timClicks | wallyworld: what is the timeline for juju scp support in k8s models? | 02:02 |
---|---|---|
wallyworld | timClicks: hopefully this cycle but it is a stretch goal as it got punted | 02:14 |
timClicks | ok | 02:45 |
wallyworld | hpidcock: if you have a moment at some point https://github.com/CanonicalLtd/juju-qa-jenkins/pull/470 | 03:10 |
hpidcock | wallyworld: sure thing | 03:30 |
hpidcock | wallyworld: lgtm | 03:44 |
wallyworld | ty | 03:45 |
timClicks | 2 easy PRs https://github.com/juju/juju/pull/11704 https://github.com/juju/juju/pull/11712 | 04:52 |
thumper | wallyworld: re: bug 1872670, if it is working on eu/us region, surely it isn | 05:15 |
mup | Bug #1872670: AWS m6g.* instances fail with latest juju <juju:Confirmed> <https://launchpad.net/bugs/1872670> | 05:15 |
thumper | isn't juju, but account issue? | 05:15 |
thumper | clearly latest/edge won't work for them | 05:15 |
thumper | but 2.8.0 stable should | 05:16 |
flxfoo | hi there | 08:59 |
flxfoo | interface for apache-solr uses deprecated reactive RelationBase, where can I find documentation to be able to define departed / joined like parts , please? | 09:00 |
stickupkid | manadart, https://github.com/juju/juju/pull/11715 | 10:47 |
stickupkid | manadart, might affect your changes you want to propose | 10:47 |
manadart | stickupkid: I also notice that `idle_subordinate_condition` does not use the declared `app_index`. | 10:51 |
stickupkid | ah good spot | 10:51 |
achilleasa | manadart: stickupkid can you take a look (and QA) https://github.com/juju/juju/pull/11716 and https://github.com/juju/juju/pull/11714? | 10:57 |
achilleasa | fun with virtual maas ;-) | 10:58 |
stickupkid | I like how in the integration tests we have .log and .txt extensions | 11:04 |
stickupkid | PICK ONE SIMON! | 11:04 |
stickupkid | I'll fix that at some point | 11:04 |
manadart | achilleasa: Yep. | 11:09 |
stickupkid | manadart, updated https://github.com/juju/juju/pull/11715, you where correct, just run the whole deploy suite and it worked fine | 11:15 |
stickupkid | deploy suite is hefty | 11:16 |
stickupkid | this annoys me https://paste.ubuntu.com/p/tX7jdGff79/ we really should sort the image server nested containers manadart | 11:18 |
stickupkid | also I've never seen the error/issue when trying to retrieve the environ | 11:20 |
manadart | stickupkid: You now have a gap in the args indexes :) | 11:22 |
manadart | stickupkid: Tell me what I am doing wrong here. I have this: https://pastebin.canonical.com/p/QZKBVhd57V/ | 11:24 |
manadart | And this query: `wait_for "network-health-xenial" "$(idle_subordinate_condition "network-health-xenial", "mongodb")"` | 11:24 |
manadart | If I fill out the generated query manually, I get what I expect: https://pastebin.canonical.com/p/dYbxvTZMV8/ | 11:26 |
manadart | stickupkid: Gonna eat. Maybe you can school me in a few. | 11:27 |
manadart | stickupkid: Got time for a quick HO? | 11:41 |
manadart | Need a review: https://github.com/juju/juju/pull/11717 | 12:42 |
jam | hey manadart. I just came across something unexpected: https://bugs.launchpad.net/juju/+bug/1883703 | 12:53 |
mup | Bug #1883703: Snap build not stripping the binary? <build> <juju:Triaged> <https://launchpad.net/bugs/1883703> | 12:53 |
jam | basically, the 'juju' in our snap is 120MB instead of 89MB and it appears it is *not* stripped. | 12:53 |
jam | I can't make it to standup, and I don't really know the priority for it, but I wanted to raise awareness | 12:53 |
manadart | jam: Ack; we will discuss. | 12:56 |
jam | manadart, it seems we use our own extension of snapcraft's go plugin | 12:57 |
jam | and it has https://github.com/juju/juju/blob/develop/snap/plugins/juju_go.py#L107 | 12:57 |
jam | which isn't supplying '-ldflags "-s -w ' | 12:57 |
hml | stickupkid: I’ve gotten this 2 of 2 times on 11715: https://pastebin.canonical.com/p/bkPRmhPRWV/ can’t confirm if it’s timing related or not | 13:48 |
achilleasa | back-port for the kvm fix: https://github.com/juju/juju/pull/11718 and https://github.com/juju/juju/pull/11719 | 13:50 |
achilleasa | QA should be easy on this one ;-) | 13:50 |
stickupkid | hml, i'll get to that after fixing the issue for manadart | 13:53 |
hml | stickupkid: k | 13:53 |
=== narindergupta is now known as narinderguptamac | ||
=== sadin999_ is now known as sadin999 | ||
=== niemeyer_ is now known as niemeyer | ||
=== nniehoff_ is now known as nniehoff | ||
=== joedborg_ is now known as joedborg | ||
stickupkid | why is it sometimes does juju status take 22 seconds on a local lxd setup | 14:19 |
stickupkid | something very strange is going on | 14:19 |
stickupkid | manadart, got a second | 14:54 |
stickupkid | another case of a random lock up when querying juju status https://paste.ubuntu.com/p/yYSnzxJ2pN/ | 15:19 |
stickupkid | see timestamp between the two different status timestamps | 15:19 |
stickupkid | should juju hang if a lxd container is down, i wonder if the controller is down? | 15:26 |
stickupkid | hml, re-writing the container integration test, but can confirm it the new stuff works for lxd and aws... | 16:22 |
stickupkid | hml, juju action wasn't working as I was targetting the wrong unit | 16:22 |
hml | stickupkid: ack | 16:23 |
hml | stickupkid: i think i have a test failure to deal with… doesn’t reproduce locally. | 16:24 |
hml | stickupkid: [LOG] 0:01.949 INFO test unit "u/0" shutting down: preparing operation "install cs:quantal/wordpress-0": failed to download charm "cs:quantal/wordpress-0" from API server: download aborted | 16:24 |
stickupkid | hml, isn't that the mock one? | 16:24 |
hml | stickupkid: unfortunately not… its in the uniter.loop code… and startup | 16:25 |
hml | no easy way to mock without rewritting, afaik | 16:25 |
hml | rewritting uniter stuff | 16:25 |
stickupkid | "cs:quantal/wordpress-0" it's just that it using the testcharms though | 16:26 |
hml | stickupkid: yeah, that’s what i don’t get… it should be using the local charm repro for test | 16:26 |
stickupkid | that's why i was wondering if it was the mock store | 16:27 |
stickupkid | hml, quick ho? | 16:27 |
hml | stickupkid: sure | 16:28 |
pmatulis | does the '--map-machines' option need to be used when deploying with both a bundle and an overlay? i'm thinking that the bundle will define the existing machines and then the overlay needs the option. or is an existing machine completely "pre-deploy"? | 20:08 |
wpk | 2/wind 343 | 20:33 |
josephillips | hey | 20:38 |
josephillips | im installing swift | 20:38 |
josephillips | on different zones | 20:38 |
josephillips | but when i perform teh relation betwen swift-storage-zonename swift-proxy | 20:39 |
josephillips | keep on missing relation proxy | 20:39 |
hml | hey wpk | 20:44 |
hml | josephillips: and no errors with the add-relation? | 20:45 |
hml | josephillips: do you have a pastebin you can share? i’m not seeing a relation “proxy” for either of those charm | 20:50 |
hml | pastebin with the output you’re seeing | 20:50 |
wpk | hml: o/ | 20:51 |
josephillips | sure hml im checking something else | 20:58 |
josephillips | give me one sec | 20:59 |
josephillips | my bad | 21:00 |
josephillips | was a bad configuraiton as block device | 21:00 |
hml | josephillips: rgr | 21:05 |
wallyworld | hpidcock: kelvinliu: tlm: see comment on bug 1875481 ... another issue with running juju agent in the workload container. the base linux may not support our jujud | 23:53 |
mup | Bug #1875481: Juju 2.8 error about unit not being the leader <k8s> <juju:Incomplete> <https://launchpad.net/bugs/1875481> | 23:53 |
hpidcock | wallyworld: I thought we were statically compiling jujud | 23:54 |
hpidcock | oh maybe that was only in snaps | 23:55 |
wallyworld | i thought we did too | 23:55 |
wallyworld | hpidcock: ah forgot to mention, there's also an issue with our snap | 23:55 |
wallyworld | bug 1883703 | 23:55 |
mup | Bug #1883703: Snap build not stripping the binary? <build> <juju:Triaged> <https://launchpad.net/bugs/1883703> | 23:55 |
wallyworld | i haven't looked closely, theory is he go plugin doesn't supply the necessary flags | 23:56 |
kelvinliu | I found all jujud images doesnot have jujuc added | 23:56 |
wallyworld | something to fix for 2.8.1 release | 23:56 |
hpidcock | kelvinliu: that shouldn't affect it, but yeah I noticed the same thing | 23:57 |
wallyworld | kelvinliu: yeah, jujuc is a separate binary that we need to package if we haven't already | 23:57 |
kelvinliu | https://pastebin.ubuntu.com/p/v9d3MRSXS5/ | 23:58 |
kelvinliu | it's weird if docker build didn't fail when jujuc was missing | 23:58 |
kelvinliu | docker should complain about ADD failed because jujuc was not existing | 23:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!