[05:13] <koolhead17> hi all
[15:26] <gary_poster> SpamapS, hazmat, hi.  we are planning to be ready for a charm review next week, Tues or Wed, so we'll try to find someone willing to work with us then.  We also plan to take time to assemble thoughtful feedback about what we've experienced with juju so far.  Our thoughts are that we ought to write that out among ourselves before the review, because we think it would be good to have a "before we've heard a lot of explanations
[15:26] <gary_poster>  of what we did wrong or misunderstood" perspective.  Then we'd send that to the juju@ list or something.  However, the only reason to do that is to try to be helpful to you all.  So, should we do something like that?  Any suggestions??
[15:31] <hazmat> gary_poster, that sounds very useful and should help us improve how to give new users a good experience and things that need to worked on. i'm hoping some of it is just better documentation, in which case i'd invite you guys to help contribute to make it better its at lp:juju/docs
[15:33] <hazmat> use case, or review, problems encountered, feedback is good
[15:34] <gary_poster> hazmat, us helping with docs: definitely willing to do that, if we all settle on that being a valuable action item for a given item
[15:34] <gary_poster> cool thanks hazmat, we'll proceed with plans.
[18:56] <koolhead17> hi all
[19:01] <jcastro> SpamapS: m_3_ can we have a G+later to catch up on charms?
[19:01] <jcastro> I have questions and ideas!
[19:05] <koolhead17> okey guys am done with my juju presentation today and there goes the slides http://www.slideshare.net/koolhead17/juju-11516934
[19:09] <m_3_> jcastro: sure man
[19:10] <SpamapS> now is good for me
[19:10] <SpamapS> jcastro: ^^
[19:10] <jcastro> ok
[19:10] <jcastro> starting up G+
[19:10] <m_3_> yeah, I guess now's a good time
[19:12] <jcastro> m_3_: SpamapS invite sent
[19:52] <jcastro> can someone review my docs mp: https://code.launchpad.net/~juju/juju/docs/+merge/90757
[19:52] <jcastro> hazmat!
[19:54] <jdstrand> hey, is it possible to specify the mirror I want to use to debootstrap with juju?
[19:56] <jcastro> hazmat: if you make a "juju-reviewers" subteam in charge of lp:juju/docs I'd be more than happy to rock that for you.
[19:56] <jcastro> and then I can fix the docs.
[20:01] <jdstrand> SpamapS: is it possible to specify the mirror I want to use to debootstrap with juju?
[20:02] <SpamapS> jdstrand: it uses apt-cacher-ng I think
[20:02] <SpamapS> jdstrand: so maybe you can tell apt-cacher-ng to use your preferred mirror?
[20:03] <jdstrand> ok, thanks
[20:03] <jdstrand> apt-cacher-ng did seem to get pulled in
[20:04] <SpamapS> hmm maybe not
[20:04] <jdstrand> backends_ubuntu.default
[20:04] <SpamapS> jdstrand: the debootstrap part only happens once per release of Ubuntu per week..
[20:04] <jdstrand> that has http://archive.ubuntu.com/ubuntu/
[20:04] <SpamapS> jdstrand: so /var/cache/lxc/{oneiric,precise,etc}
[20:05] <SpamapS> jdstrand: its possible that just setting 'MIRROR' before calling 'juju bootstrap' will actually work
[20:05] <jdstrand> interesting
[20:06] <SpamapS> jdstrand: but there's so much indirection between calling 'juju bootstrap' and calling 'debootstrap' .. no guarantees ;)
[20:06] <jdstrand> ok, I'll fiddle with it
[20:06] <SpamapS>     debootstrap --verbose --components=main,universe --arch=$arch --include=$packages $release $cache/partial-$arch $MIRROR
[20:06] <SpamapS> Thats where its called.. as part of /usr/lib/lxc/templates/lxc-ubuntu
[20:06] <jdstrand> ah
[20:07] <SpamapS> no.. its called via sudo .. so probably will be sanitized out
[20:07] <SpamapS> bummer
[20:07] <SpamapS> oh .. hmm. maybe not
[20:09] <jdstrand> sudo dpkg-reconfigure -pmedium apt-cacher-ng might be all I need
[20:10] <jdstrand> if it indeed will use it
[20:40] <jdstrand> SpamapS: I'm sorry. I feel like I am doing something wrong. I can't seem to get my juju/lxc out of state 'pending'
[20:41] <jdstrand> SpamapS: is there some juju/lxc docs I can be pointed at?
[20:41] <jdstrand> SpamapS: I don't mean to be a bother-- pointing me a docs or someone else to pester is fine
[20:41] <m_3_> jdstrand: it takes a long time for the first service to come up... it's installing an image in /var/cache/lxc
[20:42] <jdstrand> m_3_: how long is 'long'? how can I monitor its progress?
[20:42] <m_3_> if it's longer than 30mins, then there's a problem
[20:43] <m_3_> juju status, /var/log/juju/, lxc-ls, and `ls /var/lib/lxc`
[20:43] <m_3_> have you had it working before?
[20:43] <m_3_> or is this a virgin install?
[20:43] <jdstrand> no. first time is today
[20:44] <m_3_> ok, so things to check out... `groups | grep libvirt` shows something
[20:44] <m_3_> you;'ve rebooted since installing libvirt (juju lxc container)
[20:44] <jdstrand> well, I am not new to libvirt
[20:44] <m_3_> `virsh net-list --all` shows something
[20:45] <jdstrand> I did instalI did get through bootstrap ok
[20:45] <m_3_> that's another potential problem... :) ... do you have existing libvirt nets defined?
[20:45] <m_3_> ok, then we're probably just waiting for the initial lxc image to be cached
[20:46] <jdstrand> this is in a precise vm. I redefined the libvirt default network to use 192.168.123 like I normally do when testing libvirt
[20:46] <jdstrand> I am in the libvirtd group
[20:46] <hazmat> jdstrand, you can verify if the containers are running with lxc-ls if its listed twice its running
[20:46] <hazmat> jdstrand, can you pastebin the contents of $data-dir/units/master-customize.log
[20:46] <jdstrand> that bit I was not sure about (I am new to lxc)
[20:47] <jdstrand> is lxc-ls supposed to require sudo access?
[20:47] <hazmat> no
[20:47] <jdstrand> $ lxc-ls
[20:47] <jdstrand> jamie-local-0-template	jamie-local-nagios-0  jamie-local-wordpress-1
[20:47] <jdstrand> /usr/bin/lxc-ls: line 35: cd: /sys/fs/cgroup/cpuset///lxc: Permission denied
[20:47] <jdstrand> ls: cannot access jamie-local-wordpress-1: No such file or directory
[20:47] <jdstrand> ls: cannot access jamie-local-nagios-0: No such file or directory
[20:47] <hazmat> odd
[20:47] <jdstrand> so I think I may be missing another group
[20:48] <jdstrand> this vm doesn't have an admin group
[20:48] <jdstrand> (only 'sudo')
[20:48]  * jdstrand looks at lxc-ls
[20:48] <hazmat> jdstrand, that $data-dir/master-customize.log might have some useful debug information, $data-dir from environments.yaml for the local provider
[20:48] <ejat> hazmat: where to define for exposing the ports/services ?
[20:49] <m_3_> jdstrand: right, not /var/log/juju, but $data_dir/**/*.log
[20:49] <jdstrand> I don't seem to have a master-customize.log
[20:51] <jdstrand> ./jamie-local/units/wordpress-1/unit.log
[20:51] <jdstrand> ./jamie-local/units/wordpress-1/container.log
[20:53] <hazmat> jdstrand, it would be in jamie-local/units/master-customize.log
[20:54] <hazmat> ejat, invoke open-port /close-port in your hooks
[20:54] <jdstrand> hazmat: yeah, that file doesn't exist. sudo find . -name master-customize.log returns nothing
[20:55] <hazmat> jdstrand, strange indeed.. the container.log and unit.log might have useful info.. if the links there are active, then at least things are starting up
[20:55] <ejat> at any file ?
[20:56] <hazmat> ejat, from within any hook
[20:56] <_mup_> Bug #930404 was filed: timestamps in status <juju:New> < https://launchpad.net/bugs/930404 >
[20:57] <jdstrand> units/wordpress-1/container.log claims the container is running
[20:59] <jdstrand> I'm going to blow away the vm and try again
[20:59] <jdstrand> and make sure the vm is up to date precise
[21:09] <ejat> hazmat: wondering why juju asking my key passphrase ?
[21:10] <hazmat> ejat, it stuffed the public portion of the key onto the containers, if your connecting to a container via ssh, and the key has a password, ssh will prompt for it
[21:10] <hazmat> s/it/its
[21:10] <ejat> im just checking juju status
[21:11] <ejat> previous its not given me any prompt
[21:11] <hazmat> ejat, it shouldn't do that with local provider
[21:11] <ejat> ec2
[21:11] <hazmat> ah
[21:11] <hazmat> sorry mixing the streams
[21:12] <hazmat> ejat, yeah.. the ssh connection is nesc to allow the client to communicate with juju's db (zookeeper)
[21:13] <hazmat> ejat, you can specify a key without password via environments.yaml or use an ssh agent if you want a password protected key
[21:13] <ejat> noted .. me trying to destroy environment 1st then trying to redo .. btw .. i already put open-port 80/tcp in my file but its not exposing the service
[21:16] <hazmat> ejat, you also have to do juju expose <service-name>
[21:16] <ejat> already did .. but .. let me rerun the test so i can post the status
[21:18] <hazmat> ejat, which hook are you doing the open-port in?
[21:20] <ejat> since i referring to wordpress hook
[21:20] <ejat> so its in the db-relation-changed
[21:21] <ejat> hazmat: brb
[21:28] <jdstrand> ok, so it seems juju is changing the perms on /sys/fs/cgroup/cpuset///lxc
[21:28] <jdstrand> I started a new vm, made sure it was up to date, did the bootstrap and everything was ok
[21:29] <jdstrand> I then di juju deploy --repository=/usr/share/doc/juju/examples local:oneiric/wordpress
[21:29] <jdstrand> I immediately did 'lxc-ls' and it worked without error and returned jamie-local-0-template
[21:30] <jdstrand> later I ran it to see how my machine was coming up and I got:
[21:30] <jdstrand> $ lxc-ls
[21:30] <jdstrand> jamie-local-0-template	jamie-local-wordpress-0
[21:30] <jdstrand> /usr/bin/lxc-ls: line 35: cd: /sys/fs/cgroup/cpuset///lxc: Permission denied
[21:30] <jdstrand> (notice jamie-local-wordpress-0 is there now)
[21:30] <jdstrand> the permissions on /sys/fs/cgroup/cpuset///lxc are 700 now
[21:31] <jdstrand> this is with up to date precise
[21:31] <jdstrand> I have a master-customize.log now
[21:32] <jdstrand> juju status tells me that wordpress/0 is still pending
[21:33] <jdstrand> and I see the problem:
[21:33] <jdstrand> Err http://archive.ubuntu.com/ubuntu/ oneiric-updates/main libfreetype6 amd64 2.4.4-2ubuntu1.1
[21:33] <jdstrand>   Could not connect to 192.168.122.1:3142 (192.168.122.1). - connect (111: Connection refused)
[21:34] <jdstrand> that should be 192.168.123.1 in my configuration
[21:35]  * jdstrand wonders where to adjust that
[21:36] <jdstrand> providers/local/network.py looks promising
[21:37] <jdstrand> looks like the 122 could be changed...
[21:39] <jdstrand> ok, if I destory the service lxc-ls works again
[21:40] <jdstrand> but /sys/fs/cgroup/cpuset///lxc is still 700
[21:45] <ejat> hazmat: http://paste.ubuntu.com/837113/
[21:47] <ejat> state : started but open port null
[21:50] <hazmat> ejat, i'd check the unit log on the symphony unit, and verify the hook output containing open-port
[21:51] <hazmat> ejat, the relation doesn't exist, so the db-relation-changed hook hasn't been executed
[21:52] <ejat> owh ..
[21:52] <ejat> so i need to remove the db-relation ? then where is suggested to put the open-port /
[21:53] <ejat> for now .. i dont want to make a relation to db yet .. (next step)
[21:53]  * ejat since im still crawling making the charm
[21:55] <hazmat> ejat, that's a relation hook, its only going to be called when the relation is established.. if you always want the same port for the unit you can just do it install in start or even install.. alternatively in your current environment.. you can deploy a database and add a relation between it and symphony so the db-relation-changed hook is executed, and the port is opened
[21:55] <hazmat> port opening is a two step process, the user has to have had exposed the service, and a charm has to have executed a hook that calls open-port
[21:56] <hazmat> s/charm/service unit
[21:56] <ejat> so can i put the open-port in install ?
[22:02] <jdstrand> hazmat: how do I specify that my container and everything should be using the 192.168.123 network that I configured with libvirt?
[22:02] <jimbaker> ejat, if that makes sense for your charm, sure
[22:03] <ejat> jimbaker: ok .. thanks . ill try to do that way 1st and see the outcome
[22:03] <jdstrand> I'm hoping it is something in .juju/environments.yaml
[22:05] <m_3_> jdstrand: I think virbr0 is hard-coded... didn't think the 122 was though, I thought it picked it up from the default network
[22:05] <jdstrand> I don't mind the virbr0 bit
[22:05] <jdstrand> $ virsh net-dumpxml default|grep 192
[22:05] <jdstrand>   <ip address='192.168.123.1' netmask='255.255.255.0'>
[22:05] <jdstrand>       <range start='192.168.123.2' end='192.168.123.254' />
[22:06] <m_3_> is default using virbr0?
[22:06] <jdstrand> but if I use sudo lxc-netstat --name jamie-local-wordpress-0 -ie I can see it has 192.168.122.185
[22:06] <jdstrand> oh, virbr0 does have 192.168.123.1
[22:06] <jdstrand> your right
[22:06] <jdstrand> (that was eth0 that had 192.168.122.185
[22:06] <jdstrand> )
[22:07] <m_3_> so your default's on virbr0, active, and rewritten to use 123?  hmmmmm
[22:08] <m_3_> dang, that should totally be working
[22:08] <jdstrand> the host's ip is 192.168.122.185
[22:08]  * m_3_ vaguely remembers something about lxcbr0 in precise
[22:08] <jdstrand> lxcbr0 is there and has 10.0.3.1
[22:08] <jdstrand> hmm, so maybe it is something else
[22:08] <m_3_> but I tdon't think anything's listening
[22:08] <m_3_> dnsmasq is up for 123
[22:09] <jdstrand> yes
[22:09] <jdstrand> libvirt's dnsmasq is 123
[22:09] <m_3_> grrrr
[22:09] <jdstrand> lxc's is 10.0.3
[22:09] <jdstrand> well, remember, I did reconfigure my default network
[22:10] <jdstrand> I am running juju in a libvirt vm
[22:10] <m_3_> lemme spin up a precise vm and playing with rewriting default
[22:10] <jdstrand> so my 'host' in this case is a vm that is trying to spin up lxc guests
[22:11] <m_3_> right... we do this all the time, but in ec2
[22:11] <jdstrand> ok
[22:11] <m_3_> so the 122's free :)
[22:11] <jdstrand> I was trying to save the company some mony and use lxc :)
[22:12] <jdstrand> m_3_: fyi, I am using the steps I outlined in bug 930430 (which hallyn asked me to file)
[22:12] <_mup_> Bug #930430: lxc-ls requires root access after deploying an LXC instance <juju (Ubuntu):New> <lxc (Ubuntu):New> < https://launchpad.net/bugs/930430 >
[22:12] <m_3_> `find /usr/share/pyshared/juju -name "*.py" | xargs grep 122` => <sad_face>
[22:13] <m_3_> jdstrand: ok, thanks for the ref
[22:13] <jdstrand> m_3_: bummer. I was hoping I could do something like:
[22:13] <jdstrand> environments:
[22:13] <jdstrand>   local:
[22:13] <jdstrand>     type: local
[22:13] <jdstrand>     ...
[22:13] <jdstrand>     subnet: 123
[22:13] <jdstrand> or something
[22:13] <jdstrand> (in the yaml)
[22:14] <m_3_> I filed a bug for juju to create and use its own libvirt network... so it wouldn't conflict with existing ones
[22:14] <jdstrand> I thought I saw '131' floating around in there
[22:15] <m_3_> but I like the idea of it being an environments.yaml setting better
[22:15] <hazmat> jdstrand, juju will use the 'default' libvirt network
[22:15] <m_3_> really 'subnet' should be probably the virsh network name?
[22:16] <m_3_> hazmat: but he's changed 'default' to 123
[22:16] <hazmat> m_3_, shouldn't matter
[22:16] <hazmat> m_3_, we query the attributes/netinfo out, and just define the default if it doesn't exist
[22:17] <m_3_> right... his default sounds like it's up and working correctly... we're not picking it up tho... it's catching onto a 122 net first as I understand
[22:18] <jdstrand> let me get apt-cacher-ng out of the equation (it was recommended to be installed)
[22:21] <hazmat> m_3_, hm.. yeah
[22:21] <jdstrand> ok, so the image did pick up a 192.168.123 address
[22:21] <jdstrand> 192.168.123.67
[22:21] <jdstrand> (seen in syslog)
[22:21] <hazmat> jdstrand, can you ssh into it?
[22:22] <jdstrand> I tried and a get a prompt
[22:22] <hazmat> i'm curious if its resolv.conf is wrong
[22:22] <jdstrand> so that is good, but I am not automatically logged in
[22:22] <jdstrand> is there a password I should use?
[22:22] <hazmat> jdstrand, it should have setup key auth
[22:22]  * jdstrand wonders why each step forward was pain...
[22:23] <jdstrand> that is what I thought
[22:23] <jdstrand> I specifically did 'ssh-keygen -t rsa -b 2048'
[22:23] <hazmat> jdstrand, do you have a master-customize.log in your $data-dir/units directory?
[22:23] <jdstrand> hmm
[22:23] <hazmat> this time
[22:23] <jdstrand> .ssh directory has wrong perms
[22:24] <jdstrand> fixed perms, ssh -i ./.ssh/id_rsa 192.168.123.67, still password
[22:24] <hazmat> m_3_,  jdstrand, it is  hardcoded to 192.168.122.1 for the nameserver config, thats a bug
[22:25] <m_3_> jdstrand: use ubuntu@192...
[22:25] <jdstrand> durr
[22:25] <jdstrand> m_3_: thanks!
[22:25] <m_3_> hazmat: good catch
[22:25] <jdstrand> I made it in :)
[22:25] <hazmat> same for the apt-rpoxy
[22:26] <jdstrand> actually, the resolv.conf picked up nameserver 192.168.123.1
[22:26] <m_3_> local provider can get that from SSH_CLIENT maybe?
[22:27] <jdstrand> ok, so the machine is up. state is still pending and even though I exposed the service, public-address is null and no open-ports
[22:27] <jdstrand> (this is just workpress with no relations)
[22:27] <jdstrand> wordpress
[22:28] <jdstrand> juju ssh wordpress/1
[22:28] <jdstrand> 2012-02-10 16:27:55,163 INFO Waiting for unit to come up.
[22:28] <hazmat> m_3_, i wonder if its because we're installing getting nested libvirt installations because of the recommends
[22:28] <jdstrand> I can ping www.ubuntu.com from the guest
[22:28] <hazmat> jdstrand, cool
[22:28] <m_3_> hmmm
[22:29] <jdstrand> but the 'juju' commands don't seem to be working. maybe I haven't it right
[22:29] <m_3_> we definitely need to do the no-install-recommends and at least take that moving part away
[22:29]  * jdstrand tries to read up on that
[22:30] <m_3_> jdstrand: `ip addr show | grep virbr` on the instance
[22:31] <jdstrand> m_3_: in the lxc guest? nothing
[22:31] <m_3_> ok, thanks
[22:31] <jdstrand> on the kvm host running the lxc guest, I can give you that (it is 192.168.123)
[22:32] <m_3_> jdstrand: right... just checking that we weren't installing another layer of libvirt on each instance... we were doing that for a while
[22:32] <jdstrand> oh
[22:33] <jdstrand> no libvirt in the lxc instance
[22:35] <m_3_> ec2 got... um... sluggish in the last few minutes
[22:35] <m_3_> netflix spinning up on the east-coast
[22:36] <m_3_> and hulu.. and <insert fav stream svc>
[22:48] <ejat> jimbaker: i try my install script in local machine manually on precise is working fine
[22:48] <ejat> but when i try on oneiric
[22:48] <ejat> something happen that require updating the instances then its work
[22:48] <ejat> how ?
[22:49] <ejat> need to put the update & upgrade into the install ?
[22:51] <ejat> hazmat & jimbaker : is it ok if my charm to have this result : http://ec2-175-41-185-95.ap-southeast-1.compute.amazonaws.com/symfony/web/
[22:51] <ejat> then ill try to work out the db-relation
[22:56] <ejat> since its a php framework
[23:11] <_mup_> juju/purge-queued-hooks r456 committed by jim.baker@canonical.com
[23:11] <_mup_> Added purge queued hooks test
[23:12] <jimbaker> ejat, i'm not familiar with this charm, but it is exposed :)
[23:12] <ejat> yeah .. finally ..
[23:13]  * ejat need to some additional to my install hooks .. since its differ deploying through juju and manual install at the instance
[23:13] <ejat> to do*
[23:14]  * ejat wondering ... 
[23:15] <jimbaker> ejat, in general you probably don't want to open-port until the service is ready. so you shouldn't see an intermediate state like the current page. of course, it's good for debugging right now
[23:16] <ejat> yups .. agreed ..
[23:19] <ejat> its more toward file n folder permission
[23:19]  * ejat hmm ... 
[23:33] <TheTinyToon> Hi everyone - I've started to test an ubuntu private cloud with some local virtualbox instances. I've created an orchestra server and two nodes. However, upon running "juju status", I get an error connecting to the environment. --verbose shows, that orchestra tries to connect to port 2181 on the (possible?) zookeeper node, but netstat does not show anything behind that port. Any hints?
[23:34] <TheTinyToon> I've found this IRC log of this channel, where jorge seems to have the same problem: http://irclogs.ubuntu.com/2012/01/26/%23juju.txt
[23:35] <TheTinyToon> however, destroying the environment did not help me, as did installing the zookeeper-package.