[00:43] * hyperair wonders if anyone here has had issues with plymouth.. [00:44] er i mean mountall + plymouth [00:44] fsck seems to hang [00:46] bug 571707 [00:46] Launchpad bug 571707 in mountall "fsck at bootstrap is too slow" [High,Confirmed] https://launchpad.net/bugs/571707 [00:50] hyperair: I just posted a workaround in description ;) [00:51] arand: i tried that, it didn't seem to work very well for me. [00:51] arand: in particular, did you try watching your hard disk light? [00:51] arand: the hard disk light turns off and stays that way after 80% [00:51] hyperair: And spoke to keybuk, which said that it will likely take acouple of weeks before it will be addressed.. [00:51] * hyperair sighs [00:52] lucid's in a pretty sad state eh.. [00:52] hyperair: Yes, that's the same as I had I think. [00:52] arand: does your hd light come back on when you switch ttys? [00:53] hyperair: We've gone from no sound in karmic to no boot in lucid, it's not a good trade... [00:53] hyperair: I could test I guess, I'm using a virtualbox for it. [00:53] arand: no sound in karmic? [00:53] arand: i had plenty of sound. [00:55] hyperair: Yea, me too, but that seemed to be the one common bug that people saw in karmic. [00:56] arand: but what was the cause of the issue? [00:56] hmm maybe it was pulseaudio hanging, it seemed pretty prone to hanging in karmic. [00:58] hyperair: A couple of things, the old kernel on upgrade, the simple volume muted... At least my impression of karmic was that it was the "nosound" release, if naming from issues. [00:58] lol [00:58] karmic was that golden release for me.. =\ [00:58] besides intrepid [00:58] jaunty was shit. [00:59] lucid... *groan* [00:59] well i expect lucid.1 will be better. [01:02] In hindsight, plymouth might not have been the bestest... But then again, it was tried and discarded in karmic, so it made a whole lot of sense to push it now or never... [01:04] on the contrary. [01:04] we should have pushed it in karmic [01:04] that would have made for a solid LTS. [01:05] what we have now is a LTS full of partially implemented changes. [01:05] Yea, but a very unkarmic koala though... [01:06] one is plymouth. another is the initramfs-tools resume script. and yet another is the application indicators, as not all the apps have been ported over to use them yet [01:06] Hmm, fsck halts at 70% and HD activity is null. [01:06] ye [01:07] now try switching vts [01:07] and check the HD activity [01:07] does it return? [01:07] still no activity [01:08] when switching back to splash it's up to 90% [01:08] hmm [01:08] interesting [01:08] does it go past 90%? [01:08] i fell asleep at 90% and just ctrl+alt+del'd my way out [01:08] *cringe* ah, doesn't all this parallel startup bring back memories of Windows XP bootup problems? [01:09] And then switching back to VT, and it boots in 10 seconds [01:09] "ah it's hanging at the blue screen with the cursor and nothing's happening!" [01:09] So it seems it takes two VT-switches to get it past... *scratches head* [01:11] And if I switch to early, the fsck is interrupted... [01:11] jdong: oh yay we're finally reaching where windows XP was... er how many years ago? =p [01:11] lol [01:13] hyperair: ;-) [01:14] hyperair: well we still have to figure out how to actually move blocks around ;-) [01:14] jdong: blocks? [01:14] hyperair: http://jdong.mit.edu/~jdong/ureadahead-pack.svg [01:14] hyperair: that's the disk access pattern during bootup for Ubuntu ;-) [01:14] whoa. [01:14] hyperair: Windows since XP on idle will move all boot blocks together to the front of the disk. [01:14] interesting. === nobawk|away is now known as nobawk [01:15] hyperair: and Vista/7 takes it further by using some fancy machine learning algorithm to lay out the disk [01:15] ._. [01:15] interesting. [01:15] (and prefetch things into cache before they're asked for) [01:15] sounds like they're ahead of us. [01:15] so hehe it's interesting to see Ubuntu walk down that path. [01:15] heh [01:16] hyperair: well why else does a 20GB Windows 7 install boot pretty much the same speed as a 1.2GB Ubuntu install? :-/ [01:16] jdong: good point. [01:16] well we still boot much faster than vista and xp though [01:16] \o/ [01:16] despite not using far-fetched algos [01:17] well sure, but that's like saying my lawnmower gets better fuel economy than a ferrari ;-) [01:17] i imagine if we did something like that, we'd break our 10-second goal [01:17] jdong: well said =p [01:18] so we should have something like this for the next gsoc [01:18] hyperair: yeah I think that'll be our next step. at least the ability to statically use ext4 defrag APIs to move boot files together. [01:18] basically some hack that allows a userspace program to tell the kernel to shift these blocks where [01:18] hyperair: and in the future, the ability to dynamically do this via btrfs [01:18] ooh e4defrag huh [01:18] yay btrfs [01:20] Aw.. what on earth, it doesn't work with normal VT-switching, I need to ude ctrl+al+arrow, and do it twice for it to work properly, ...the fish!? [01:22] Hang on, not even that, just arrowkeys works... I wonder if plymouth interprets arrowkeys in some special fancy way.. [01:24] the "C" doesn't work for me [01:28] hyperair: No, seems like plymouth itself is almost gone at this point [01:31] arand: heh? [01:34] hmmm i discovered something. [01:35] it seems that X has a tendency to hang under some circumstances if started without "splash" [01:42] Horray for more bugs! [03:26] polemic users always make me a sad panda. [03:27] lol === rgreening_ is now known as rgreening === nobawk is now known as nobawk|away [11:06] Rhonda: The wesnoth packages you had posted are not accessible anymore. === nobawk|away is now known as nobawk === traveller__ is now known as traveller === blueyed_ is now known as blueyed === yofel_ is now known as yofel === nobawk is now known as nobawk|away [18:08] Rhonda: there? === NCommand1r is now known as NCommander === NCommander is now known as Guest43091 === Guest43091 is now known as NCommander [20:12] anyone from the ppa nvidia vdpau team here? [21:22] My package contains a twisted module. I install it with the distutils » setup() » py_modules variable, and it goes to /usr/share/pyshared/twisted/plugins/myplugin.py. But in order for it to work a symlink is also needed in /usr/lib/python2.6/dist-packages/twisted/plugins/myplugin.py. What's the normal way to create that symlink? I can't find it in any docs... [22:58] any git guru's round ?