=== gerald is now known as Guest98176
=== adam_g_out is now known as adam_g
[04:34] <apb1963> i'm on a 64 bit machine (ubuntu 14.04) and I just noticed that I do not have libc6-amd64 installed....  shouldn't it have been installed by something?
[04:56] <tmpRAOF> apb1963: No; the libc6 that you've got installed will be the 64bit version.
[04:56] <tmpRAOF> (Assuming you used the 64bit installer)
[09:25] <Odd_Bloke> Hello kernel friends; someone has pointed out that the latest image we pushed to GCE has linux-image-3.16.0-30-generic installed, but linux-image-extra-3.16.0-30 isn't available in the repos.
[09:26] <apw> Odd_Bloke, for utopic images ?  you are saying that the archive doesn't have a linux-image-extra for that version ?
[09:27] <apw> Odd_Bloke, how are they determining that there isn't one?
[09:27] <Odd_Bloke> apw: For trusty.
[09:27] <Odd_Bloke> We have linux-image-extra-3.16.0-30 3.16.0-30.40~14.04.1 installed.
[09:28] <apw> ? that is -extra ?
[09:28] <Odd_Bloke> Argh, copy-paste fail.
[09:28] <Odd_Bloke> We have linux-image-3.16.0-30-generic 3.16.0-30.40~14.04.1.
[09:29] <apw> I see that in the archive pool on archive.ubuntu.com linux-image-extra-3.16.0-30-generic_3.16.0-30.40~14.04.1_amd64.deb
[09:29] <apw> and a similar one for _i386, so it definatly exists
[09:29] <apw> as i say what are they typing that tells them they don't have it
[09:29] <apw> Odd_Bloke, ^
[09:31] <Odd_Bloke> apw: http://paste.ubuntu.com/9953137/ (immediately after an update)  This _could_ be a GCE mirror problem; so let me dig on that.
[09:32] <apw> Odd_Bloke, yeah, i am suspicious it is mirroring lag, but hten how did they get their -30 kernel ?
[09:32] <Odd_Bloke> Nope, GCE archives are resolving to archive.ubuntu.com still.
[09:33] <Odd_Bloke> apw: We built the image with the kernel on it.
[09:37] <apw> Odd_Bloke, that is a -proposed kernel ... so your apt-sources need -proposed enabled to get -extras at that version
[09:38] <Odd_Bloke> Oh, hm.
[09:39] <Odd_Bloke> apw: That makes sense; thanks for your help. :)
[09:39] <apw> Odd_Bloke, np :)
[09:42] <cristian_c> tseliot, where can I find the newest mainline kernel?
[09:42] <tseliot> cristian_c: http://kernel.ubuntu.com/~kernel-ppa/mainline/
[09:43] <cristian_c> tseliot, ok, but what one between them?
[09:44] <tseliot> v3.19-rc6-vivid
[09:44] <cristian_c> tseliot, ok, but:
[09:44] <cristian_c> <tseliot> in that case, upstream developers will only be interested in the upstream kernel, not in the ubuntu specific release (where we apply our own patches)
[09:49] <Odd_Bloke> apw: How long until that kernel is expected to migrate?  Wondering if suggesting this guy just waits is a viable option...
[09:50] <apw> Odd_Bloke, it is likely very soon, they were waiting on security signoff only i think last i looked
[10:25] <tseliot> cristian_c: those are unmodified upstream kernels, regardless of the ubuntu release name attached
[10:25] <tseliot> cristian_c: have a look at this: https://wiki.ubuntu.com/Kernel/MainlineBuilds
[11:04] <cristian_c> tseliot, ok
[12:01] <infinity> Odd_Bloke: Why are you building images against -proposed?
[12:01] <infinity> Odd_Bloke: Please, with sugar on top, stop that. :/
[12:57] <melmoth> Hola. On a 14.04 the hpsa module is in the linux-image-extra packages... How to have this module available during install time ?
[13:15] <Odd_Bloke> infinity: I'm trying to work that out at the moment.
[13:30] <apw> melmoth, from the CD installer ?
[13:30] <apw> melmoth, from the alternate or live ?
[13:31] <melmoth> good question; i m not the one makine the install apw, but i bet my customer is using the "regular" server cd
[13:32] <melmoth> looks like his drive is not recognised and according to the model he is using, wich is certified, the controller is suppose to use this module
[13:32] <apw> melmoth, anyhow being in the extra package is not a relevant question i don't think, as that does not define what is on the install image
[13:32] <melmoth> ahhhh.
[13:32] <melmoth> i was assuming whatever is in there is not in available during install time.
[13:33] <apw> melmoth, but we'd need to know what image they are using, so we can investigate
[13:33] <melmoth> ok. thanks.
[13:43] <apw> melmoth, right now we don't even know which series they are installing
[13:55] <smb> apw, If it is a 14.04 server cd image, the hpsa module would be in the block device udeb on cd. Right now I am not sure how to get the installer to ask you about it. Just trying again with expert mode enabled
[14:05] <infinity> smb: The installer isn't meant to ask at all, it should just install the udeb and probe.
[14:07] <smb> infinity, yeah and for utopic I seem to have seen something about block modules flying by. Though I had to manually udpkg the udeb before it would find the hpsa module on a different console
[14:08] <infinity> smb: The second half of that sentence doesn't really make much sense.  If the block-modules udeb wasn't unpacking, we'd have hundreds of bug reports.
[14:09] <infinity> Not to mention that elmo would kill me if hpsa wasn't working from the installer.
[14:10] <smb> infinity, really? ahci is built-in... anyway just what I see right now with vms. Something about failing to open the modules.buildin.bin when trying to modinfo
[14:10] <infinity> smb: Which CD are you using?
[14:11] <smb> infinity, once the trusty release one. now the trusty .1 server-amd64 and at some point the modinfo works now... confused
[14:13] <smb> infinity, in parallel I ran the utopic release server 64bit iso and still fail while the installer is asking about partitioning. And for that expert mode also seems ignored...
[14:15] <infinity> smb: I'm willing to believe utopic is confused.  Less willing to believe that about trusty, given how many hpsa controllers we have in the DC.  Unless literally all of them are still running precise.
[14:15] <infinity> smb: (And if utopic is broken, I'm also not going to fix it)
[14:16] <smb> infinity, yeah. all agreed. I think I want to repeat all my steps with a trusty cd (release and . release) without expert mode and see how those compare exactly
[14:16] <infinity> smb: Yeah.  If you can reproduce in a fully normal install, I want syslogs.
[14:17] <infinity> smb: Also try a trusty daily to see if we've fixed it by accident for 14.04.2
[14:18] <infinity> It's possible the missing modules.builtin is confusing something, but I doubt it.  I assume that's a red herring in your debugging.
[14:18] <smb> infinity, will do that if I find it a problem in .1
[14:31] <smb> infinity, Ok, pebkac... It helps if one does not move around those four letters. damn
[14:37] <smb> Bah, how I consistently typed hpsa here and hspa when it came to modinfo is beyond me
[14:37] <infinity> smb: Heh.  And the new name even makes sense!  HP Smart Array.
[14:37] <smb> apw, anyway the hpsa module is included if they use server cd images
[14:37] <Odd_Bloke> Right, I've worked out why we were building from proposed; it was so we could hit the GCE deadline and include the fix for bug 1390604.
[14:37] <ubot5> bug 1390604 in linux (Ubuntu Vivid) "Outbound TCP Throughput drops to zero for several drivers" [High,Fix released] https://launchpad.net/bugs/1390604
[14:37] <infinity> smb: As opposed to the old Compaq module which was something ridiculous with 37 Cs.
[14:38] <smb> was that that cciss thing?
[14:38] <infinity> smb: Yeah.
[14:38] <infinity> smb: I don't remember what it stood for.
[14:38] <infinity> smb: But same hardware, same driver, just pre-buyout rebranding.
[14:39] <smb> infinity, of course it does not say. still around for the other pci ids I would think
[14:39] <infinity> Odd_Bloke: Hrm, fair enough.  Were the images actually built with -proposed (ie: you may have gotten other -proposed packages in them too?), or was the kernel special-cased?
[14:39] <Odd_Bloke> infinity: Proposed was enabled, kernel was installed, proposed was disabled.
[14:40] <infinity> Odd_Bloke: Alright.  Still pretty icky, but I guess it could have been worse. :/
[14:40] <infinity> I'm going to break my "don't release on Fridays" rule to unbreak you.
[14:44] <Odd_Bloke> The fix we needed was in 3.16.0-28 and 3.16.0-29 is now what we'd get without proposed, so I'm going to stop using the proposed kernel in new GCE image builds.
[15:12] <apw> smb, thanks for looking into that
[15:45] <infinity> Odd_Bloke: -30 was just promoted, so if you hold off a bit, future images should be happy.
[15:46] <ogra_> how could it get into that image though ... given it wasnt in the archive 
[15:46] <infinity> ogra_: It was in proposed, and he was intentionally enabling proposed for kernels. :/
[15:46] <ogra_> sounds like an issue with the way this image was built 
[15:47] <ogra_> ah
[15:47] <infinity> ogra_: We've covered this.
[15:47] <ogra_> right, if it was intentional ... 
[15:50] <Odd_Bloke> infinity: Thanks!
[15:55] <Odd_Bloke> infinity: How can I track the promotion?  Is `rmadison linux-image-3.16.0-30-generic` my best bet?
[16:31] <infinity> Odd_Bloke: Yeah.
[17:00] <Odd_Bloke> infinity: Our GCE build process no longer uses proposed kernels, and we've pushed fixed images out the door. Thanks for your help!
=== swordsmanz is now known as hugbot