=== hggdh_ is now known as hggdh [07:25] uhh, my 3.8.X kernels do not boot, instead they give me kernel panic [07:25] or no panic but stackdump + no boot [07:26] Any hints how to get the text that i see on the screen to a file so i can search the issue? [08:34] #u+1 [11:42] Hiyas all === shadeslayer is now known as kubot1 === kubot1 is now known as shadeslayer [12:19] Hi [12:24] BluesKaj: can you spare me a moment? My 13.04 kernel 3.8.X is segfaulting on every boot but 3.5.X works fine, i should check what is wrong but taking photo from the display sounds stupid, any suggestions? I have checked kernel.log but its not there (as it hapens on the boot i guess .. ) [12:25] susundberg: since 13.04 is released you should ask in #ubuntu [12:28] susundberg, if it's segfaulting then it's difficult to know what's breaking , best to purge the 3.8 kernel and use the 3.5 [12:28] k1l: yeah i guess, but the question how to get kernel dump to a text file is not really release related anyhow .. [12:28] BluesKaj: thanks [12:29] I kind of though to file report but i am not going to take photos of the screen ;) [12:29] susundberg: so you should prefere the main channel. yes. === ChanServ changed the topic of #ubuntu+1 to: Welcome to #ubuntu+1, the channel for discussion of pre-release versions of Ubuntu. | 13.04 has been released!! http://www.ubuntu.com/getubuntu/releasenotes | https://wiki.ubuntu.com/SaucySalamander/ReleaseSchedule [15:49] have saucy installed as vmware machine. kernel 3.9 very nice [15:56] do you think .10 or .11 will be the released kernel? [16:09] johnjohn101: 3.10.x look at https://wiki.ubuntu.com/SaucySalamander/ReleaseSchedule Aug 15: 12.04.3 [18:04] Good evening. [18:13] Anca-Emanuel: Where do you see 3.10.x there? [18:13] The kernel freeze happens on October, 3rd [18:22] bekks: kernel 3.10 will take at least 60 days to develop. That will be july. [18:24] bekks: then they need something to deliver for the hardware enablement stack for LTS 12.04.3 [18:26] so will 12.04.03 have the 3.10 kernel? [18:27] Anca-Emanuel: And? 60 days after end of July we will be at 3.11 possible - end of september :) [18:27] hmm... They aren't supposed to update the kernel version in the stable series? [18:27] Anca-Emanuel: So its not that clear which kernel version will be used, isnt it? [18:28] Ah, I see a precise-updates option for a newer kernel than 3.2 [18:28] johnjojn101: 12.04.1 have 3.5 kernel; 12.04.2 have 3.8 kernel from raring... guess [18:28] i have the 3.5 kernel on 12.04.whatever now [18:29] Anca-Emanuel: 12.04.2 is already release, should have the 3.5 kernel. [18:51] for a simple quick check of system activity, ctrl+esc [19:09] are we getting gcc 4.8? [19:11] its already in [19:13] woot, was going to try to slam it into 12.04 but didn't get around to it, now i have a chance to test!! [19:14] is there a metapackage for getting all the updated packages at once? [19:14] for instance the kernel won't update by default, apparently [19:19] gcc 4.8 is in [19:21] is it bad that when ever i share folders between two ubuntu machines, i use samba to do it? [19:22] no, I'd use samba in every case. It seems to be the most productized solution. [19:22] meaning you can basically right click on a folder icon and get it to work, maybe, without having to puzzle out anything much [19:22] just like windows I guess [19:22] yeah, windows solution for ubuntu [19:23] though some people seem to dislike it when something ordinarily complex becomes rather easy. [19:23] doesn't seem right. yeah, i click share on one folder and then use connect to server on another machine and voila, ez file transfer [19:24] well it's from the windows world and not the linux world. [19:24] I had more trouble getting file sharing to work on windows today... mostly because it's a locked down windows machine and the GUI didn't work, but "net use t: \\blahblah\blaa /permsistent:yes /user:blaa" did [19:25] i still use net use and my coworkers say i'm too old fashioned [19:27] alankila: kernel should update automatically as long as you have linux-image-generic installed [19:27] * yofel runs 3.9.0-0 [19:28] yofel: it doesn't acquire the precise-update, only new kernels from precise I think... [19:28] oh, you were talking about precise... [19:29] Anyway it doesn't actually matter to me in practice [19:31] kernel updates seems really only necessary for hardware [19:32] yeah these are virtual systems [19:32] I'm keen on getting new kernel versions because I run btrfs. Risky as hell though I haven't lost data yet and I have terabytes on it. but I'm about to start using it on a server with customer data, which means I'm going to do backups extra carefully. [19:38] alankila, as long as you do snapshots and have some unused space and whatnot you're ok [19:39] yeh that's what I think [19:40] though I do wonder about my sanity. Still, the experience with btrfs has been positive so far and I've used it on a backups volume due to the snapshot capability for long time [19:45] you should post your findings somewhere and share the knowledge [19:46] long being over a year, but still there's generally some 10 GB of new stuff going in every day and snapshots are kept from days to 1 month with roughly exponential backoff in how the old snapshots get deleted [19:47] I think you could easily torture btrfs for a few days for similar level of exercise of the filesystem, so my accumulated experience is not very valuable [19:50] plus properly constructed test would be able to validate that the filesystem has kept the data correctly. [19:51] the fact I have a backup volume that appears to hold files but which I basically never need means it's not exactly a convincing demonstration of the filesystem's stability. But it has run actual virtual machines for months too and without any issues [19:51] these virtual machines would crash on data corruption pretty soon [20:02] yeah, good luck with all that [20:02] though only 3.8 kernel provided reasonable performance for VM disk images [20:03] btrfs is attractive in many ways in that it provides atomic snapshots that allow continued writing ... just such a pain in the ass with the risk that it might crash or corrupt itself somehow [20:03] the feature set is not bad, and it's right in the kernel, and works with everything. [20:04] like grub. Better than xfs even. You are always playing with fire if your /boot is on XFS for instance. If you reboot too quick after a new kernel, there's stuff in journal which grub isn't capable of reading [20:04] and I dare not even try ZFS for the time being [20:05] i would love to see an article about zfs on ubuntu [20:05] The reason why I mention XFS is that it's another serious contender for a VM host because it has supported a version of hole punch ioctl for a long time. It used to be completely XFS-specific but it allows fstrim on guest to free image range on the host side [20:05] not even sure what that means. [20:07] it means VM disk images can be say 100 GB large but only consume like 1 GB of space. If you do a lot of stuff like touch a lot of disk blocks, they grow, even if the real data contained on the guest FS is still 1 GB [20:07] with support for fstrim and filesystem capable of hole-punching, that space can be freed. [20:08] only used when its used [20:08] I went from 10 GB disk images to 2 GB disk images when I trimmed them [20:08] for instance loop mount can be performed on the raw disk image and fstrim can be performed on the loop mount nowadays [20:09] it will be translated by the kernel's loopback mount driver to hole punch ioctls [20:09] qemu 1.4 users can enjoy ata trim support on the IDE driver [20:12] enabling the trim was difficult though, because it requires passing a parameter discard_granularity=512 to the IDE device description and libvirt doesn't currently support this parameter, so passing it was difficult. I ended up writing a shell wrapper that notices the -device argument and appends it when ide-disk is being defined [20:12] it wasn't pretty but whatever [20:13] annoyingly, the libvirt people even validate that the paramters used in the domain xml specifications can be interpreted correctly by the invoked qemu so they forbid using a comma as value inside. I tried many hacks. For once, someone had engineered this competently. [20:14] I almost quit using libvirt that day and I keep on thinking maybe I should have thrown it to the river. All it really does is start my VMs at boot and stop them at shutdown. That's not a whole lot of useful work, but I hate writing init.d scripts [20:15] who is running saucy yet ;D [20:17] i have it [20:17] is it useable? [20:17] im installing it right now [20:18] so far so good. doesn't look like much has changed since 13.04. 3.9 kernel/ gcc 4.8 is all i know about [20:18] There is any Ubuntu developers here ? Do you plan to support f2fs ? http://ubuntuforums.org/showthread.php?t=2137837&page=2 [20:18] ya thought so [20:19] don't worry, they'll figure out how to break something! [20:19] Anca-Emanuel: usually if its in debian its gonna be in ubuntu [20:20] Anca-Emanuel: if you want something in ubuntu the best way is to get it into debian [20:21] what file system is that for? [20:21] new filesystem for smartphone flash the way it looks to me [20:23] right for samsung fones [20:24] wondering if perl 5.16 will be the default perl [20:26] johnjohn101: f2fs is for flash based memory devices. Early benchmarks: http://www.phoronix.com/scan.php?page=article&item=linux_f2fs_usb3&num=1 [20:28] isn't that in the kernel? [20:28] or not yet? [20:29] http://www.phoronix.com/scan.php?page=news_item&px=MTMxMTU [20:30] looks like 3.9 kernel [20:30] then itll be in saucy [20:30] maybe already is [20:30] or you have to update kernel in raring [20:34] jtaylor: what ? f2fs ? yes it is included in kernel, you can read the docs here: https://github.com/torvalds/linux/blob/master/Documentation/filesystems/f2fs.txt [21:27] hachre: http://cards.linaro.org/browse/CARD-277 sent mail to linaro-dev and Wook Wookey (debian bootstrap expert for arm64) === Jikan is now known as Jikai === Jikai is now known as Jikan === bazhang_ is now known as bazhang === Jikan is now known as Jikai === Jikai is now known as Jikan === Jikan is now known as Jikai === Jikai is now known as Jikan