[00:05] <roaksoax> bigjools: around?
[00:05] <bigjools> roaksoax: in body, perhaps
[00:05] <roaksoax> bigjools: heh :) just coming back to life?
[00:05] <lifeless> bigjools: the flesh is here the spirit is lacking
[00:05] <lifeless> ?
[00:06] <bigjools> I am sat at my desk but nobody is home!
[00:06] <bigjools> sleepless nights do that :(
[00:06] <roaksoax> boomer
[00:07] <roaksoax> bigjools: so there's a bug in packaging
[00:07] <bigjools> annoying, because it's the first night in ages where it's not been roasting hot
[00:07] <bigjools> roaksoax: /o\
[00:07] <bigjools> A BUG, OH NOES!
[00:07] <roaksoax> bigjools: when cluster controller puts MAAS_URL as http://localhost/MAAS
[00:07] <roaksoax> which breaks
[00:07] <roaksoax> pxe booting
[00:07] <bigjools> that would yes
[00:07] <roaksoax> (enlistment, commissioning)
[00:07] <roaksoax> bigjools: other than that tested maas/1.2 in raring
[00:08] <roaksoax> and works like a charm
[00:08] <bigjools> woohoo
[00:08] <bigjools> did syslinux get fixed?
[00:08] <roaksoax> bigjools: nah
[00:08] <bigjools> you used quantal's :)
[00:08] <roaksoax> bigjools: nah
[00:08] <bigjools> heh
[00:08] <roaksoax> bigjools: :) the only thing is that it didn't localboot
[00:08] <roaksoax> which is external from maas
[00:08] <bigjools> you used magic pixie dust
[00:09] <bigjools> ah
[00:09] <roaksoax> bigjools: which im gonna look at tomorrow
[00:10] <roaksoax> bigjools: now im gonna upload that to raring
[00:10] <roaksoax> bigjools: as soon as what i proposed for merging gets uploaded
[00:10] <bigjools> roaksoax: nice
[00:10] <bigjools> when's the next techboard meeting?
[00:10] <roaksoax> bigjools: now since lp:maas/1.2 is 1.2, does it make sense to upload as maas-1.2~bzrXYZ?
[00:11] <roaksoax> bigjools: yes after I upload, wait for tech board meeting
[00:11] <bigjools> roaksoax: +1 on the version
[00:12] <roaksoax> bigjools: so when I upload trunk, i was thinking on maas-1.2+bzrXYZ
[00:12] <bigjools> mmmmm
[00:12] <bigjools> I don't know
[00:13] <bigjools> you were talking about a version bump for raring
[00:13] <roaksoax> bigjools: yeah so i thought about this ~|+ since you mentioned you werent sure about the version bump
[00:13] <roaksoax> bigjools: we could call it 1.3~
[00:14] <bigjools> sounds ok
[00:14] <bigjools> we can make a 1.3 branch when releasing into 13.04
[00:14] <roaksoax> ok cool
[00:14] <roaksoax> i'll do that then
[00:14] <roaksoax> bigjools: so I'd just like https://code.launchpad.net/~andreserl/maas/ipmi_versioning_support_lp1086162 and https://code.launchpad.net/~andreserl/maas/ipmi_versioning_support_lp1086162_1.2 in
[00:14] <roaksoax> to upload to raring :)
[00:15] <bigjools> one sec!
[00:17] <roaksoax> bigjools: alright so it iwll be 1.2+ (instead of 1.2~)
[00:17] <bigjools> yeah makes sense
[00:19] <roaksoax> cool
[00:19] <roaksoax> alright so TB is next :)
[00:20] <roaksoax> bigjools: and if they accept we will make it before 12.04.2
[00:20] <bigjools> roaksoax: both approved with a small change
[00:21] <roaksoax> cool thanks
[02:45] <roaksoax> bigjools: the hjenkings issue seems unrelated
[02:45] <roaksoax> WebDriverException: Message: u'Message: \'Can\\\'t load the profile. Profile Dir: /tmp/tmpcmDMBI Firefox output: Xlib:  extension "RANDR" missing on display ":1014".\\n*** LOG addons.xpi: startup\\n*** LOG addons.xpi: checkForChanges\\n*** LOG addons.xpi: No changes found\\n\' \n-------------------- >> begin captured stdout << ---------------------\n    \n    Starting Firefox\n\n--------------------- >> end captured stdout << ----------------------'
[02:45] <bigjools> roaksoax: yeah
[02:45] <bigjools> it's broken yet again
[02:46] <roaksoax> there was a jenkins breakeage due to a security update (or so i heard)
[02:46] <roaksoax> bigjools: do you want me to fix the formatting issue?
[02:46] <bigjools> roaksoax: please
[02:47] <bigjools> I don't think this is related to that, it's having trouble starting an X frame buffer
[02:48] <roaksoax> bigjools: to the security update?
[02:48] <bigjools> I think so
[02:49] <roaksoax> yeah that's what I mean,t it might be related to the security update
[02:50] <roaksoax> bigjools: this has already been merged btw: https://code.launchpad.net/~andreserl/maas/ipmi_versioning_support_lp1086162_1.2/+merge/143194
[02:50] <bigjools> ok
[02:51] <roaksoax> bigjools: https://code.launchpad.net/~andreserl/maas/fix-formatting-issue-1.2/+merge/143226
[02:51] <roaksoax> https://code.launchpad.net/~andreserl/maas/fix-formatting-issue/+merge/143227
[02:51] <roaksoax> i'm approving them myself
[02:51] <bigjools> no worries
[02:51] <bigjools> I wonder what the heck jenkins and tarmac are doing then?
[02:52] <roaksoax> yeah
[02:52] <roaksoax> did it fail in trunk? or unle in 1.2?
[02:52] <bigjools> waaaat
[02:53] <bigjools> wtf
[02:53] <bigjools> it said it wasn't merging it, and then did!
[02:53] <roaksoax> yeah it did merge it
[02:54] <bigjools> not good
[02:54] <roaksoax> indeed
[02:54] <roaksoax> things went crazy
[02:55] <bigjools> I emailed matsubara
[02:56] <roaksoax> ack
[15:37] <roaksoax> rvba: howdy!! I was wondering what's that "generator" option in pserv.yaml
[15:42] <rvba> roaksoax: hi, let me check.
[15:42] <rvba> roaksoax: it's the url used by the TFTP server to contact the MAAS API to fetch the PXE configs.
[15:43] <roaksoax> rvba: ok cool
[15:43] <roaksoax> so there's a bug in packaging
[15:43] <roaksoax> on upgrade it replaces maas_cluster and sets MAAS_URL as localhost
[15:43] <roaksoax> localhost/MAAS
[15:43] <roaksoax> which cases PXE to break
[15:46] <rvba> roaksoax: Looks like a bug indeed… I'm surprised that Julian did not witness that when he did the upgrade testing…
[15:47] <roaksoax> rvba: so this should help: http://paste.ubuntu.com/1534585/ makes sense?
[15:48] <roaksoax> argh didn't work
[15:51] <roaksoax> i know why!
[16:24] <spideyman> Hi
[16:25] <spideyman> where does the config template path come from? /usr/share/provisioningserver/pxe?
=== matsubara is now known as matsubara-lunch
[16:42] <roaksoax> spideyman: huh?
[16:43] <roaksoax> spideyman: i think you want /usr/share/pyshared/provisioningserver/pxe
[16:43] <roaksoax> rvba: still around?
[16:44] <rvba> roaksoax: yep
[16:44] <roaksoax> rvba: could you pelase review: https://code.launchpad.net/~andreserl/maas/update_maas_url_on_upgrade_raring/+merge/143352
[16:44] <roaksoax> https://code.launchpad.net/~andreserl/maas/update_maas_url_on_upgrade_precise.sru/+merge/143350
[16:44] <roaksoax> https://code.launchpad.net/~andreserl/maas/update_maas_url_on_upgrade_quantal/+merge/143351
[16:44] <roaksoax> same change on all
[16:44] <roaksoax> has been tested works as expected
[16:44] <rvba> ok
[16:44] <roaksoax> rvba: thanks!! :)
[16:45] <spideyman> roaksoax: ah, thanks sorry
[16:56] <rvba> roaksoax: when does the bug manifests itself exactly?
[16:56] <rvba> roaksoax: simply when the cluster package gets upgraded?
[16:56] <roaksoax> rvba: yes
[16:57] <rvba> k, ta
[16:57] <roaksoax> rvba: say you have raring installed, and then you upgrade, maas_cluster.conf ends uip with MAAS_URL=http://localhost/MAAS
[16:57] <roaksoax> so when a node PXE boots
[16:57] <roaksoax> it tells it look for your preseed in url=http://localhost/MAAS
[16:57] <roaksoax> which is wrong :)
[16:59] <roaksoax> rvba: so the fix is basically 1. "update MAAS_URL always (install and upgrade from the debconf db)" or if nothing is set in 1. check whether the cluster has been installed alongisde the region, and use the value set in MAAS_DEFAULT_URL
[17:00] <rvba> All right, your fix seems sane but I'll recreate the problem first just for safety.
[17:00] <roaksoax> sure
[17:00] <roaksoax> rvba: the package that fixes this is in ppa:maas-maintainers/experimental
[17:01] <roaksoax> rvba: and I saw this after upgrading all my MAAS systems (2) and being unable to enlist
[17:07] <rvba> roaksoax: why did the upgrade change the original value?
[17:09] <roaksoax> rvba: the root cause of this is that in 1.2 branch and trunk the url used for the preseeds is the value in maas_cluster.conf (MAAS-URL), however in previous versions (currently released) thius is obtained from DEFAULT_MAAS_URL
[17:11] <roaksoax> rvba: this is the commit http://bazaar.launchpad.net/~maas-maintainers/maas/1.2/revision/1332
[17:16] <rvba> roaksoax: so if we're upgrading and the region controller is on another machine, the value of the url will be in te debconf db right?
[17:20] <roaksoax> rvba: yes
[17:21] <roaksoax> rvba: so on upgrade, it will be upgraded correctly
[17:21] <rvba> k
[17:30] <rvba> roaksoax: I'm sorry but I don't really understand the problem… if the cluster is installed alongside the region, then http://localhost/MAAS is a valid url to use to reach the MAAS server.
[17:31] <roaksoax> rvba: yes, now try to enlist a node
[17:31] <roaksoax> rvba: it will PXE boot, then it will fail to find the meta-data server
[17:31] <roaksoax> because the argument that passes the url for the metadata
[17:32] <roaksoax> will be http://localhost/MAAS/metadata/XYZ
[17:37] <rvba> I've installed a 1.2 package from scratch and I still have MAAS_URL=http://localhost/MAAS in /etc/maas/maas_cluster.conf
[17:37] <rvba> (I can't test enlistment just new because there is a problem with the lab)
[17:38] <roaksoax> rvba: ok so add ppa:maas-maintainers/experimental and sudo apt-get update && sudo apt-get upgrade
[17:38] <roaksoax> that should upgrade MAAS
[17:38] <roaksoax> and MAAS_URL should be http://XYZ/MAAS
=== matsubara-lunch is now known as matsubara
[17:44] <rvba> roaksoax: it's really weird, the url passed to the node is still computed using DEFAULT_MAAS_URL for the nodes in the master cluster's network.
[17:59] <rvba> roaksoax: I'm sorry but I need to think a little bit more about the bug you've fixed, I've summarized what I think about the problem on the MP but I need to step out now.  I'll have another look at this tomorrow.
[18:00] <rvba> roaksoax: sorry for the delay but maybe we will find a more elegant way to fix this.  I have to talk to Julian tomorrow morning about something else also we will have a look at your MP too.
[18:03] <roaksoax> rvba: cool, but i'm gonna upload this regardless
[18:03] <roaksoax> cause this "fix" only works when region controller is installed
[18:03] <rvba> roaksoax: ok
[18:04] <roaksoax> otherwise
[18:04] <roaksoax> upgrades will be broken
[18:04] <roaksoax> rvba: i'll catch up with julian about this tonight
[18:05] <roaksoax> rvba: but thanks for taking a look at it :)
[22:56] <spideyman> Hi
[22:57] <spideyman> I just updated from the latest quantal maas-maintainers daily build archive and I can no longer set kernel_opts for nodes
[22:58] <bigjools> spideyman: can you file a bug describing it please
[22:58] <spideyman> bigjools: I'm not sure if it's something I missed, I also don't see the global kernel paramaters Web UI option anymore
[22:59] <spideyman> bigjools: was there a change to this?
[22:59] <spideyman> bigjools: sure, although I'm not sure if it's a bug, or something I didn't update/install
[23:00] <bigjools> what version did you upgrade from?
[23:00] <spideyman> well, both before and after said 1.4.1
[23:01] <bigjools> the option is there somewhere, I can't remember offhand where though!
[23:02] <spideyman> it was here: http://localhost:8080/MAAS/settings/
[23:02] <spideyman> and tag.create(kernel_opts="<whatever>") isn't excepted
[23:03] <bigjools> if you are around later I can take a look