=== freeflying is now known as freeflying_away
=== freeflying_away is now known as freeflying
[04:30] <jtv> bigjools: are you free for a review by any chance?  It's https://code.launchpad.net/~jtv/maas-test/ssh-key/+merge/194488 — has two controversial points that I'd like your opinion on even if you can't do a full review.
[04:31] <bigjools> jtv: sure
[04:31] <jtv> thx
[04:38] <bigjools> jtv: reading your cover note, I see another bad side effect of using Mock.patch
[04:39] <bigjools> it's evil if not used carefully
[04:39] <bigjools> mocking out in setUp() is particularly crappy IMO
[04:43] <jtv> Yes, it's saddening how dangerous it is.  Thanks for spotting...
[04:48] <bigjools> jtv: in case you didn't see yet, I agree with Raphers on the ssh key
[04:58] <jtv> bigjools: yes, saw it — I thought /tmp made sense for avoiding leaving garbage after deinstallation.
[04:59] <bigjools> jtv: I don't think it's garbage in this instance
[04:59] <jtv> (And perhaps a bit of general principle, security-wise, but that doesn't really matter here)
[04:59] <bigjools> no, no security issues here
[05:09] <bigjools> jtv: so one controversy decided, and I am about to remove the other entirely
[05:12] <bigjools> jtv: read the last review email and let's see what you think
[05:16] <jtv> I just saw it...  Can assert_has_call do partial matches?
[05:16] <bigjools> yes
[05:16] <bigjools> use Mock.ANY
[05:17] <jtv> Ahhh...  Keep struggling with Mock's documentation
[05:17] <bigjools> it's not great
[07:20] <bigjools> would anyone care for a moderately simple review? https://code.launchpad.net/~julian-edwards/maas/ntp-server-dhcp/+merge/194659
[07:21] <bigjools> jtv or gmb?
[07:21] <jtv> Really want to get my branch back in shape here right now...  Need a bit more time to resolve conflicts etc.
[07:22] <bigjools> why why why is a blank line at the end of a file lint
[07:22] <bigjools> sigh
[07:23] <jtv> bigjools: it'll only cause diff pollution when it gets removed etc.
[07:23] <bigjools> then don't remove it?
[07:24] <jtv> Somebody will, at some point, perhaps by accident.
[07:24] <bigjools> and "pollution" ... ?
[07:24] <jtv> You know: start new class underneath, change your mind, remove the line, remove blank line...
[07:24] <jtv> Yes, that silly little change will show up in review diffs.
=== CyberJacob|Away is now known as CyberJacob
[07:35] <freeflying> 2013-11-11 07:31:48 ERROR juju supercommand.go:282 could not access file 'bootstrap-verify': gomaasapiot error back from server: 401 UNAUTHORIZED
[07:36] <freeflying> this was after I destroyed a mchines, and redeploy it
=== CyberJacob is now known as CyberJacob|Away
[10:13] <gnuoy> allenap, thanks for the response to Bug#1249435 , I've updated it in return. I think you're right about the cause but it points at a bug in the way the maas cli define kernel_opts for tags when none are explicitly defined.
[11:02] <allenap> gnuoy: I saw that, ta :) At core it's an issue in the database field definition; the field can be set to either null or '', but other code assumes that only null equals undefined, but we should consider '' as undefined too. Better would be to prevent one of those possibilities.
[11:03] <gnuoy> allenap, thanks, I think the bug is still valid then
[11:03] <allenap> gnuoy: Definitely; I've just triaged it to High.
[11:03] <gnuoy> tip top
=== freeflying is now known as freeflying_away
=== freeflying_away is now known as freeflying
[13:09] <AskUbuntu> How to make my nodes online in maas? | http://askubuntu.com/q/375412
=== CyberJacob|Away is now known as CyberJacob
[20:47] <Azendale> bigjools: Is now a good time to take it up again?
[21:44] <wintrmute> anyone know if its safe to edit /etc/resolv.conf directly?  http://maas.ubuntu.com/docs/configure.html#manual-dhcp says to do it but the file comments say changes will be overwritten
[21:46] <wintrmute> nevermind, its using resolvconf
[21:53] <Azendale> wintrmute: if you haven't already figured out, /etc/resolvconf/resolv.conf.d/head
[21:54] <wintrmute> Azendale, yep, just did that.  but it adds the server to the end of the list.  how does resolv.conf work, does it fail-over if a name doesnt resolve with the first nameserver?
[21:57] <wintrmute> Azendale1, yep, just did that.  but it adds the server to the end of the list.  how does resolv.conf work, does it fail-over if a name doesnt resolve with the first nameserver?
[21:57] <Azendale1> wintrmute: I believe so
[21:57] <wintrmute> kk ty
[21:57] <Azendale1> wintrmute: did you regerate the config by running sudo resolvconf -u
[21:58] <wintrmute> Azendale1, i restarted the resolvconf service via /etc/init.d
[21:58] <wintrmute> the resolv.conf file shows the head
[21:58] <wintrmute> or at least i ran the resolvconf init script
[22:00] <Azendale1> wintrmute: resolvconf is about generating the file /etc/resolv.conf
[22:00] <Azendale1> wintrmute: I don't remember the exact way /etc/resolv.conf works, but that's probably because I haven't had much trouble. man resolv.conf should tell you about the format
[22:01] <wintrmute> Azendale1, thanks.  i just read the man page, i think it works the way we're describing
[22:02] <Azendale1> wintrmute: np, did that get it working for you?
[22:03] <wintrmute> Azendale1, not sure yet, ill let you know ;)  setting up the maas region controller and first cluster atm
[22:03] <Azendale1> wintrmute: ok, sounds good
[22:04] <wintrmute> i have an existing dhcp/dns server for my LAN and im having the region controller run dhcp/dns for pxe nodes i start... just have to get the two to play nicely together
[22:05] <wintrmute> i defined a PXE class for both and have one server ignore the class and the other allocate for it
[22:05] <wintrmute> at least for dhcp.  dns should hopefully "just work" with the proper order of nameservers in resolv.conf
[22:06] <Azendale1> wintrmute: so, the classes are based on mac address I assume?
[22:06] <wintrmute> Azendale1, nah i just defined it by PXE boot.  i dont use that for anything but maas nodes.  this is a home lan.
[22:06] <wintrmute> i can add macs of course if necessary
[22:06] <wintrmute> or use mac prefixing
[22:08] <Azendale1> wintrmute: I'm not an expert. Just when I first started with MaaS I had two dhcp servers conflicting, and of course the nodes would not PXE boot. But now you know if you have that problem to just try one DHCP server at a time :)
[22:11] <wintrmute> Azendale, yep.  and it works ;)
[22:12] <Azendale> wintrmute: awesome. I'll probably keep that in mind next time I set up MaaS
=== freeflying is now known as freeflying_away
[23:42] <Azendale> bigjools: Are you there? Is now a good time to work on the MaaS DHCP debugging we were doing Friday?
[23:42] <bigjools> Azendale: hi, I am here but a little busy.  how long are you around?
[23:44] <Azendale> bigjools: I'm pretty flexible. I'm in the UTC-0800 timezone and should be up til about 23:00, (so that would be 0900 UTC I think)
=== freeflying_away is now known as freeflying
[23:45] <bigjools> 0700 :)
[23:45] <bigjools> ok will ping you a bit later
[23:45] <Azendale> bigjools: ok, sounds good (my brain isn't good at time math yet)
[23:46] <bigjools> I've been doing it a few years now....
=== CyberJacob is now known as CyberJacob|Away
[23:46] <bigjools> not to mention I went from +0000 to +1000