[00:01] <arosales> winael, thanks for being a champion :-)
[00:02] <arosales> keep posting your feedback here and to the list. It sounds like utlemming updated the docs too
[00:02] <arosales> utlemming, was that the juju.u.c/docs ?
[00:02] <utlemming> arosales: no, it was my blog post
[00:02] <mxc> everything is in the same region
[00:03] <arosales> mxc, looks like a post @ juju.provider.azure environ.go:406 picked tools is failing .  . . I am still investigating . . .
[00:03] <arosales> mxc, ack and if it were different you would get a slightly different error
[00:03] <arosales> utlemming, does it makes sense for the juju.u.c/docs?
[00:07] <arosales> mxc and to confirm in your .juju/environments.yaml file you only have "admin-secret, management-subscription-id, management-certificate-path, and management-certificate-path " set
[00:08] <mxc> yes
[00:08] <mxc> i also have the and location set
[00:09] <mxc> and storage-account-name
[00:09] <arosales> ah yes my buffer missed the last ones
[00:10] <mxc> are there any restrictions on the admin secret?
[00:10] <mxc> and should that match anything?
[00:11] <arosales> mxc, not that I know of.  You could set it or have "juju generate-config --show" make one for you
[00:11] <winael> utlemming: I to the sshuttle thing but now to connect my service that Juju knows as 10.0.3.23 what can I do ?
[00:13] <mxc> arosales: what version of ubuntu are you using?
[00:14] <utlemming> winael: I don't follow the question....can you clarify?
[00:14] <arosales> mxc, for the client or server?
[00:14] <mxc> the machine that I run
[00:14] <mxc> "juju bootstrap" on
[00:15] <arosales> mxc, 13.04
[00:15] <mxc> i've been trying 1310
[00:15] <mxc> would 1204 or 1304 be a better option?
[00:15] <winael> I use the sshuttle command as it's written on your post, and now, I just want to see my mediawiki in my browser
[00:15] <arosales> you should be ok at 13.10
[00:16] <arosales> in this case an API call to Azure is failing
[00:16] <mxc> yes
[00:16] <winael> As I understand I opened a ssh tunel for 10.0.3.0/24 network
[00:16] <arosales> so your install looks ok, and you are at the latest
[00:16] <mxc> is there anyway to get a stack trace of the error or having juju log every http request without recompiling
[00:17] <utlemming> winael: are you using the browser?
[00:17] <winael> my host machine browser yes
[00:17] <utlemming> winael: if so, it should show you the IP address. To access it, just put the 10.0.3.xxx in your brower and it should should up
[00:18] <arosales> mxc, without specifically inserting probe points in the code I am uncertain although we could ping the devs in #juju-core to confirm if there is an easy way
[00:19] <winael> no just a chromium page saying that no data was received
[00:23] <mxc> juju-core is empty
[00:24] <winael> Well doesn't matter I'll take a look later. I have to sleep I have to get up in 4 hrs...
[00:26] <arosales> winael, post to the list and we can try to get back there too
[00:26] <arosales> mxc, sorry I didn't grok juju-core is empty
[00:27] <winael> I'll try to do it tomorrow morning
[00:29] <mxc> np
[00:29] <arosales> mxc, were you trying to compile juju core?
[00:29] <mxc> trying hard not to
[00:39] <mxc> i gotta run
[00:39] <mxc> thanks for your help
[00:39] <mxc> the AU question is here: http://askubuntu.com/questions/388419/juju-bootstrap-fails-in-azure-badrequest-the-affinity-group-name-is-empty-or
[00:54] <arosales> I'll catch you on the list mxc.
[00:55] <arosales> I wasn't able to reproduce from a stand alone instance in Azure either for other folks following . . .
[00:55] <arosales> perhaps it is cert issue . .
[00:55]  * arosales will follow up later
[11:54] <ashipika> hi all..
[11:54] <ashipika> ERROR missing CA cert or auth-key
[11:54] <ashipika> where should i start looking?
[11:54] <ashipika> where should these two things be defined?
[12:20] <Luca> ls -latr
[12:28] <ashipika> @luca: where?
[12:29] <Luca> ashipika: focus was in the wrong window :)
[12:29] <ashipika> sorry :)
[12:29] <Luca> np :)
[13:05] <marcoceppi> ashipika: is this the machine you originally bootstrapped the environment with?
[13:06] <ashipika> marcoceppi: no worries. i found the error.. mistakenly named null.jenv as null.yaml..  totally my mistake
[13:06] <marcoceppi> ashipika: you shouldn't need to touch the .jenv files at all
[13:06] <ashipika> well.. in my case..
[13:20] <ashipika> ERROR: cannot put charm: ... x509: certificate has expired or is not yet valid.. does that imply that mongodb certificate on the bootstrapped host is not valid?
[16:20] <X-warrior> What am I supposed to do when a service is being listed on juju status, but the machine that the service was, is destroyed? :S
[16:24] <X-warrior> What am I supposed to do when a service is being listed on juju status, but the machine that the service was, is destroyed? :S
[16:31] <X-warrior> marcoceppi:
[16:40] <marcoceppi> X-warrior: juju destroy service?
[16:40] <marcoceppi> X-warrior: was the machine removed outside of juju?
[16:41] <X-warrior> marcoceppi: nope
[16:42] <X-warrior> I used, destroy-service and then it showed up as "    life: dying"... I wait for a while... and nothing changes... then I used juju destroy-machine ... and the machine was destroyed... but the service is still showing as  "life: dying "
[16:42] <X-warrior> :S
[16:48] <marcoceppi> X-warrior: what version?
[16:48] <marcoceppi> Someone had a similar issue
[16:48] <X-warrior> marcoceppi: 1.16.3
[16:48] <marcoceppi> X-warrior: 1.16.4 was released, but I don't think it addresses this issue.
[16:49] <X-warrior> uhmm
[16:50] <X-warrior> is it possible to deploy a charm with another name?
[16:51] <marcoceppi> X-warrior: yes
[16:51] <X-warrior> I mean I'm receiving "2013-12-10 16:49:51 ERROR juju supercommand.go:282 cannot add service "logstash-indexer": service already exists"
[16:51] <marcoceppi> juju deploy <charm> <alias>
[16:51] <marcoceppi> X-warrior: can you pastebin your entire juju status?
[16:51] <marcoceppi> I might know what's going on
[16:52] <X-warrior> marcoceppi: http://pastebin.com/g74FKzdK
[16:53] <marcoceppi> X-warrior: can you `juju remove-relation logstash-indexer elasticsearch`
[16:53] <X-warrior> I will try again, but I already tried that
[16:53] <marcoceppi> oh :\
[16:56] <X-warrior> marcoceppi: same result : (
[16:57] <X-warrior> btw, using juju deploy --repository=charms/ --constraints="mem=1G cpu-power=1 arch=amd64" local:precise/logstash-indexer ALIAS gives me error: invalid service name "ALIAS"
[17:01] <marcoceppi> X-warrior: don't use all caps?
[17:02] <X-warrior> oh it can't have - or numbers on it I guess
[17:02] <X-warrior> I updated the name and it worked
[17:02] <X-warrior> x)
[17:46] <X-warrior> what is the folder on machine that has the charm files?
[17:51] <X-warrior> found it ty
[17:51] <X-warrior> ;D
[17:52] <webbrandon> Is this critical?  I am running "juju set amazon" and getting :ERROR Get : 301 response missing Location header.
[17:52] <webbrandon> oh yeah its critical
[17:52] <webbrandon> cant bootstrap
[17:54] <webbrandon> what in the heck did I do?? I didn't touch anything to the configuration!!! garr.. breathe.... now fix....
[18:04] <jcastro> http://askubuntu.com/questions/369231/juju-mysql-adding-units-vs-adding-new-service-with-relation
[18:04] <jcastro> I have a bounty on this question if someone wants 200 points!
[18:09] <webbrandon> im getting a bug issue that appeared a year ago https://bugs.launchpad.net/juju-core/+bug/1083017
[18:09] <_mup_> Bug #1083017: Cannot bootstrap with public-tools in non us-east-1 region <ec2> <juju-core:Fix Released by dave-cheney> <https://launchpad.net/bugs/1083017>
[18:10] <webbrandon> I am getting when trying to read the environments.yaml file it seems
[18:11] <jcastro> marcoceppi, hey so what was the tldr on that IS rewrite of the squid charm?
[18:11] <mgz> webbrandon: can you get the actual response body, so we see the url that's failing and why?
[18:16] <webbrandon> mgz: http://paste.ubuntu.com/6552412/
[18:18] <mgz> webbrandon: yeah, you need a better tool, as our logging there seems to suck
[18:18] <mgz> wireshark maybe?
[18:39] <webbrandon> mgz: okay, I installed it.  Not familiar with this,  I assume I link a particular item up to the pipe?
[18:43] <mgz> webbrandon: you should just be able to capture the interface, if there's not too much else going on, then rerun `juju set`
[18:48] <marcoceppi> jcastro: no idea?
[18:48] <jcastro> well it was in review last week
[18:48] <jcastro> that's the one that was essential a rewrite iirc?
[18:49] <marcoceppi> jcastro: let me check
[18:49] <marcoceppi> I can only keep track of so many charms ;)
[18:49] <jcastro> yeah
[18:49] <jcastro> this one is just weird because if they're using the one not in store in production then we need to either merge their work, or drop ours and promote theirs
[18:49] <jcastro> but having 2 is not good
[18:49] <webbrandon> mgz: looks like I can't will keep looking into using it but maybe there is a bigger issue
[18:50] <marcoceppi> jcastro: was it squid?
[18:50] <marcoceppi> I don't see any merges for it recently
[18:51] <jcastro> squid reverse proxy
[18:51] <mgz> webbrandon: did you elevate?
[18:53] <webbrandon> mgz:  Doc said it was supposed to be ran as root if installed though package manager but it wasnt.  started as root user. CAn now see interfaces
[18:53] <mgz> webbrandon: or see /usr/share/doc/wireshark-common/README.Debian it seems
[18:53] <mgz> but root is fine(ish) for a one-off
[18:54] <marcoceppi> jcastro: I guess it was merged, the last merge was sidnei's branch in early November, the other two pending are < 20 lines of code
[18:55] <jcastro> yeah but you would have remembered merging it wouldn't you have?
[18:55] <jcastro> it was like 6000 lines
[18:56] <sidnei> jcastro: what's the question?
[18:57] <jcastro> so the IS guys are apparently using a forked version of the squid reverseproxy charm
[18:58] <jcastro> and they had a MP a long long time ago that was essentially a rewrite
[18:58] <jcastro> but I don't think it ever got merged
[18:59] <jcastro> it looks like the main feature they want is SSL
[19:02] <sidnei> jcastro: i think it's merged yes, i pushed most of it.
[19:03] <jcastro> kirkland, ok so I think the issue here is probably why they can't get SSL working
[19:06] <sidnei> jcastro: pushed the latest from webops into https://code.launchpad.net/~sidnei/charms/precise/squid-reverseproxy/trunk/+merge/198451
[19:06] <sidnei> it's only a few small changes
[19:06] <jcastro> k
[19:07] <jcastro> sidnei, ok so we're not out of sync, they might just be running an older version of the charm?
[19:07] <sidnei> jcastro: they being? sorry im missing some context.
[19:07] <jcastro> I'll send you an email
[19:08] <sidnei> k
[19:11] <webbrandon> mgz: Is this what you were looking for http://paste.ubuntu.com/6552655/ ?  That is the result of running "juju set amazon"
[19:13] <webbrandon> paste bin is acting funny right now.  you need to give it a moment or it will look like there are 8k of blank lines
[19:18] <mgz> webbrandon: urgh, I'm silly, https means this approach isn't enough. will have to add better logging to a juju binary for you
[19:21] <webbrandon> mgz: not sure I follow.  What are the https issues I have?
[19:22] <mgz> I can't see the response that gives the error, because of course it's encrypted :)
[19:23] <webbrandon> ahhh.  I see
[19:24] <webbrandon> this all happened after I started and stoped a bootstrap.  Did nothing to the setup or system between bootstrap events
[19:25] <webbrandon> why can't scratching my head fix it
[19:25] <mgz> it may well be a weird config thing
[19:27] <webbrandon> I can I make sure all of juju is removed and try a reinstall?  I tried this once but something muct be left behind
[19:28] <webbrandon> how can I make sure
[19:28] <webbrandon> ?
[19:29] <mgz> webbrandon: I doubt it's juju itself
[19:29] <webbrandon> rebooting
[19:30] <mgz> you may want to try moving your ~/.juju dir out of the way and recreating your configuration though
[19:30] <webbrandon> I changed nothing though between a working bootstrap.
[19:30] <webbrandon> mgz: tried that already
[19:30] <mgz> it could also be not your end
[19:31] <webbrandon> I doubt my rsa need to be reset but gona try a full reinstall and new rsa
[19:52] <mxc> quick question, if I want to deploy a charm from a local repository, does that repository have to be on the machine I'm running the juju commands from or the "machine 0" that is managing the juju cluster?
[19:56] <webbrandon> okay nothing worked. gonna submit bug
[19:56] <mgz> webbrandon: thanks
[19:57] <mgz> please mention region/other provider config that's not standard
[20:38] <webbrandon> yeah fixed
[20:52] <lazypower> mxc: did you get an answer to your charm deployment question?
[20:53] <mxc> yeah
[20:54] <lazypower> mxc: the way I understand it is, the repository has to be on the machine you're executing the juju commands from - this triggers a sequence of caching on Machine0 - which caches all the charms in use in your juju landscape. Sorry if this is a duplicate answer or incorrect, but its how I understand it to work.
[20:54] <lazypower> ah ok. my bouncer must have missed the answer in its cache. Glad you got it sorted *hattip*
[20:59] <mxc> the problem was that my local charm wasn't in a precise directory
[21:09] <lazypower> aahhh ok.
[21:09] <lazypower> yeah, release directories are important.
[21:09] <lazypower> I found that out the hard way too
[22:27] <aquarius> marcoceppi, ping about postgresql backups in discourse
[22:27] <marcoceppi> aquarius: ack
[22:27] <aquarius> wait, ignore me
[22:27] <aquarius> ha!
[22:28] <aquarius> postgres is backing up
[22:28] <aquarius> on the postges machine
[22:28] <aquarius> which is not the discourse machine
[22:28] <aquarius> heh :)
[22:28] <marcoceppi> aquarius: all of these statements are correct!
[22:28] <aquarius> marcoceppi, the stuff in /var/lib/postgresql/backups/ is .dump files -- can I just scp them to another machine?
[22:29] <aquarius> they're not some sort of live backup?
[22:29] <marcoceppi> aquarius: you can scp them
[22:29] <marcoceppi> it's safe
[22:29] <marcoceppi> aquarius: they're snapshots
[22:30] <webbrandon> pretty sure I found a bug in juju-local.  Run it by here before I submit it.  I am getting with "juju status" after launching charm instance on machine 1: agent-state-info: '(error: container "mind-local-machine-1" is already created)'.  > The whole status at http://paste.ubuntu.com/6553558/
[22:31] <marcoceppi> webbrandon: are you on 1.16.4 ?
[22:31] <marcoceppi> webbrandon: nvm, you are!
[22:31] <webbrandon> yes
[22:31] <webbrandon> :)
[22:32] <marcoceppi> webbrandon: this looks the local environment wasn't cleaned up from last time
[22:32] <marcoceppi> webbrandon: what happens if you juju destroy-environment then reboostrap?
[22:32] <webbrandon> I went into var/lib/juju/containers and made sure everything was removed before bootstrapping
[22:33] <webbrandon> It does it again
[22:33] <webbrandon> marcoceppi: its like it is creating the mind-local-machine-1 for both my bootstrap and my charm
[22:36] <marcoceppi> webbrandon: that's a bug
[22:36] <webbrandon> figured
[22:36] <webbrandon> I will report it
[22:38] <webbrandon> hmm.  just need to find the juju-local branch
[22:38] <webbrandon> lol
[22:40] <marcoceppi> webbrandon: there is no juju-local
[22:40] <marcoceppi> report it against juju-core project
[22:43] <webbrandon> just figuring that out.  thank you marcoceppi. again.......
[22:55] <webbrandon> marcoceppi I am noticing it says its running 13.04 and I am on 13.10.  COuld this be my issue maybe?
[23:10] <webbrandon> So why was juju gernate-config created if juju init does the same?  Is it the extra freatures like --show?  Just trying to learn random info, no reason.
[23:32] <aquarius> OK. My postgres machine, deployed with juju, backs up things into /var/lib/postgresql/backups, which is great.
[23:32] <aquarius> I would like those backup files to also end up on another machine somewhere. How should I do that?
[23:34] <aquarius> I could, for example, generate a new ssh key on the postgres machine, for the postgres user, and then run "scp --newer /var/lib/postgresql/backups/* othermachine:/backups-from-juju" in a cron job. But then I've generated an ssh key *on that machine* -- if I scale postgres out with juju, or redeploy or something, then I won't have any backups.
[23:34] <aquarius> I could run that cron job on othermachine and scp *into* my juju machine to copy the backup files back, but I hardly know what the name of my juju machine *is*, because I don't have to
[23:35] <aquarius> how am I meant to set this sort of thing up?