=== Spads_ is now known as Spads
=== iceyec_ is now known as iceyec
=== scuttlemonkey is now known as scuttle|afk
tasdomaswhat is the purpose of amulet.raise_status(amulet.PASS) ?07:49
Odd_BlokeHow large a disk does the bootstrap node need?11:05
Odd_BlokeWould 10GB be sufficient?11:05
marcoceppiOdd_Bloke: that should be sufficent11:31
Odd_Blokemarcoceppi: Great, thanks.11:31
marcoceppitasdomas: that more or less exits the test and prints a PASS line11:31
tasdomasmarcoceppi, I'm asking because if I did that, the test was marked as failed11:32
marcoceppirofl, I don't think that's the intention of that command11:32
=== natefinch is now known as natefinch-afk
Odd_Blokemarcoceppi: If you have a few minutes: https://code.launchpad.net/~daniel-thewatkins/charm-helpers/lp1370053/+merge/26086412:25
marcoceppiOdd_Bloke: ack, taking a look12:26
Odd_Blokemarcoceppi: Thanks muchly!12:31
Odd_BlokeThat should let me sort out issues in the ubuntu-repository-cache merge proposals.12:31
Odd_BlokeThen you can merge those. ;)12:31
tasdomasmarcoceppi, bundletester is a great tool!12:36
marcoceppitasdomas: it is very nice, we hope to have it be what `charm test` and `juju test` run in the near future12:37
tasdomasmarcoceppi, that would be nice - charm test fails for me (using local provider)12:39
coreycbjamespage, hey, can you take a look at my response to your comment here?  A charm-helpers fix is needed for upgrade-charm:  https://code.launchpad.net/~corey.bryant/charms/trusty/quantum-gateway/end-of-life/+merge/26503512:46
=== natefinch-afk is now known as natefinch
beisnermarcoceppi, it appears that amulet.SKIP is unexpectedly invoking --set-e behavior.  end result is a failed test run instead of a single skipped test.  I don't generally use amulet.SKIP or amulet.PASS explicitly, but dosaboy has a percona-cluster amulet test in flight which does, and that's where we are seeing this.13:12
beisnermarcoceppi, according juju test -h this ^ would appear to not be the expected behavior.13:12
beisnermarcoceppi, i've not dug into the code, but could it be that any amulet.raise use takes us down a fail path (given tasdomas is also seeing amulet.PASS raises cause a fail)?13:14
tasdomasbeisner, that's what I saw - I just assumed I was using it wron13:17
=== scuttle|afk is now known as scuttlemonkey
g3narook,, is there a charm that can build and configre a docker-registry server ?13:42
g3naroor should i somehow package a puppet class that does setup for this13:42
mbruzekg3naro: I am not aware of a charm that sets up a docker-registry server.  That is a good candidate for a new charm *nudge* *nudge*14:08
mbruzekg3naro: charms can be written in puppet or chef, or any thing that runs on Ubuntu14:08
mbruzekalso centos14:09
mbruzekg3naro: If you want to search for charms use this url: https://jujucharms.com/14:10
mbruzekIt searches charms that are in a personal namespace and ones that are in the recommended section of the charm store.14:11
g3naroahh ok great!14:17
marcoceppibeisner: it shouldn't PASS -> FAIL is determined by the test runner14:37
marcoceppiPASS exits with code 100, juju test has a flag to determine how to handle SKIP (as OK, or problem)14:38
beisnermarcoceppi, raising a SKIP or PAAS causes juju test to exit non-zero.14:38
marcoceppinot sure how bundletester handles this14:38
beisnerPASS even14:38
marcoceppibeisner: well that's not the expected outcome14:38
marcoceppiand is the first I've heard of this. Do you have an example of the test/cahrm causing this so I can repro?14:39
beisnerdosaboy, ^14:39
marcoceppistub: ping14:39
marcoceppistub: actually, going to PM14:39
dosaboymarcoceppi: https://code.launchpad.net/~hopem/charms/trusty/percona-cluster/min-cluster-size/+merge/26550214:40
dosaboytry running 'make test'14:40
dosaboyyou'll see that most tests skip since they require a vip14:40
dosaboyyet they all get marked as fail14:40
dosaboyall the ones that skip that is14:40
beisnerdosaboy, even though this may be unexpected behavior, i would recommend that we allow the test to fail if it's not fed everything it needs (and not even try to SKIP it).14:45
beisnerdosaboy, otherwise a binary pass/fail in ci for a merge proposal may not always mean the same thing.14:45
marcoceppidosaboy beisner do you have output? I'm getting skips as expected14:47
marcoceppiwhat version of charm-tools are you using?14:47
beisnermarcoceppi, one diff:   use --set-e14:47
beisnermarcoceppi, uosci runner injects that into the makefile so that if something fails, we have an environment to pull logs from14:47
dosaboybeisner, marcoceppi: i don;'t use --set-e and i still get "failed: 5" which is all those that skip14:48
marcoceppibleh, I hate how half baked this tool is, it should really just always get logs14:48
beisnerotherwise if a test fails, the juju environment is torn down before juju test exits, and we don't really have data to inspect.14:48
dosaboycharm-tools 1.5.1-0ubuntu1~ubuntu14.04.1~ppa114:48
marcoceppithis is stupid logic doing stupid things14:51
marcoceppiI'll open a bug for juju test and get a patch release out14:51
* beisner coffees..14:52
marcoceppiso, I found the problem with skip15:18
apuimedombruzek: ping15:24
apuimedombruzek: how are you?15:24
mbruzekI am well.  How are things with you?15:25
apuimedombruzek: not too bad15:26
apuimedoI did the change you suggested15:26
mbruzekgreat I will review it today15:26
apuimedombruzek: should I change it to needs review?15:28
mbruzekthank you for the followup15:29
marcoceppiapuimedo: any reason why you're using the precise version of cassandra and not the new trusty version?15:43
apuimedomarcoceppi: the trusty version was not there when I satarted with this15:44
apuimedoI guess I should move to it15:44
apuimedoand check if it can do what I need to15:44
marcoceppiapuimedo: it's available now, you should just be able to deploy the trusty version instead15:44
marcoceppiit's a lot smarter about how to handle clustering and versions of cassandra15:45
apuimedoI'm very glad to hear that15:45
apuimedomarcoceppi: I can't see about choosing a specific version15:46
marcoceppiapuimedo: how specific do you need? this supports more generic options like datastax vs apache and has support for apache cassandra 2.0 and 2.1 atm15:46
apuimedoI need to install 2.0.115:47
apuimedofrom datastax15:47
marcoceppiyou mean datastax enterprise or just the cassandra 2.0.1?15:47
apuimedo        #   - deb http://debian.datastax.com/community stable main15:48
marcoceppiapuimedo: you can set that in the config.yaml15:50
apuimedomarcoceppi: yes, but I need an older package from that source15:51
apuimedoI guess I would have to patch this new cassandra to allow you to specify a specific version15:51
marcoceppiapuimedo: I see15:51
marcoceppiapuimedo: well, you can continue to use the precise version, don't get me wrong15:51
marcoceppijust wanted to make sure you were aware of teh trusty version15:52
apuimedoI would like to use the trusty15:52
apuimedoso probably what we can do is15:52
apuimedowe get this merged for precise15:53
apuimedoand then for the next release I patch the new cassandra charm to accept version15:53
apuimedoand move our bundle to use it15:53
marcoceppithanks for the all the work so far!15:53
mbruzekapuimedo: What package_version can I set the cassandra charm to?16:17
apuimedoI'll give you the config16:18
apuimedombruzek: http://paste.ubuntu.com/11920807/16:19
g3narowe come in peeeeace16:19
coreycbjamespage, gnuoy: can one of you review this?  it's needed by neutron-gateway in order to allow upgrading from quantum-gateway to neutron-gateway.  https://code.launchpad.net/~corey.bryant/charm-helpers/upgrade-charm/+merge/26545017:15
pmatulisare log levels for debug-log and juju-log the same? TRACE, DEBUG, INFO, WARNING, ERROR19:19
=== ddellav_ is now known as ddellav
marcoceppipmatulis: probably19:51
sebas5384lazyPower: ping23:22

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