[01:30] babbageclunk: can you please take a look at https://github.com/juju/1.25-upgrade/pull/29 when you have a moment? [01:30] axw: sure [01:57] babbageclunk: thanks [01:58] axw: sorry, meant to ping here but got distracted! [03:54] axw: if you get a chance, here's a PR which adds to the processing of suspended relations. i'd also like to catch up with you before my EOD for a quick chat https://github.com/juju/juju/pull/7862 [03:55] wallyworld: will take a look after lunch [03:56] wallyworld: and will be free to chat then too [03:56] sgtm [04:22] I think I'm though many of the bundle hurdles [04:22] but this last one is kinda biig [04:23] I think I have come across a wonderful undocumented feature [04:23] and I need to unpack what it means [04:33] awesome [05:18] wallyworld: reviewed, and free to chat when you are [05:18] axw: ok, give me 5 [05:29] axw: standup ho? [05:29] wallyworld: okey dokey [06:00] wallyworld: with the dyingOrDead thing, what I was getting at is: if the relation is Dying and also Suspended, won't it fail to be removed? doesn't the Suspended state mean the departed/broken hooks won't be run? [06:05] axw: if it is already suspended, the hooks won't run as they already have been, but the uniter still removes the relation from its state [06:06] ah yes, already run, of course [06:08] * wallyworld afk to buy more coffee, it's an emergency === frankban|afk is now known as frankban [09:54] axw: ping [10:21] I was trying to figure out why it takes so long to rebuild the tests for state, as we end up importing apiserver [10:21] it turns out that it is because we are using provider/dummy, but we seem to only be using it for storage [10:21] is there another way we could test those, or possibly move it to a different dir? then go test for state would have far less to rebuild [10:22] hm, maybe I'm wrong [10:22] looks like internalStateSuite also uses dummy === frankban is now known as frankban|afk [18:19] Dmitrii-Sh, jhobbs: i think i might have hit bug 1717590 the other day trying to run the amulet tests for the percona-cluster charm - i hit with juju based on develop, but not with 2.2.4 - i didn’t look into what the cause was, focused on other things. nothing else was running in the model. [18:19] Bug #1717590: percona-cluster unit constantly fires hooks, blocks hacluster unit from being setup hacluster charm:Incomplete> [20:00] hml: hmm, good info. Might need to try to loop-test it