| infinity | BenC: Want to rebase to 3.8.0-22.33 and close #1181305 in your changelog? | 04:12 |
|---|---|---|
| infinity | BenC: You might have to then give me a debian package (or a git tag) to sponsor to the kernel PPA. | 04:12 |
| BenC | infinity: working on it…should be about an hour | 04:27 |
| infinity | BenC: Shiny. | 04:28 |
| infinity | zequence: You have three more lowlatency tracking bugs. The fun never ends: | 04:30 |
| infinity | https://bugs.launchpad.net/ubuntu/+source/linux-lowlatency/+bug/1181103 | 04:30 |
| ubot2 | Ubuntu bug 1181103 in Kernel SRU Workflow "linux-lowlatency: <version to be filled> -proposed tracker" [Medium,In progress] | 04:30 |
| infinity | https://bugs.launchpad.net/ubuntu/+source/linux-lowlatency/+bug/1181062 | 04:30 |
| ubot2 | Ubuntu bug 1181062 in Kernel SRU Workflow "linux-lowlatency: <version to be filled> -proposed tracker" [Medium,In progress] | 04:30 |
| infinity | https://bugs.launchpad.net/ubuntu/+source/linux-lowlatency/+bug/1181021 | 04:30 |
| ubot2 | Ubuntu bug 1181021 in Kernel SRU Workflow "linux-lowlatency: <version to be filled> -proposed tracker" [Medium,In progress] | 04:30 |
| infinity | zequence: When you get those up in your PPA, I can sponsor the copies if Andy's not around. | 04:32 |
| === akashj87_ is now known as akashj87 | ||
| zequence | infinity: cool. I'll have the kernels built sometime today. I'll ping you | 07:08 |
| infinity | zequence: Awesome, thanks. | 07:10 |
| BenC | infinity: whoops…fell asleep…tag is PPC-Ubuntu-3.8.0-10.16 | 14:32 |
| BenC | Give it a sec to sync... | 14:33 |
| BenC | infinity: done | 14:36 |
| BenC | infinity: for meta, just use linux-meta-ppc master | 14:37 |
| BenC | infinity: Hope I'm doing the workflow bug correctly | 14:57 |
| zequence | infinity: all lowlatency have been uploaded to the PPA now | 17:25 |
| lifeboy | Anyone here over the weekend that can give advice on loading the latest rc kernel on 12.04 to get a patch? | 19:46 |
| lifeboy | I run 12.04 64bit server with libvirt and QEMU-KVM and have run into the bug 1071322 and need to patch it, but don't quite see how. | 19:47 |
| ubot2 | Launchpad bug 1071322 in linux (Ubuntu) "Quantal kernel for precise: starting KVM VM causes 'vmwrite error'" [Medium,Confirmed] https://launchpad.net/bugs/1071322 | 19:47 |
| lifeboy | The patch has been applied to kernel 3.7-rc6 and upwards. Is there a way in which I can easily install that kernel? Or could I apply a patch for it? If so, where do I find the patch? | 19:47 |
| lifeboy | Reading through the kernel team's MainlineBuilds page at https://wiki.ubuntu.com/Kernel/MainlineBuilds?action=show&redirect=KernelTeam%2FMainlineBuilds I installed the amd64 3.7.7-999-generic debs from http://kernel.ubuntu.com/~kernel-ppa/mainline/prestable/linux-3.7.y/2013-02-13-raring/ which allows me to boot my server and the bug mentioned has been fixed. | 20:15 |
| lifeboy | That's all fine, but it means that I will have to apply future kernel updates manually now, or will the the update manager install new kernel updates post 3.7.7-999 when 12.04 has reached that kernel version? | 20:17 |
| infinity | lifeboy: You might consider installing linux-generic-lts-raring if you're happy running newer kernels on precise. | 20:25 |
| lifeboy | infinity: Would the lts refer to long term support? | 20:50 |
| infinity | lifeboy: lts-raring, lts-quantal, etc refer to it being "the raring kernel backported to the LTS release". | 20:51 |
| lifeboy | Maybe I should ask, is the bug I refered to a problem or can I just ignore it and stay with the stock kernel? | 20:52 |
| infinity | I have no idea. I know lots of people run KVM on precise (including us, in production). | 20:52 |
| lifeboy | I don't really want to run newer kernels. Something's bound to break at some stage... | 20:52 |
| infinity | Well, if you're already running lts-quantal, you're already on the HWE track anyway. | 20:53 |
| infinity | At least, the bug report is on 3.5.0, were you as well? | 20:53 |
| lifeboy | I'm really just testing to see if I can make that error go away and will roll back to stock standard once I've tested | 20:54 |
| infinity | Yes, but what is "stock" to you? | 20:54 |
| infinity | Like I said, the bug report was on 3.5.0 (ie: lts-quantal), what were you running when you ran into it? | 20:54 |
| lifeboy | Whatever the kernel is that ubuntu installs via update manager | 20:54 |
| infinity | That depends on how you installed. | 20:55 |
| lifeboy | I was on 3.5 yes | 20:55 |
| infinity | Okay, right. So, you're on the HWE track anyway. | 20:55 |
| lifeboy | I did a server install from an ISO | 20:55 |
| lifeboy | HWE? | 20:55 |
| infinity | (The "stock" precise kernel is 3.2.0) | 20:55 |
| infinity | HardWare Enablement. | 20:55 |
| lifeboy | Oh wait, sorry, I got confused. My desktop is on 3.5 (quantal), the server was 3.2 (precise) before I installed 3.7.7. | 20:56 |
| infinity | Kay. Well. The way forward is one of three things: 1) Ingore the error, maybe it doesn't matter too much, 2) comment on that bug report that you've seen the error on 3.2 as well, and include logs, and suggest pulling the fix back to the 3.2 kernel, 3) use the lts-raring (3.8) kernel and wash your hands of it. | 20:58 |
| zequence2 | infinity: Seems like I've made a mistake with providing the correct version for the build. Sorry, didn't double check before uploading | 21:00 |
| lifeboy | Infinity: I need advice then. This is an LTSP server for an office which we are re-installing (on 12.04). I don't want to play with stuff that will cause problems later, so if I go for lts-raring (3.8), would that be considered risky? | 21:00 |
| infinity | zequence2: Hrm? What mistake? I reviewed it all before I copied, and it looked fine to me. | 21:00 |
| infinity | zequence2: (Well, other than you feeding the wrong '-v' to genchanges and getting enormous .changes files, was that what you meant?) | 21:00 |
| zequence2 | infinity: Yeah. two of them had too long changes | 21:01 |
| infinity | zequence2: Yeah, I noticed that but oh well. | 21:01 |
| zequence2 | Was baby sitting a 4 year old, and was kind of in a hurry, but I still should have double checked. | 21:03 |
| lifeboy | infinity: Would that be considered risky? | 21:04 |
| infinity | lifeboy: Are you sure you saw the exact same bug on 3.2? | 21:04 |
| infinity | lifeboy: The commit that introduced this bug happened after 3.2 was released. | 21:04 |
| lifeboy | Let me check my logs | 21:04 |
| infinity | zequence2: No big deal. Wasn't worth me asking you to reupload for. A few big emails to -changes won't kill anyone. :) | 21:05 |
| infinity | zequence2: The sources looked fine, which is what matters. | 21:05 |
| lifeboy | infinity: Dang! I was wrong about being mistaken... | 21:07 |
| lifeboy | May 17 08:20:32 parow kernel: [28030.165662] vmwrite error: reg 401e value a023a000 (err 12) | 21:07 |
| lifeboy | May 17 08:20:32 parow kernel: [28030.165670] Pid: 25778, comm: kvm Not tainted 3.5.0-28-generic #48~precise1-Ubuntu | 21:07 |
| lifeboy | So I'm on HWE, so would it be safe to install lts-raring 3.6 then? | 21:09 |
| infinity | lifeboy: Okay, so. If that's a server that isn't running X, doesn't need fancy drivers, etc, the simplest solution might be for you to switch to the 3.2 kernel. | 21:09 |
| lifeboy | s/3.6/3.8/ | 21:09 |
| lifeboy | No, it's running X! | 21:09 |
| lifeboy | LTSP runs X in each session | 21:09 |
| infinity | ... not locally. | 21:10 |
| lifeboy | And if I have users connecting via nx to that server, that's a local X session, not? | 21:11 |
| lifeboy | Or do you mean X on the console? | 21:12 |
| infinity | Unless it's rendering on the server's video card (it's not), it's not a local session from the POV of what I was referring to. | 21:12 |
| lifeboy | Ah, I'm with you. | 21:12 |
| infinity | The HWE stack is specifically about hardware drivers, after all. | 21:12 |
| lifeboy | So if go back to 3.2 and but a hold on the linux kernel package that it doesn't want to upgrade to a later kernel version all the time, it should work fine? | 21:14 |
| infinity | No reason to hold it. | 21:15 |
| lifeboy | Will it stay on 3.2? | 21:15 |
| infinity | Yes, if you remove the 3.5/3.7 kernels you installed. | 21:15 |
| infinity | But if that was installed with 12.04.2 media and has the HWE X stack, you'll want to switch that too. | 21:15 |
| lifeboy | It unfamiliar territory for me: How do I switch away from the HWE X stack then? | 21:17 |
| infinity | Something like: "apt-get --purge install linux-generic xserver-xorg-lts-precise" | 21:17 |
| infinity | And then you'll also need to purge all the 3.5.0 and 3.7 kernels you installed. | 21:17 |
| infinity | "dpkg -l linux\*3.[57]\* | awk '/^i/ {print $1}' | xargs dpkg -P" | 21:19 |
| infinity | Ish. | 21:19 |
| lifeboy | Ah, ok, no rocket science then. For some reason the media I used for this server is different, probably the alternative CD instead of the standard server... :-( | 21:19 |
| infinity | Err. | 21:19 |
| infinity | "dpkg -l linux\*3.[57]\* | awk '/^i/ {print $2}' | xargs dpkg -P" | 21:19 |
| lifeboy | thanks infinity. Let me see how it goes. | 21:21 |
| lifeboy | infinity: For a server with LTSP & libvirt which kernel is better to run generic, server or virtual? | 21:39 |
| infinity | lifeboy: It's all the same kernel. | 21:40 |
| lifeboy | So it's just the tools that get install with it that differ? | 21:40 |
| infinity | lifeboy: No, it's just that in the previous LTS, they were different kernels, so we provided transitional metapackages. | 21:41 |
| infinity | lifeboy: Installing "linux-server" gets you the linux-generic kernel. | 21:41 |
| lifeboy | So if install linux-image-generic I'll be fine | 21:41 |
| lifeboy | ? | 21:41 |
| infinity | Or linux-generic, if you want headers too. | 21:41 |
| infinity | But yes. | 21:41 |
| lifeboy | It will install linux-image-3.2.0-43-generic or linux-3.2.0-43-generic then. | 21:42 |
| infinity | linux-image-3.2.0-43-generic and linux-headers-3.2.0-43-generic if you install "linux-generic", and just linux-image-3.2.0-43-generic if you install "linux-image-generic" | 21:43 |
| infinity | Either way, having the metapackages installed will keep you up-to-date. | 21:43 |
| infinity | You want the headers if you also build out-of-tree modules (via dkms, for instance). | 21:43 |
| infinity | Like nvidia/fglrx. | 21:43 |
| lifeboy | There were quite a few thing broken running 3.7.7 Firefox couldn't find its profiles amongst others... | 21:44 |
| lifeboy | No, I don't need special modules | 21:44 |
| infinity | zequence2: All built and copied to -proposed. Should hit mirrors in ~1h. | 21:49 |
| zequence2 | infinity: thanks | 21:50 |
| infinity | BenC: FWIW, the "prepare-package*" tasks auto-complete when the package is built successfully in the PPA. Confuses the bot a bit if you set them Fix Released. ;) | 22:15 |
| infinity | BenC: Anyhow, cloning git and such to do a quick review, and if it's all good, I'll upload. | 22:15 |
| infinity | BenC: *poke* | 22:39 |
| infinity | BenC: Your changelog looks suspect. It's missing the last two master revisions (sure, I realise those end up being a no-op, but having the CVE and bug refs there is nice), and you should have a bit with your tracking bug at the top. | 22:42 |
| infinity | BenC: See the bit above [ Ubuntu: 3.8.0-22.33 ] in this changelog: https://launchpadlibrarian.net/140174115/linux-lowlatency_3.8.0-22.15_source.changes | 22:43 |
| BenC | infinity: I/m not sure what you mean | 22:50 |
| BenC | I did add the tracking bug to the top (I'm pretty sure) | 22:51 |
| BenC | infinity: Make sure you check debian.ppc/changelog and not something else | 22:51 |
| BenC | linux-ppc (3.8.0-10.16) raring; urgency=low | 22:51 |
| BenC | * Release Tracking Bug | 22:51 |
| BenC | - LP: #1181305 | 22:51 |
| BenC | etc... | 22:51 |
| ubot2 | Launchpad bug 1181305 in Kernel SRU Workflow prepare-package-meta "linux-ppc: 3.8.0-10.16-proposed tracker" [Medium,In progress] https://launchpad.net/bugs/1181305 | 22:51 |
| infinity | Hrm. | 22:52 |
| infinity | Crusty tag issue? | 22:52 |
| BenC | infinity: Hold on, maybe I didn't --force when I pushed --tags | 22:52 |
| BenC | # git push --force --tags | 22:52 |
| BenC | Everything up-to-date | 22:52 |
| BenC | Nope, should be good | 22:52 |
| infinity | Possibly not. Of course, that doesn't help if I already had the old tag. | 22:53 |
| infinity | There's no way to ask git to force others to update tags. | 22:53 |
| BenC | git fetch --tags > | 22:53 |
| BenC | ? | 22:53 |
| infinity | This was a fresh clone anyway, though... | 22:53 |
| infinity | So it should be right. | 22:53 |
| BenC | Yeah, because I pushed it last night | 22:54 |
| BenC | github tag perusing is less than optimal | 22:55 |
| BenC | https://github.com/benmcollins/ubuntu-raring-powerpc/blob/PPC-Ubuntu-3.8.0-10.16/debian.ppc/changelog | 22:55 |
| BenC | That definitely doesn't match what I pushed | 22:55 |
| infinity | Matches what I have here, though. | 22:56 |
| BenC | infinity: Fixed…my commit --amend didn't pick up my changelog revision | 22:57 |
| BenC | Other than that, the tree is essentially the same | 22:57 |
| infinity | BenC: Fixed where? :P | 22:59 |
| BenC | infinity: tag PPC-Ubuntu-3.8.0-10.16 | 23:00 |
| BenC | Fixing ppc branch now, but use the tag | 23:00 |
| infinity | Ahh, yeah, it wasn't on the branch, hence my confusion. | 23:00 |
| infinity | Floating tags drive me batty. | 23:00 |
| infinity | Anyhow, testbuilding here out of paranoia, but will upload to the PPA later today. | 23:01 |
| BenC | Ok, thanks | 23:01 |
| BenC | infinity: floating tag fixed for your sanity's sake | 23:02 |
| infinity | \o/ | 23:03 |
| infinity | Now I just need to wait 5 hours for this test build. | 23:03 |
| infinity | I need a machine like sagari in my bedroom. | 23:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!