[19:47] <ogra_> stgraber, could we deleteimage 16 and 17 ?  seems we have a busted udev
[19:48] <ogra_> (touch that is indeed)
[19:49] <stgraber> ogra_: in -proposed?
[19:49] <ogra_> yeah
[19:49] <ogra_> trusty-proposed
[19:49] <stgraber> so deleteimage is a bad idea since it won't downgrade users and will break the next upgrade
[19:50] <stgraber> but you can copy image 15 as image 18
[19:50] <ogra_> oh, indeed
[19:50] <ogra_> do i do that without -k ?
[19:51] <stgraber> right, with -k it'd fail because the version already exists, so just do copy-image trusty-proposed trusty-proposed <device> 15
[19:51] <ogra_> k
[19:52] <stgraber> I never actually tried doing that, but it's what I recommended to asac as the rollback procedure when something goes wrong, so if that fails somehow, I'll just fix it (but I don't think I put any check in place to prevent a copy from within the channel)
[19:53] <ogra_> its running, lets see
[19:53] <stgraber> oh yeah and it's probably going to be slow as it'll be computing some missing deltas in the process (from 17 to 18)
[19:54] <ogra_> thats fine, i just want to make sure people dont upgrade to the broken one ... if it takes time thats ok
[19:59] <stgraber> ogra_: did you get any error? it looks like it failed for mako at least
[19:59] <ogra_> whoops, didnt watch it, yeah
[19:59] <ogra_> http://paste.ubuntu.com/6389999/
[20:01] <stgraber> ogra_: ok and it looks like the index is a bit messed up as a result... I'll revert what just happened and I'll figure out what went wrong and do the copy myself in a bit
[20:01] <ogra_> thanks
[20:20] <stgraber> ogra_: ok, I think I fixed copy-image, running it now
[20:20] <ogra_> great, thanks ... and sorry for bothering you on a sat. with that
[20:21] <stgraber> another problem this triggers is that import-images will try to re-import the current image as build 19 now since it's technically more recent than our revert is at build 18...
[20:21] <stgraber> so I turned of import-images for now
[20:21] <stgraber> *off
[20:21] <stgraber> and added a TODO item to have it cope with reverts properly
[20:22] <stgraber> (it should notice that the latest cdimage available has already been imported in the past and so refused to publish it again)
[20:24] <stgraber> ogra_: done, the indexes now look reasonable on the public server
[20:27] <stgraber> ogra_: I ended up doing something slightly cleaner actually, import-images still runs but the trusty-proposed channel has been marked as manual in the config, so won't try to auto-import new cdimage images
[20:28] <stgraber> ogra_: let me know once we've got a good build again and I'll flip it back to auto
[20:39] <ogra_> might be tomorrow ... depending if ricardo actually invests time into udev
[20:40] <ogra_> if he gives up (i think he shouldnt put to much time into it on a weekend) it might be monday after pitti fixed it
[20:42] <ogra_> the worst bit is fixed though ... people will be safe
[20:55] <stgraber> ok. I'll be spending most of tomorrow unloading and setting up servers in a DC but should be on IRC in the afternoon (my time) so just ping here and I'll do the config switch and a debug run to make sure everything goes as planned
[22:08] <xnox> ogra_: is there a bug # about the udev issue?
[22:18] <ogra_> xnox, not yet, i was waiting for ricardo, he wanted to take a look (probably its a quick fix, but the diff is sadly huge)
[22:19] <ogra_> xnox, http://paste.ubuntu.com/6389687/ ... thats /dev on mako with the new udev
[22:19] <xnox> =/
[22:20] <ogra_> (all labels gone)