[21:27] <lifeless> wgrant: why does deleting a ppa supersede its publications ?
[22:49] <huwshimi> The monday morning channel silence...
[22:50] <huwshimi> Which makes me fear that:
[22:50] <huwshimi> a) I'm not connected, or
[22:50] <huwshimi> b) it's not Monday
[22:50] <lifeless> its a public holiday in vic/sa/tas
[22:50] <huwshimi> lifeless: Ah of course
[22:50] <lifeless> thumper and I are here
[22:52] <jelmer> huwshimi: also, it's not Monday :-P
[22:52] <lifeless> jelmer: yes it is :)
[22:52] <thumper> lifeless: what about queensland?
[22:52] <lifeless> thumper: not a holiday
[22:52] <lifeless> thumper: AFAIK
[22:54] <lifeless> oh, and act
[23:04]  * mwhudson waves
[23:05]  * thumper waves at mwhudson
[23:37] <wgrant> lifeless: Possibly to make it easier to expire, although it fails to let them expire anyway.
[23:38] <wgrant> lifeless: If you want to make it fast, just grab all the binary and source pubs separately and delete them.
[23:38] <wgrant> Rather than letting the sources find and delete their binaries.
[23:38] <lifeless> wgrant: I'm wondering why we bother
[23:38] <wgrant> Or make it batch-call getBuiltBinaries, since that's reasonably efficient when not called hundreds of times.
[23:39] <lifeless> wgrant: isn't just flagging it deleted sufficient ?
[23:40] <wgrant> lifeless: That won't let the files expire.
[23:40] <wgrant> But as it stands they won't expire anyway.
[23:40] <wgrant> Due to another bug.
[23:40] <lifeless> wgrant: so the point is to free uplibrarian space
[23:40] <lifeless> ?
[23:41] <wgrant> I believe so.
[23:41] <lifeless> wgrant: anyhow, the reason I ask is we have a timeout with 9 second spent in a single update query marking the rows expired
[23:41] <wgrant> Hmm, we must be down about 5 appservers by now.
[23:41] <lifeless> 435 rows
[23:41] <wgrant> lifeless: Yeah, I saw that.
[23:41] <lifeless> wgrant: you think more are dying again ?
[23:42] <wgrant> lifeless: Appservers hang... it's a normal expected thing :/
[23:42] <lifeless> wgrant: not entirely
[23:43] <lifeless> wgrant: we had a bunch exit
[23:43] <wgrant> lpnet1/2/6/10 are dead at the moment.
[23:43] <lifeless> wgrant: what tells you that ?
[23:43] <wgrant> lifeless: They had connections killed over the weekend.
[23:43] <lifeless> wgrant: elmo started them up again around 5am your time
[23:43] <lifeless> wgrant: see #is backscroll
[23:43] <wgrant> Ah, handy.
[23:45] <lifeless> wgrant: you're on leave today?
[23:45] <wgrant> lifeless: Indeed.
[23:45] <lifeless> kk
[23:46] <lifeless> can I interest you in qa of Revision 12581 can not be deployed: needstesting
[23:46] <wgrant> Is that poppy?
[23:46] <lifeless> StevenK: 12580 needs qa too
[23:46] <lifeless> wgrant: yes
[23:46] <lifeless> we get those fixed and us losa can deploy a fix for search performance
[23:47] <wgrant> I already mostly QA'd that before it landed for just this reason, so I'll quickly retest now.
[23:47] <lifeless> looks like we're still doing memcache on bugtask:+index
[23:55] <lifeless> :(
[23:55] <lifeless> https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1898L1962#statementlog is full of get/sets
[23:57] <wgrant> lifeless: NullBugTask:+indec
[23:57] <wgrant> s/c/x/
[23:57] <wgrant> Not BugTask:+index.
[23:58] <lifeless> argh
[23:58] <lifeless> righto
[23:59] <wgrant> lifeless: You can qa-ok the DSD thing if you want. The altered method is not called yet.