=== spammy is now known as Guest80454
=== spammy is now known as Guest29965
=== frankban|afk is now known as frankban
kjackalGood morning juju world!07:51
juju_worldGood morning kjackal07:56
=== spammy is now known as Guest93444
=== Guest93444 is now known as spammy
tzondoes anybody knows hot to fix hook failed "update-status"10:48
tzonI cannot resolbe10:48
kjackalhi tzon10:52
kjackalyou have an update status that is failing, right?10:52
tzonkjackal, yeah it says hook failed "update-status" on nova-cloud-controller10:53
kjackalah, nova-cloud-controller not something I know anything about but I will try to help as much as I can10:55
kjackalso, lets see, you are not in a state where the unit is in an error state10:56
tzonyeah it is in error state10:57
kjackalwhen you do a juju resolved --retry <unit> you are left again with an error10:57
kjackalwhat do the logs say?10:57
kjackaljuju debig-log10:57
kjackalyou leave the logs running in a console and you fire a resolved --retry10:58
kjackalyou could filter errors of ony the failing unit with something like this: juju debug-log --include unit-mysql-011:00
tzonits gets me an error with the include11:08
tzonI used juju debug-log | grep nova-cloud11:09
tzonbut I did not get any results11:09
kjackalhm.... what kind of an include error?11:11
tzonsorry I dont get you11:12
tzonwith the resolved --retry it says that it is already resolved11:12
kjackalok so if it says it is already resolved then your unit should not be in an error state11:13
kjackalcloud you double check that11:13
tzonyeah I checked it again it is in error state11:15
tzonmaybe its a bug?11:15
kjackaldoesn't seem right. Could be a bug, but it is surprising...11:17
kjackalit is a rather basic "usecase": error state -> resolve --retry11:18
tzonyeah I have resolve similar issues in the past with this way but I have no idea whats going on now11:19
tzonalso I have another service that is in executing state and running update 2 days now11:20
tzonthis also is not normal11:20
kjackal2 days! Super strange I would expect things to expire after 2 hours or something11:22
kjackalSo do you know how to trigger a update-status hook?11:23
kjackaljuju run --unit namenode/1 'hooks/update-status'11:23
kjackalthis could save you some time11:23
kjackaltzon: ^11:24
tzonok I will give it a shot11:27
tzonit just suck when I running it :)11:33
kjackalCool that means that the hook runs OR that another hook is now running11:35
infinityplusbhi folks, after upgrading to 2.0-beta11 I have issues running any juju commands. Is there a way to "reset/reboot" juju?11:37
=== BlackDex_ is now known as BlackDex
tzonfinnaly I got error timed out :/12:06
babbageclunkinfinityplusb: do you have a bootstrapped controller?12:08
infinityplusb@babbageclunk: I do have a controller present when I do `juju list-controllers` but if I try and get details about the model with `juju models` it hangs12:10
infinityplusbjuju also hangs when I do `juju status` so I can't see what is happening12:11
babbageclunkinfinityplusb: What about when you run `juju status --debug`?12:12
infinityplusb@babbageclunk: ah, that gives me ... something. It seems there is something "amiss" with lxd. I don't seem to have permissions to the charms that are already deployed ...12:14
babbageclunkinfinityplusb: Want to put the output into a pastebin?12:15
=== spammy is now known as Guest78769
=== dpm_ is now known as dpm
infinityplusb@babbageclunk: http://pastebin.com/f5Gi51J312:18
infinityplusbwhich is odd, cause if I do a `groups` command, I can see I am in the lxd group12:20
babbageclunkinfinityplusb: ok, it seems like you can't connect to the container running the controller? Can you ssh to ubuntu@
infinityplusb@babbageclunk: via `juju ssh ...` no, but I can just via regular ssh12:23
babbageclunkinfinityplusb: hmm. Inside the container can you see jujud running?12:25
infinityplusb@babbageclunk: if I do a `service jujud status` it returns it as "inactive (dead)" ... probably not a good sign12:27
babbageclunkinfinityplusb: no, doesn't sound great!12:27
babbageclunkinfinityplusb: What can you see in /var/log/juju?12:27
infinityplusb@babbageclunk: many many errors - a lot similar to "juju.rpc server.go:576 error writing response: write tcp> write: connection reset by peer"12:31
infinityplusb@babbageclunk: and lots of "broken pipe" messages12:32
babbageclunkinfinityplusb: Is that on the controller host?12:32
infinityplusb@babbageclunk: yup, in the "machine-0.log"12:33
babbageclunkinfinityplusb: Maybe put it in a pastebin again? (This kind of stuff ends up using a lot of pastebins. ;)12:36
=== Guest78769 is now known as spammy
babbageclunkinfinityplusb_: Does the controller have anything deployed? Is the ip address it's trying to write to the host's?12:37
infinityplusb_@babbageclunk: it's like 300k lines long. I'll pastebin it somewhere :P And I *think* there is stuff deployed, but I can't do a `juju status` to see what is deployed where.12:37
babbageclunkinfinityplusb_: So was this bootstrapped with a previous beta of juju?12:38
infinityplusb_@babbageclunk: ... maybe. I (stupidly) didn't check if anything was up before updating12:40
babbageclunkinfinityplusb_: That might be part of the problem - I think we've had some backwards incompatible changes in the latest beta, could you try bootstrapping a new controller and see if you get the same issue?12:41
infinityplusb_@babbageclunk: if I try a new bootstrap, I get a permission erro about being in the lxd group (which I am).12:43
babbageclunkinfinityplusb_: That's weird. I don't really know much about lxd permissions - might be best to email the juju list? Sorry not to be too much help.12:44
infinityplusb_@babbageclunk: nah that's cool. I've learned some new things along the way. Thanks for trying. I'll keep digging. :)12:45
babbageclunkinfinityplusb_: ok, good luck!12:46
neiljerramMorning all!13:19
neiljerramIf I've written a new layer XYZ, how do I publish it, so that some other charm can say "includes: [ 'layer:XYZ' ]"13:21
kjackalhi neiljerram you could/should go and register your layer at http://interfaces.juju.solutions/13:28
neiljerramkjackal, I see, thanks.  What about during development?  Is that just a matter of putting the layer code under ${LAYER_PATH} ?13:30
kjackalneiljerram: yes putting your layer under {LAYER_PATH} will work13:31
neiljerramkjackal, Many thanks.13:31
kjackalneiljerram:  note that charm build will first look in you {LAYER_PATH} and then try to grab a layer from a remote repo. That means that your local copy will shadow anything else that might be outthere13:33
* D4RKS1D3 Hi14:59
=== Tristit1a is now known as Tristitia
=== frankban is now known as frankban|afk
=== spammy is now known as Guest70034
=== Guest70034 is now known as spammy
cargonzafyi: openstack irc meeting in #ubuntu-meeting this week. check out the details : https://wiki.ubuntu.com/ServerTeam/OpenStackCharmsMeeting17:44

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!