[11:07] <jtv> rvba: just having a look at bug 1092265...  one thing I don't understand is that the node's log mentions "raring" all over the place, not "precise."  I thought you said the node was being set up with Precise?
[11:07] <ubot5> bug 1092265 in MAAS "juju setup fails on raring" [Critical,Triaged] https://launchpad.net/bugs/1092265
[11:08] <jtv> Oh, that log is not from the node, is it?
[11:08] <rvba> jtv: it's booting precise.  What log?
[11:08] <rvba> The Jenkins' log?
[11:09] <rvba> That's the report from the installation of the MAAS server… on raring.
[11:09] <jtv> Yeah...  I was looking at the wrong log there.
[11:10] <rvba> jtv: I'm trying to debug this in the lab right now… so far, no luck…
[11:10] <jtv> The console screenshots seem to suggest that the TFTP request times out, or fails in some other way, and the node falls back to local boot.  Is that normal?
[11:11] <rvba> That's not what I see.  I see the request go through ok (i.e. reach the webserver fine).
[11:12] <jtv> The webserver?  But it's supposed to go to the cluster controller, right?
[11:13] <rvba> Node -> pserv -> webapp and then back
[11:13] <jtv> Oh, depending on whether it's the config or the actual image of course...
[11:14] <jtv> So what you get to see on the webserver is the cluster controller's API request for PXE config information.
[11:14] <rvba> Yes
[11:15] <rvba> This is the pxe config used: http://paste.ubuntu.com/1491555/
[11:15] <rvba> Looks fine.
[11:17] <jtv> Do we know chain.c32 is still OK?
[11:18] <rvba> No.  It definitely looks like the problem is in there.
[11:19] <rvba> Any idea how to debug this?
[11:19] <jtv> TFTP server logs.  See if the requests _after_ the one for the config are happening.
[11:21] <jtv> I guess we'd expect requests for chain.c32, the kernel, and initrd.  And here's hoping they didn't change the naming convention for those again!
[11:25] <rvba> rarg, I need to clean up all the debugging statements I put in there first…
[11:27] <rvba> http://paste.ubuntu.com/1491568/ That's the log
[11:28] <rvba> No sign of the other files being transferred.
[11:30] <jtv> I do see chain.c32 being mentioned...  but not the images  :(
[11:30] <rvba> Indeed
[11:33] <rvba> jtv: If I replace the file chain.c32 with the one from the syslinux package on *quantal*, the boot seems to work.
[11:34] <jtv> Gah
[11:40] <jtv> rvba: the syslinux in Quantal is 2.x, and upstream is at 4.x.  Good chance of a major change in Raring I guess.  :(
[11:41] <rvba> Judging by the size of the files only, there is a major change.
[11:41] <jtv> I think all you can really do at this point is punt to the server peoples.
[11:41] <rvba> Yep, I've added a comment on the bug.
[11:42] <jtv> And I just got my marching orders — time to hop on the scooter and scoot over to the other village.
[11:44] <jtv> Let's just hope the new syslinux doesn't require any painful changes on our end.  See you tomorrow!
[11:44] <rvba> jtv: why do you say that the version of syslinux is 2.x in quantal, it says 2:4.05+dfsg-6
[11:44] <rvba> ?
[11:44] <rvba> http://paste.ubuntu.com/1491595/
[11:44] <jtv> Oh!  Argh!  I read the "2" and ignored the rest.
[11:44] <jtv> Upstream is at... 4.06?
[11:45] <rvba> http://paste.ubuntu.com/1491596/
[11:45] <rvba> The versions are not that different.
[11:45] <rvba> But still, we have a problem :)
[11:45] <jtv> Clearly!
[11:46] <rvba> Anyway, I'll poke the server people.
[11:46] <jtv> What does the "2:" stand for, in the version number?  I probably should know this, but it's not the sort of thing I deal with regularly any more.
[11:46] <rvba> I don't know either.
[11:48] <jtv> Oh well.  The upstream wiki did say something about a major change.
[11:48] <jtv> In 4.06.
[11:48] <jtv> And I really must run — nn!
[11:48] <rvba> nn jtv
[15:49] <rvba> roaksoax: smoser:  Hi guys.  Would you mind having a look at bug 1092265?  I think I identified where the problem is (see my comments on the bug) but this looks like something you guys would know how to fix it.
[15:49] <ubot5> bug 1092265 in MAAS "Nodes fail to boot from local disk on raring" [Critical,Triaged] https://launchpad.net/bugs/1092265
[16:32] <roaksoax> robbiew: i'll do some testing but seems problem of newer chain.c32 as you mention
[16:32] <roaksoax> so there's probably a bug in a code or who knows what would it be
[16:33] <robbiew> huh?
[16:33] <rvba> roaksoax: cool, thanks for looking into it.
[16:36] <roaksoax> robbiew: err sorry :) was for rvba :)
[16:37] <roaksoax> rvba: maybe something changed in newer syslinux and template needs some additional info in order for it to boot
[16:38] <rvba> roaksoax: the pxe config used is this: http://paste.ubuntu.com/1491555/
[16:40] <roaksoax> rvba: yeah seems right
[16:40] <rvba> And simple :)
[16:40] <roaksoax> rvba: i have to re-setup environments so it will take me a few hours
[16:43] <rvba> roaksoax: all right.  I'll be off soon so please send me an email or comment on the bug if you make progress or want me to test things.
[16:44] <roaksoax> rvba: sure thing
[16:44] <rvba> ta
=== ppetraki is now known as ppetraki-busy
[17:32] <alperkanat> hey there.. i'm trying to test maas with Ubuntu Server 12.10 on virtualbox.. there are 2 vm's each have 2 nics on themselves. 1 is open to public network (internet) and the other is open to private network (192.168.57.x with no DHCP server)
[17:33] <alperkanat> i've enabled DHCP/DNS on MAAS
[17:33] <alperkanat> the node seems to find PXE
[17:33] <alperkanat> but TFTP gives a timeout
[17:33] <alperkanat> any ideas?
[18:01] <roaksoax> a/win 14
=== ppetraki-busy is now known as ppetraki
[21:46] <bstillwell> I'm working on setting up MAAS on quantal and when attempting to PXE boot a node, it gets hung up trying to download pxelinux.cfg/default
[21:46] <bstillwell> Any idea what I might have missed?
[21:47] <bstillwell> I've configured MAAS to handle both dhcp and dns.
[21:49] <bstillwell> I don't see a /var/lib/maas/tftp/pxelinux.cfg at all.
[22:07] <bstillwell> What part of MAAS creates pxelinux.cfg?
[22:29] <roaksoax> bstillwell: check that /etc/maas/maas_local_settings.py has the correct IP for MAAS_URL
[22:31] <bstillwell> roaksoax: looks good
[22:34] <bstillwell> I did find one issue with the docs so far.  When running maas-import-pxe-files, you may need to set both http_proxy and https_proxy.
[22:34] <bstillwell> maas-import-ephemerals uses https
[22:35] <bstillwell> Oh, and this page points to other pages that don't exist:
[22:35] <bstillwell> http://www.ubuntu.com/download/help/install-ubuntu-cloud
[22:35] <bstillwell> specifically steps 4 & 5
[23:00] <bstillwell> ahh, figured it out
[23:00] <bstillwell> I disabled i386 and armhf so I wouldn't have to download the largish ephermal images.
[23:00] <bstillwell> Looks like i386 is needed
[23:03] <bstillwell> looks like the node is trying to download stuff from archive.ubuntu.com now, but it isn't using the proxy like it should...
[23:03] <bstillwell> Is that configurable somewhere?
[23:11] <bigjools> it's set in the preseed
[23:11] <bigjools> by default it uses a local squid-deb-proxy
[23:12] <bstillwell> Where's the preseed file at?
[23:16] <bstillwell> found them: /usr/share/maas/preseeds
[23:22] <bstillwell> This message needs to be updated:
[23:22] <bstillwell> You can boot this node using Avahi-enabled boot media or an adequately configured dhcp server. See https://wiki.ubuntu.com/ServerTeam/MAAS/AvahiBoot for instructions.
[23:22] <bstillwell> That link doesn't work
[23:29] <bigjools> gah
[23:29] <bigjools> I think a lot of these bugs were fixed, it's just not released yet
[23:30] <bigjools> see https://launchpad.net/maas/+milestone/12.10-stabilization
[23:33] <bstillwell> looks like lots of fixed bugs.  :)
[23:57] <bstillwell> So on this page it says the updated juju packages will be in main after the release of precise:
[23:57] <bstillwell> https://maas.ubuntu.com/docs/quantal/juju-quick-start.html
[23:57] <bstillwell> seems like it should be updated too