=== woutervb1 is now known as woutervb
thumperwallyworld: https://github.com/juju/juju/pull/1102701:32
thumperalmost gave myself a heart attack there01:37
thumperforgot that I had manually stopped the controller on two nodes01:37
thumperand was then wondering why they didn't upgrade01:37
thumperhpidcock: trade ya02:12
thumperhpidcock: there was a pending forward port from nammn_de102:12
hpidcockYeah it had outstanding conflicts02:12
thumperhpidcock: does this bring forward the same one?02:13
hpidcockI think anastasiamac looked at https://github.com/juju/juju/pull/11027 for you02:14
thumperoh cool02:14
thumperthanks anastasiamac02:14
hpidcockSo I guess I owe anastasiamac a PR now02:14
anastasiamachpidcock: from u, i'll take it with interest... 2 prs :D02:15
=== stub` is now known as stub
zeestratMorning! I have 2 charmhelpers PRs ready for review: https://github.com/juju/charm-helpers/pull/403 and https://github.com/juju/charm-helpers/pull/40408:37
nammn_de1hpidcock thumper yup,  while i was away some things got merged and ola, new conflicts ;/09:30
nammn_de1stickupkid: morning. ptal? https://github.com/CanonicalLtd/juju-qa-jenkins/pull/34809:31
stickupkidnammn_de1, lovely jubbly09:31
stickupkidmanadart, ping09:32
manadartstickupkid: Pong.09:32
stickupkidmanadart, quick ho, about my latest PR09:33
nammn_de1rick_h: can you oing me if you are around? got 2 things09:38
stickupkidnammn_de1, he's not online for another 3 hours09:39
nammn_de1stickupkid: i know just heads up, maybe to early to be doing a headsup09:40
nammn_de1stickupkid: see you are rewriting the firewallrules, how did you come across this?09:43
stickupkidnammn_de1, I refactored it as we don't use the firewall service directly, instead we use the firewallDoc inside a txn.Op10:04
nammn_de1stickupkid: roger10:04
=== salmankhan1 is now known as salmankhan
stickupkidCR anyone: I've had this locally for a few weeks, others might find it useful https://github.com/juju/juju/pull/1102911:15
nammn_de1stickupkid: done11:49
manadartstickupkid: This is the change I talked about in the HO. I have separated it from the big patch. https://github.com/juju/juju/pull/1103011:51
stickupkidmanadart, cool, let me check11:52
=== bdx0 is now known as bdx
hmlachilleasa: review please https://github.com/juju/juju/pull/1103113:19
achilleasahml: will do after standup if that's OK?13:20
hmlachilleasa:  np13:20
rick_hnammn_de1:  sure thing, hop into daily early or do post-daily?13:23
nammn_de1rick_h stickupkid: just lookiung through that I realized, we already do have a local controller cache. I suspect we could just remove the caching there.14:38
nammn_de1What do you think?14:38
nammn_de1The only advantage I can see is, if juju process itself hangs and we want to access the controllers14:38
rick_hnammn_de1:  if which process is hanging?14:43
nammn_de1rick_h: apologies, maybe this pr makes it more clear what I meant https://github.com/juju/juju/pull/1103215:16
rick_hnammn_de1:  looking15:27
nammn_de1rick_h: the bash completion is far from perfect, but this change should make it a little easier to use/less prone to show some errors because of the cache15:33
nammn_de1stickupkid: https://github.com/juju/juju/pull/11033 removing the not needed texts from set firewallrule15:33
nammn_de1rick_h: so testing for me was fine the whole day. But happy to get others testing. If we want to take a more conserative approach i would suggest just to change the error message to be empty  or do a simple retry: https://github.com/juju/juju/blob/develop/etc/bash_completion.d/juju#L273 because15:50
rick_hnammn_de1:  yea, just my own reservations as it's not something I play with much and has such an impact on filks15:51
nammn_de1rick_h: more conserative approach would be that one: https://pastebin.canonical.com/p/sMWVC4S2gC/16:09
nammn_de1it fails because it could not access the created cache in time and fill it. Kind of racy16:09
rick_hnammn_de1:  yea, that was my original thought was that if you hit a cache miss you prime the cache kind of thing but understand where you're going.16:09
nammn_de1rick_h: was my initial thought as well, just thought that we already had one and thus could just remove it16:10
nammn_de1but iam happy to propose this patch16:10
rick_hnammn_de1:  my main concern is that IS has complained to thumper about juju tab completion taking time and I was worried about calling out to Juju for any completion. It's just the one controller call but I'm curious on delay/timing feelings for end users16:11
rick_hnammn_de1:  that's why I'm just curious for other folks to try with different data in their controllers output to sanity check making a change that wfm but might have issues/delays for others with more data/etc16:11
nammn_de1rick_h: makes sense to me, We can let the pr open and see how other folks prime in16:12
rick_hnammn_de1:  kind of thing that tests well/fine with one or two controllers but might have unforseen issues with other folks16:12
rick_hnammn_de1:  yea I think we could land it with a request to test given it'll only show in the edge snap16:12
rick_hnammn_de1:  and recruit some feedback16:12
rick_hnammn_de1:  I don't think it'll get much feedback otherwise since folks will need to adapt their local completion scripts16:12
nammn_de1and if its too slow then maybe just remove/rewrite the "error message" could already help16:12

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