| === freeflying_away is now known as freeflying | ||
| hobbyBobby | hi all, struggling to set up virtual box and I have a sos report here if you're willing to help | 02:05 |
|---|---|---|
| hobbyBobby | *with juju | 02:05 |
| hobbyBobby | from what I can tell it can't resolve the mongo address because of some networking issue | 02:06 |
| hobbyBobby | wow im daft | 02:18 |
| rick_h_ | thumper: the config hook issue from last night was because I ran a make command that didn't background. So it never 'committed' and wouldn't rerun | 02:53 |
| rick_h_ | thumper: just fyi, I was a moron that is all :) | 02:54 |
| thumper | :) | 03:24 |
| === mwhudson is now known as zz_mwhudson | ||
| === _mup__ is now known as _mup_ | ||
| === bloodearnest_ is now known as bloodearnest | ||
| === zz_mwhudson is now known as mwhudson | ||
| === mwhudson is now known as zz_mwhudson | ||
| === timrc is now known as timrc-afk | ||
| melmoth | hola... i m trying to remove a ceph-osd unit. it fails when it try to destroy the realtion with ceph-mon;: | 10:41 |
| melmoth | 2014-01-28 10:40:05 ERROR juju.worker.uniter uniter.go:350 hook failed: fork/exec /var/lib/juju/agents/unit-ceph-osd-2/charm/hooks/mon-relation-departed: cannot allocate memory | 10:41 |
| melmoth | the cpeh-osd box has 500M of ram | 10:42 |
| melmoth | what can i do ? | 10:42 |
| jamespage | melmoth, not much - you could do a juju terminate-machine --force # | 12:34 |
| jamespage | that will kill it outright for you | 12:35 |
| melmoth | i destroyed all environment, and redeployed again with 1G of ram on esd node | 12:35 |
| melmoth | jamespage seems to work better | 12:35 |
| jamespage | melmoth, yeah - anything complicated needs something bigger than 500M | 12:35 |
| jamespage | melmoth, we bumped the default instance memory in our openstack qa environment for that reason | 12:36 |
| jamespage | 900M | 12:36 |
| === CyberJacob is now known as CyberJacob|Away | ||
| === timrc-afk is now known as timrc | ||
| === freeflying is now known as freeflying_away | ||
| bloodearnest | anyone know of any charms that can utilise round-robin DNS? | 15:16 |
| rick_h_ | bloodearnest: how would that work? Wouldn't you just assign the cnames in dns and point at the exposed front end services? | 15:55 |
| bloodearnest | rick_h_: it's for the frontend that I want it :) | 15:56 |
| bloodearnest | rick_h_: I was thinking that when a new unit is added, it adds it's ip to the pool of ips in the DNS server | 15:56 |
| rick_h_ | bloodearnest: oic, so you'd need a dns server charm that could accept a realtion to web app charms then | 15:56 |
| bloodearnest | and when the unit is removed, it pops it's ip from the pool | 15:56 |
| rick_h_ | right | 15:57 |
| bloodearnest | rick_h_: in theory, the DNS wouldn'y need to be charmed (as DNS tends to be non-juju-env infrastructure), but it'd need an api for a charm to use | 15:57 |
| bloodearnest | I guess | 15:57 |
| rick_h_ | bloodearnest: https://jujucharms.com/sidebar/~hazmat/precise/aws-route53-2/#bws-readme looks interesting | 15:58 |
| bloodearnest | not 100% convinced rr-DNS is the best solution for this problem, just seeing if anything has already been done | 15:58 |
| rick_h_ | bloodearnest: I think honestly most people just put haproxy in fron of it | 15:59 |
| rick_h_ | and then use the realtions to that to control adding/removing | 15:59 |
| bloodearnest | rick_h_: sure, but this is aimed at scaling 1M+ long-lived https connections | 16:00 |
| bloodearnest | s/https/tls, it's not http | 16:00 |
| rick_h_ | bloodearnest: ah k | 16:00 |
| bloodearnest | some http clients don't support rr-DNS well anyway, so we may have to DIY it | 16:01 |
| bloodearnest | rick_h_: tell you all about it in Cape Town next week :) | 16:03 |
| rick_h_ | bloodearnest: sounds good | 16:03 |
| * bloodearnest has a list of juju gui feature ideas to throw at someone... :) | 16:04 | |
| * rick_h_ ducks | 16:05 | |
| jcastro | negronjl, https://bugs.launchpad.net/charms/+source/mongodb/+bug/1267222 | 16:11 |
| _mup_ | Bug #1267222: Race condition when deploying a simple replica set <audit> <mongodb (Juju Charms Collection):New for negronjl> <https://launchpad.net/bugs/1267222> | 16:11 |
| jcastro | hey if you're too busy to fix that if you could post where you think the problem is I can have one of our dudes take a look? | 16:11 |
| jcastro | arosales, mbruzek, lazyPower, marcoceppi: I have a newish proposal for the audit | 16:13 |
| jcastro | so like, we sorted the list by downloads, but we're now getting into the "not top 10". | 16:14 |
| jcastro | IMO we should be smarter on targetting now, so for example next on the list is "phpmyadmin", I'd like to skip that and go to ganglia next. | 16:14 |
| marcoceppi | jcastro: sounds good to me | 16:14 |
| marcoceppi | feel free to re-order the list as you see fit | 16:15 |
| jcastro | like, go after the big services, nothing against myphpadmin | 16:15 |
| lazyPower | Agreed, if you need to change my priority queue i have no problems with it. | 16:15 |
| jcastro | marcoceppi, I thought about that but I do want to leave it sorted by downloads, because people are using it, and I get that | 16:15 |
| lazyPower | jcastro, actually if you want you can assign a batch of 3 or 4 charms so i have "up next" targets | 16:16 |
| jcastro | but like, logstash-agent is #93 and that's probably way more important than phpmyadmin | 16:16 |
| jcastro | lazyPower, I am putting "should be higher priority" in the notes column | 16:17 |
| jcastro | then I'll just start on their readmes, so "follow the green" I guess? | 16:17 |
| jcastro | lazyPower, hadoop worked fine for me btw | 16:18 |
| lazyPower | thats so strange | 16:18 |
| jcastro | is it not working for you at all? | 16:18 |
| lazyPower | Which version of juju are you running jorge? | 16:18 |
| lazyPower | no, its got an inline hack to parse the machine id, and since my LXC containers do not generate FQDN's its failing at that point in the install hook. | 16:19 |
| lazyPower | This may be specific to my setup thats causing the failure, and I feel like this is a growing problem, that probably needs addressed this week. | 16:19 |
| jcastro | lazyPower, 1.17 | 16:22 |
| === rogpeppe2 is now known as rogpeppe | ||
| jcastro | lazyPower, my local provider is broken so I was on AWS | 16:23 |
| lazyPower | aaahhh see, i said it works in cloud providers :) | 16:23 |
| * jcastro is on trusty | 16:23 | |
| jcastro | oh, I missed that! | 16:23 |
| lazyPower | glad its +1'd though | 16:24 |
| lazyPower | the tests for that are going to be hairy | 16:24 |
| arosales | jcastro, agreed and some charms may not have had the life span to attract the higher number, but if you know ones we should hit first perhaps highlight those in orange | 16:26 |
| jcastro | yeah, like I said, nothing wrong against phpmyadmin, but like, we should be wailing on things like cassandra and so on | 16:27 |
| arosales | jcastro, I got your feedback addressed in https://code.launchpad.net/~a.rosales/charms/precise/jenkins/update-metadata-readme/+merge/203181 | 16:28 |
| arosales | lazyPower, mbruzek good time to do some +1's on the queue this morning | 16:36 |
| lazyPower | arosales, already ahead of you :) Working through OpenMRS as we speak | 16:36 |
| arosales | lazyPower, mbruzek hpcc and openvpn-* are also good candidates | 16:37 |
| mbruzek | arosales, take from the top of the review queue or the red tracking box in leankit? | 16:38 |
| lazyPower | good question ^ | 16:38 |
| lazyPower | arosales, jcastro can i get another +1 on this? the charm is solid from my testing standpoint. https://bugs.launchpad.net/charms/+bug/1125869 | 16:38 |
| _mup_ | Bug #1125869: New charm: postfix <Juju Charms Collection:New for jose> <https://launchpad.net/bugs/1125869> | 16:38 |
| arosales | mbruzek, lazyPower generally from the top as those have been in the queue the longest | 16:39 |
| jcastro | looking | 16:39 |
| hazmat | bloodearnest, rick_h_ that aws-route53 charm isn't rr dns fwiw.. but route53 can do it, and low latency routing across multiple backends. | 16:41 |
| bloodearnest | hazmat: yeah, I'd love to use route53, but it's ec2 only, and my target is openstack. idk of an equivalent/similar service for openstack | 16:43 |
| hazmat | bloodearnest, route53 is not ec2 only. | 16:44 |
| hazmat | its dns | 16:44 |
| rick_h_ | hazmat: right, I was thinking more an example of doing DNS tweaks based on relations via a charm | 16:44 |
| hazmat | there are some ec2 specific provisions granted, but nothing prevents general use | 16:44 |
| hazmat | rick_h_, thats exactly what the aws-route53 charm does | 16:44 |
| lazyPower | oh brilliant | 16:45 |
| rick_h_ | hazmat: right | 16:45 |
| hazmat | most of my aws charms back into https://github.com/kapilt/awsjuju .. a couple of other goodies in there | 16:45 |
| hazmat | but i'd label it pre-alpha | 16:46 |
| hazmat | except the automated backup stuff | 16:46 |
| hazmat | which is just alpha :-) | 16:47 |
| bloodearnest | hazmat: huh, TIL. Maybe an option then. | 16:55 |
| bloodearnest | thanks | 16:55 |
| jose | jcastro: did you get to take a look at the postfix charm? | 17:35 |
| jcastro | not yet no | 17:35 |
| jcastro | I see it on the list though | 17:35 |
| jose | well, I'll wait then :) | 17:35 |
| jamespage | adam_g, I switched all the icehouse branchs to openstack-charmers btw - https://code.launchpad.net/~openstack-charmers | 17:48 |
| === zz_mwhudson is now known as mwhudson | ||
| === mwhudson is now known as zz_mwhudson | ||
| === CyberJacob|Away is now known as CyberJacob | ||
| === _mup__ is now known as _mup_ | ||
| === zz_mwhudson is now known as mwhudson | ||
| === gary_poster is now known as gary_poster|away | ||
| === freeflying_away is now known as freeflying | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!