/862680 > [01:44] hazmat: Yep, works [01:44] niemeyer, cool [01:47] <_mup_> juju/trunk r372 committed by gustavo@niemeyer.net [01:47] <_mup_> Revert example changes introduced with lxc-omega so that they [01:47] <_mup_> work again on EC2. [01:57] <_mup_> juju/unit-with-addresses r404 committed by kapil.thangavelu@canonical.com [01:57] <_mup_> merge trunk and resolve conflict [02:01] oh.. the store work is in [02:01] awesome!! [02:01] * hazmat does a dance [02:02] hmm [02:03] i keep getting incompatible protocol errors [02:03] niemeyer, unit tests broke with 372 [02:04] hmmm ssh keys not found [02:04] for local provider [02:04] needs to be mocked [02:10] hazmat: For the incompatible errors you'll need to use the trunk on juju-branch [02:11] hazmat: The version was bumped with the store changes [02:11] ic [02:14] Hmmm.. butler seems to have jumped a revision [02:19] Yeah, I broke it [02:22] niemeyer, i don't see how exactly.. all your changes appear to be in formulas [02:22] hazmat: I broke butler itself, in the ftests [02:22] hazmat: It was jumping revisions [02:22] niemeyer, ah [02:23] hazmat: Unrelated to the above issue [02:23] niemeyer, so trunk examples are using the metadata sever? [02:23] hazmat: Yeah, I reverted it so tests could pass [02:23] hazmat: We have to find a way to make both happy [02:23] and also have to fix the unittests so they don't rely on install ssh keys [02:23] installed [02:24] I've put it to rerun the tests since 369 [02:24] unittests will likely all be broken [02:25] since I removed the fake id keys [02:25] ssh id_*sa keys [02:25] The ec2-wordpress one should pass on 372 [02:28] i'll add the unit-info cmd to resolve charm addressing.. but realistically that's not till tomorrow evening [02:29] very happy to see the repository work go in [02:42] hazmat: That's awesome, thanks a lot [02:42] hazmat: Yeah, quite exciting isn't it? [02:42] hazmat: Almost there! [02:47] feels very disconcerting to login into byobu [02:48] hmm.. odd serialization of the charm id to escaped form in zk [02:50] hazmat: It's almost like a url encoding [02:50] niemeyer, yeah.. but its not needed for the zk node name [02:50] hazmat: It is needed, due to slashes [02:50] ah [02:52] hazmat: Agreed about byobu.. Eric Hammond is also complaining publicly about it [02:57] Alright, past bed time here.. [02:58] My talk is first in the morning tomorrow [02:58] See you all tomorrow! [03:17] <_mup_> juju/trunk r373 committed by kapil.thangavelu@canonical.com [03:17] <_mup_> merge unit-with-addresses [r=niemeyer][f=859308] [03:17] <_mup_> Units detect and record their public/private addresses into the unit state, [03:17] <_mup_> in a provider specific manner. [03:20] <_mup_> juju/lxc-provider-rename-local r406 committed by kapil.thangavelu@canonical.com [03:20] <_mup_> merge pipeline and resolve conflict [03:21] <_mup_> juju/local-origin-passthrough r406 committed by kapil.thangavelu@canonical.com [03:21] <_mup_> merge pipeline and resolve conflict [03:24] <_mup_> juju/trunk r374 committed by kapil.thangavelu@canonical.com [03:24] <_mup_> merge unit-with-addresses [r=niemeyer][f=859308] [03:24] <_mup_> Command line tools use unit addresses (status, ssh, debug-hooks, etc). [03:43] <_mup_> juju/trunk r375 committed by kapil.thangavelu@canonical.com [03:43] <_mup_> merge lxc-provider-rename-local [r=niemeyer][f=860982] [03:43] <_mup_> Rename lxc provider to local provider. [03:57] <_mup_> juju/unit-relation-with-addr r407 committed by kapil.thangavelu@canonical.com [03:57] <_mup_> unit relations are automatically created with the unit's private address [04:06] <_mup_> juju/unit-relation-with-addr r408 committed by kapil.thangavelu@canonical.com [04:06] <_mup_> merge trunk [04:50] <_mup_> juju/local-origin-passthrough r409 committed by kapil.thangavelu@canonical.com [04:50] <_mup_> update juju-create to use jujuorigin for bzr branches instead of a separate jujusource var [04:52] <_mup_> juju/local-origin-passthrough r410 committed by kapil.thangavelu@canonical.com [04:52] <_mup_> merge latest env-origin [04:56] <_mup_> Bug #862987 was filed: Local provider should respect juju-origin < https://launchpad.net/bugs/862987 > [05:20] <_mup_> juju/local-origin-passthrough r411 committed by kapil.thangavelu@canonical.com [05:20] <_mup_> use default origin util func if none specified [14:47] anyone know if Gustavo's around today? [14:54] rog: I expect he'll be around at some stage, but he's theoretically not [14:56] j/win 17 [14:56] aerrr [14:57] fwereade: ah, ok. i thought he was back today from the conference [14:58] rog: hm, I thought it was still on today, maybe he's travelling [15:00] ah, you're right, it's still on. [15:00] didn't think it was. [15:10] Hi all [15:12] is this canonical juju's user channel ? [15:14] $ juju deploy --repository=examples mysql [15:14] Charm 'cs:natty/mysql' not found in repository https://store.juju.ubuntu.com/charm [15:14] 2011-09-30 15:13:49,717 ERROR Charm 'cs:natty/mysql' not found in repository https://store.juju.ubuntu.com/charm [15:14] this command is what's written when I follow the tutorial [15:14] is the charms repository down , or isn't yet up ? [15:14] <_mup_> Bug #863374 was filed: deploy documentation is out of date < https://launchpad.net/bugs/863374 > [15:14] ok [15:14] thanks [15:15] the ticket doesn't mention where is the repository [15:15] " Recent changes to repository structure and deploy args are not recorded in the documentation." [15:15] there's no more repo ? [15:15] or none , for now ? [15:19] fwereade: ^ ? [15:19] xerxas, ha, sorry, I was just fixing the docs [15:19] ;) [15:19] xerxas: $ juju deploy --repository=examples local:mysql [15:20] xerxas: if you're not running oneiric, you may need to do: [15:20] $ juju deploy --repository=examples local:mysql [15:20] [Errno 2] No such file or directory: '/home/ubuntu/examples/natty' [15:20] 2011-09-30 15:20:18,818 ERROR [Errno 2] No such file or directory: '/home/ubuntu/examples/natty' [15:20] > xerxas: $ juju deploy --repository=examples local:oneiric/mysql [15:20] ok , waiting ;) [15:21] $ juju deploy --repository=examples local:oneiric/mysql [15:21] [Errno 2] No such file or directory: '/home/ubuntu/examples/oneiric' [15:21] 2011-09-30 15:20:55,711 ERROR [Errno 2] No such file or directory: '/home/ubuntu/examples/oneiric' [15:21] I'm I supposed to retrieve the repo locally ? [15:21] ah: yu don't have the examples repository in your working dir? [15:23] xerxas: I'm not sure where you got juju from, but there should be an "examples" directory *somewhere*, which should have a subdir called "oneiric" [15:24] ok [15:24] in usr/share/doc , i suppose [15:25] I got it from the ppa as written in the doc [15:25] anyway, thanks for helping ;) [15:25] and , I think the charm repository doesn't have charms that fits my need anyway [15:28] xerxas: hold on a mo, I'm just installing it myself in the hope I can track it down [15:28] xerxas: the examples repository might not, but we have plenty of other charms :) [15:28] xerxas: what are you looking for? [15:33] xerxas: https://code.launchpad.net/charm [15:39] Heyo [15:39] heya niemeyer [15:42] Argh.. really bad connectivity here :( [15:43] bcsaller: Heyo [15:43] niemeyer: hi [15:44] niemeyer: hiya! [15:45] bcsaller: Had to revert the example changes yesterday so the ftests could be happy [15:45] bcsaller: The modified examples were not working in EC2 [15:45] rog: Yo! [15:46] rog: Sorry for not giving you much feedback this week man.. it's being well beyond active [15:46] rog: Promise to do a better job next week [15:47] niemeyer: no probs. i do hope you might come around to my changes! :-) [15:48] rog: I like them from one side, but the problem of trashing information is a real one [15:48] rog: I don't want to have to rollback to the current version because we need details that are being dumped [15:48] niemeyer: there's no information being dumped, honest. [15:48] <_mup_> Bug #863400 was filed: examples repository is not installed from PPA < https://launchpad.net/bugs/863400 > [15:49] niemeyer: the code will instantly panic if it happens... which i'll bet it won't. [15:50] Erm.. [15:50] Bad connectivity indeed :( [15:51] rog: The change isn't improving things enough to justify dropping information in the protocol [15:52] niemeyer_: if the paths start actually meaning something, then it's easy to change the channel type to reflect that. the main thing is separating session status events from node status events. [15:53] niemeyer_: currently the code checks on every protocol message that the path is what is expected. so we're guaranteed that no information is being dropped. [16:04] fwereade: I'not looking for something special [16:04] I'm just testing juju [16:04] ;) === hazmat` is now known as hazmat [16:05] xerxas: I've looked into it, and for some reason the examples repository (which should be installed to /usr/share/doc/juju) isn't there at the moment [16:05] xerxas: it's being worked on [16:05] ok [16:06] thanks for the information [16:06] xerxas: if you want to test quickly, you can always "bzr branch lp:juju" and use the examples repo in there [16:06] xerxas: otherwise, expect an update to the PPA soon [16:06] ok [16:07] thanks so much [16:07] xerxas: a pleasure :) [16:07] fwereade, also when committing a breaking change on trunk, (version increment) its probably nice to manually trigger the ppa build [16:08] hazmat: thank you, good to know; er, how do I do that? [16:08] hmm.. i guess it only happened because i was running trunk and deploying ppa [16:09] * hazmat looks for the build link [16:10] fwereade, https://code.launchpad.net/~juju/+recipe/juju from here (its linked off the trunk code view) [16:11] a stable ppa will help some as well [16:12] hazmat: thanks :) [16:16] fwereade, did you discuss with niemeyer a timeline for getting the remote end operational? [16:16] fwereade, great work btw. its very exciting to see this stuff getting done [16:16] hazmat: afraid not, all I know is that niemeyer's working on it when he can [16:17] hazmat: thanks :) [16:19] oh, bother, gtg: nn everyone, happy weekends :) [16:22] fwereade: have a good one! [16:23] fwereade, cheers [16:23] * hazmat relocates to a better table [17:06] i'm off to enjoy the sunshine. see y'all monday. [17:08] http://arstechnica.com/business/news/2011/09/google-devops-and-disaster-porn.ars [17:08] surge conference writeup [17:08] rog, have a good one [17:14] <_mup_> juju/unit-info-cli r412 committed by kapil.thangavelu@canonical.com [17:14] <_mup_> amp api for unit-get cli [17:16] hazmat: nice. XML for the lose. [17:18] rog, its always the cascading failures that burn [17:29] Two more tests broken when run inside clean chroot's [17:30] you guys should really be running the test suite with HOME=`mktemp -d /tmp/foo.XXXXX` [17:30] SpamapS, sounds like a good idea [17:30] http://paste.ubuntu.com/699960/ [17:30] jimbaker: I'm kind of surprised WTF didn't pick that up [17:31] or maybe it has [17:31] I see fails since 366 [17:31] SpamapS, yeah, i noticed too [17:33] SpamapS, yeah... i've seen those on wtf but its been unstable (wtf) [17:34] that's worth a bug report [17:34] trying to make the 'unit-get' cli happen so we can stop hard-coding address in formula [17:34] just found elmo at surge [17:35] jimbaker, is juju-origin ready to merge? [17:39] i've two branches in dev that i'd like to see go in that our based on it [17:39] SpamapS, so 361 is uploaded, do we get to push any new things in? [17:40] i'm a little confused on the schedule for uploads [17:40] yes [17:40] one more upload. :) [17:40] hazmat: technically we can upload all the way up to release day.. but as of yesterday, with FinalFreeze, we are at the release team's pleasure. [17:47] <_mup_> Bug #863499 was filed: local provider tests fail with an empty home directory < https://launchpad.net/bugs/863499 > [17:52] hazmat, juju-origin is waiting on approval. i believe it's ready, that's why it's in review [17:53] hazmat, and i did see your branch that's waiting on it, looks good in my review of it [17:53] hazmat, just about to head to lunch, but i'll complete the review right after that [17:55] jimbaker: The ec2-wordpress ftest is hanging after a defer error and never returning [17:55] jimbaker: Will paste the output on an error [17:55] Erm on a bug [17:55] Just pushed a fix up for bug 863499 .. would be awesome if that got a fast track for reviews so I can enable running the test suite on the daily build PPA [17:55] <_mup_> Bug #863499: local provider tests fail with an empty home directory < https://launchpad.net/bugs/863499 > [17:56] SpamapS: Will check it out right away [17:56] cool thanks. :) [17:57] that will help streamline the process as we move into the final upload to 11.10 [17:57] jimbaker: https://bugs.launchpad.net/juju/+bug/863510 [17:57] <_mup_> Bug #863510: destory-environment errors and hangs forever < https://launchpad.net/bugs/863510 > [17:57] <_mup_> Bug #863510 was filed: destory-environment errors and hangs forever < https://launchpad.net/bugs/863510 > [17:57] SpamapS: Hah, sweet.. I was going to ask bcsaller about this one [17:57] SpamapS: You rock [17:58] it will be using the same debian/ dir as Ubuntu. :) [17:58] niemeyer, thanks for this [17:58] jimbaker: No worries.. don't know exactly what is going on there, but this blocked the ftests since yesterday evening [17:58] Will also fix bug 863400 [17:58] <_mup_> Bug #863400: examples repository is not installed from PPA < https://launchpad.net/bugs/863400 > [17:59] jimbaker: It's hanging due to the unhandled error for sure, but reason why it's erroring isn't clear [18:00] niemeyer, ok, i'll think about this over lunch (have some friends waiting on me here) [18:00] biab [18:00] jimbaker: Thanks [18:09] <_mup_> juju/unit-relation-with-addr r409 committed by kapil.thangavelu@canonical.com [18:09] <_mup_> setup a unit address by default, so relations have some useful value there by default, fix up hook tests that where checking raw value to include address info [18:11] <_mup_> juju/trunk r376 committed by gustavo@niemeyer.net [18:11] <_mup_> Merge fix-failing-local-tests branch by Clint [r=niemeyer] [trivial] [18:11] <_mup_> This fixes tests so that they don't depend on installed user ssh keys. [18:11] <_mup_> Bug #863526 was filed: Juju agents do not handle reboots < https://launchpad.net/bugs/863526 > [18:16] Jim Fulton on stage right now [18:16] hazmat: I filed bug 863526 but now I'm wondering if it may be a duplicate [18:16] <_mup_> Bug #863526: Juju agents do not handle reboots < https://launchpad.net/bugs/863526 > [18:19] SpamapS: the "production" tag is a neat idea [18:24] <_mup_> juju/unit-relation-with-addr r410 committed by kapil.thangavelu@canonical.com [18:24] <_mup_> fix up additional test fallout from relations with unit addresses [18:24] niemeyer: thanks, I've been trying to be very discerning in adding it.. only things that we *cannot* live without for production. Nice to haves are different. [18:24] niemeyer: btw, is bug 712476 valid anymore with the new repo work? [18:24] <_mup_> Bug #712476: Ensemble deploy should have a --force option for reuploading a formula < https://launchpad.net/bugs/712476 > [18:24] niemeyer, nice [18:25] niemeyer, i think my friend alan is out there as well, he wanted to come out and give tribute to dorneles [18:27] SpamapS, do you think cross-az is required for prod? [18:27] hazmat: no [18:27] hazmat: two environments and config settings can work around that [18:28] hazmat: I think its pretty important though! [18:28] for whizbang awesomeness [18:31] <_mup_> juju/unit-info-cli r414 committed by kapil.thangavelu@canonical.com [18:31] <_mup_> merge pipeline and resolve conflict [18:48] <_mup_> juju/unit-info-cli r415 committed by kapil.thangavelu@canonical.com [18:48] <_mup_> add unit-get cli [18:52] <_mup_> juju/unit-info-cli r416 committed by kapil.thangavelu@canonical.com [18:52] <_mup_> update examples to use unit-info to get addresses [18:54] <_mup_> juju/unit-info-cli r417 committed by kapil.thangavelu@canonical.com [18:54] <_mup_> update the php example.. not sure what its around for [19:59] bcsaller, have you tried running local provider recently? [20:23] <_mup_> juju/local-origin-passthrough r413 committed by kapil.thangavelu@canonical.com [20:23] <_mup_> fix some local provider problems [20:52] negronjl: hey Juan... can you pls change owner of lp:charm/mongodb to charmers? trying to push name changes [20:52] m_3: sure...give me a sec [20:53] negronjl: danke [20:57] m_3: done [20:57] m_3: let me know if it works [20:59] negronjl: worked... um... like a charm [21:00] bu-dumpdum [21:03] and that made him quit the channel? heh [21:03] tough room [21:59] hmm. using 'juju deploy --repository=$REPO local:$charm', $REPO/oneiric/charm1 deploys find but $REPO/oneiric/charm2 doesnt. get a charm not found error on the second [22:09] <_mup_> juju/local-origin-passthrough r414 committed by kapil.thangavelu@canonical.com [22:09] <_mup_> unit container deploy pulls origin from environment [22:09] adam_g, is it possible there's a metadata.yaml parse error on charm2? [22:10] that breaks the repo find algorithm as i recall [22:20] hazmat: ah, it was some bad strings in config.yaml [22:20] adam_g, yeah.. i went back and forth if that should break things or not [22:21] adam_g, probably worthwhile for us to at least log a message [22:21] if we're in verbose mode [22:23] * hazmat loves wifi hurtling down the train tracks [22:26] hazmat: i found "ScannerError: ScannerError() [22:26] " [22:27] in the tracing i was doing.. not sure if thats catchable or what [22:27] * adam_g <- python n00b [22:28] adam_g, we can definitely catch it and report it, its just not clear that the log would be useful per se against it [22:29] adam_g, could you pastebin the config file [22:29] almost done fixing up the local provider, i'll have a look at that next [22:31] hazmat: http://paste.ubuntu.com/700121/ [22:31] the issue was the description of 'virt-type' wasn't quoted [22:32] and the : was throwing off the parse [22:32] it'd certianly be useful to just point out that there was in error parsing config.yaml, or metadata.yaml [22:42] hi bcsaller [22:42] hazmat: hey, power was out for a while, just came back on [22:43] bcsaller, noticing a few problems in the local provider, some from the rename, some with the origin stuff, just curious if you been able to run it recently [22:44] not quite sure where its going wrong doing the origin stuff, adding a new log for the customize script [22:44] haven't tried it with the origin stuff [22:45] yeah.. the origin stuff is new, but even before then i found some other problems in the omega merge, some broken path handling split between the container and the juju home location [22:45] hazmat: but niemeyer reverted the hostname stuff in the examples [22:45] to an ec2 only version [22:45] so his test pass [22:45] bcsaller, i've got a unit-get cli [22:45] implemented [22:46] which fixes charms to get addresses in cross-provider manner [22:46] if you didn't change the formula though that could be it [22:46] ahh [22:46] also populates unit relation settings with private address by default [22:46] so charms don't need to do that anymore [22:46] going to try and spend next week on the upstartifying everything and handling disconnect/reconnects [22:47] just wanted to see the unit-get working with local provider, when i hit some of these issues [22:51] I'll test it soon [22:58] bcsaller, does something delete the juju-create script in the container? [22:59] no, but currently I think its just written to tmp [23:01] bcsaller, so the customize is only run on the master ? [23:01] hazmat: yeah, and then clone rewrite the hostname [23:02] don't want to have to apt-get install all the stuff for each node [23:06] debugging this on a mobile hotspot is still a bit painful [23:08] hazmat: is that because of the initial download? [23:08] bcsaller, it looks like using a chroot to run juju-create still leaks a bunch of env variables which are problematic [23:08] SpamapS, its setup with apt-cacher-ng, but i'm doing a bzr branch for origin [23:09] although i switched that out to a lightweight checkout [23:09] that always helps. :) [23:09] truthfully, we should be doing bzr export [23:09] so its about as fast as it can be [23:09] SpamapS, well its a dev thing anyways.. so having the option to link it back is useful still imo [23:09] i've definitely done some remote debug/tests/commits [23:10] hazmat: should we write an upstart job that deletes itself at the end? [23:10] I 'spose [23:10] bcsaller, its fine.. i was just wondering cause i wanted to execute it.. but chroot in, wouldn't give me the same fs.. i ran into a pty allocation problem ssh'ing.. just curious if it was still there if i could manage to ge tin [23:11] bcsaller, looks like the problem is no setuptools for the develop on a branch [23:11] back in a moment, cafe caboose trip [23:49] SpamapS, do you know how to get lots out of upstart? [23:49] s/logs [23:54] hazmat: indeed I do [23:54] ... [23:55] hazmat: initctl log-priority debug works [23:55] hazmat: you might just want 'initctl log-priority info' [23:55] hazmat: debug is a bit ridiculous [23:55] hazmat: if you want the programs' logs.. you have to redirect their stdout/stderr.. [23:56] SpamapS, thanks