18:00:51 #startmeeting 18:00:51 Meeting started Fri Jun 14 18:00:51 2013 UTC. The chair is davej. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:51 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:51 #meetingname Fedora Kernel meeting 18:00:51 #meetingtopic Fedora Kernel meeting 18:00:51 The meeting name has been set to 'fedora_kernel_meeting' 18:00:56 #addchair jwb 18:01:09 it's that time again... 18:01:19 whee 18:01:31 jforbes is out this week. I suspect todays meeting will be short. 18:02:08 usual drill: release run-through. 18:02:08 * nirik is lurking around 18:02:27 #topic F17/18 status 18:02:46 F17, F18, and F19 are all on 3.9.5 18:02:49 well 18:03:08 F18 and F19 are on 3.9.5 now. F17 is stuck again, as usual but 3.9.5 is in updates-testing 18:03:15 3.9.6 is built for all 3 but not filed as an update yet 18:03:58 not much else to say on 3.9.x I guess 18:04:06 a few people have said 3.9.5-[23]01 doesn't boot, but that seems to be the usual "missing initramfs" kind of issues 18:04:38 I've kinda tuned out of the EFI bug that got reported, what's the status on that ? 18:04:56 "the EFI bug" 18:04:59 should be fixed now 18:05:15 ok, good to know 18:05:19 i think someone is still hitting issues on an SGI UV machine, but it's being tracked in a separate bug 18:05:29 and tbh, not really a concern for Fedora 18:05:32 probably not many users with those machines anyway 18:06:06 ok, that it for 17-19 ? 18:06:16 believe so. normal karma plea for f17 18:06:38 if we don't get karma soon, we'll have to just cheat the system and pimp it ourselves 18:06:42 yep 18:06:52 ok 18:06:53 davej: same as it ever was 18:06:54 * nirik did +1 it, but that might have been the previous one 18:07:17 #topic rawhide status 18:07:44 so rawhide is a bit behind. I've been getting stuff fixed upstream in the hope that the next -rc doesn't suck 18:07:57 because the stuff I've seen this last week or so has been pretty scary given we're at rc5 18:08:16 (RAID5 breakage, multiple XFS breakage, RCU shitting itself left and right) 18:08:22 its been ok on my laptop, but nothing too fancy here. 18:08:32 i don't think it's behind. it's on rc5. just no git snapshots i a while 18:08:35 and a bunch of stuff I've forgotten. (I've had to start writing it down on my blog to keep it straight) 18:08:48 jwb: yeah, that's kinda what I meant. I'm not doing daily builds 18:09:06 It's working well for me. Some stalls when disk use was high (that may or may not have been kernel related) seem to have gone away. 18:10:29 another point is that rawhide has debug off, because I've only been doing builds on the rc release. part of the reasoning behind this is that I'm seeing enough problems in my own testing without needing more bugs right now. 18:11:01 * nirik is happy for the no debug. Makes things zippy 18:11:25 when they aren't crashing ;) 18:11:27 so, I've got a handle on what still needs fixing, and I'm going to keep nagging people upstream to get stuff merged 18:11:59 that's about it for rawhide from me 18:12:45 #topic testing project 18:13:42 Any issues or concerns with those test machines? :) 18:13:49 as I mentioned, jforbes is off, so he's not around to give status on this this week, but it's kinda stalled due to numerous problems we've had just keeping the boxes up and running. When he's back next week, hopefully we can sort that out. 18:14:03 ok 18:14:17 I've had problems even logging into them 18:14:22 :( 18:14:31 well, happy to help poke them, etc... 18:14:41 you can find infra folks in #fedora-noc if you need us 18:14:52 I sent smooge some mail about it, he was poking at it yesterday 18:15:20 but thanks for the offer, might take you up on it if we don't see any progress 18:15:32 * nirik nods. 18:16:17 on the side from this, I've been doing some work on some scripting to test various IO configurations. Fear my ugly bash: https://github.com/kernelslacker/io-tests/blob/master/setup.sh 18:16:35 that already tripped up the raid5 bug I mentioned above. 18:17:02 what else.. continued work on the fuzzing tools, various improvements there 18:17:26 and jwb got some of the IO stuff from ltp into our test cases 18:17:37 io, fs, ipc, and something else i can't remember 18:17:46 probably for the best 18:17:53 cool. 18:17:57 too much ltp exposure is known by the state of California to... something 18:18:03 signal testing maybe. anyway, the whole LTP thing is build and available, just need to look at what all we want 18:18:07 and how long it runs 18:18:14 would i/o schedulers be interesting? or since cfq is default and few people change it, not so? 18:18:40 nirik: yeah, been trying to think of the best way to do that. maybe pick one at random for each workload 18:18:54 because cycling through every possibility increases the runtime of the tests a lot 18:19:00 yeah. 18:19:10 (it already took ~15 minutes on my 3-disk setup to run that script) 18:19:19 i figured those might be more suited for performance testing 18:20:05 jwb: yeah, we don't really see crashes there that often. still, nice to have more coverage testing 18:20:56 so think that's about it for testing stuff for now 18:21:06 #topic open floor 18:22:34 anyone ? Or are we closing early ? 18:23:08 close 18:23:28 looks like it. thanks for coming everyone. 18:23:35 #endmeeting