[01:00] <bigjools> jtv: quick pre-imp?
[07:13] <rvba> jtv: gmb: The following (taken from the most recent failure in the test lab) looks like a fallout from the recent 'label' work: http://paste.ubuntu.com/7118396/
[07:14] <gmb> rvba: Damn.
[07:15] <gmb> rvba: I need to step out for a little while right now but I'll glom onto this as soon as I get back. Best to email it to jtv too as he knows more about the label stuff than I.
[07:16] <rvba> gmb: okay.  I'll have a look now;  let's talk about it when you're back.
[07:17] <gmb> rvba: Cool, thanks.
[07:17]  * gmb -> back in a wee while
[07:33] <rvba> gmb: Could you please have a look at: https://code.launchpad.net/~rvb/maas/bug-import-eph-crash/+merge/211673
[08:16] <bigjools> rvba: TypeError: install_image_from_simplestreams() takes at least 4 arguments (4 given)
[08:16] <bigjools> lol?
[08:16] <bigjools> :)
[08:20] <bigjools> rvba: want me to review that?
[08:21] <rvba> bigjools: sure
[08:22] <rvba> bigjools: that's if you rather do that than talk to me :)
[08:22] <bigjools> rvba: ok let's try and see if 3g holds up
[08:22] <bigjools> I'll call you
[08:34]  * gmb returns
[08:34] <gmb> Thanks for fixing that rvba. Sorry I had to be afk for the duration.
[08:34]  * gmb didn't time it that way. Promise.
[08:37] <rvba> gmb: no worries
[08:39] <rvba> gmb: building a package with the latest trunk now
[08:44] <gmb> rvba: Is there any card from the new import script work that you'd like me to pick up in particular, or should I just grab something at random?
[09:07] <bigjools> rvba: did my change fix ipmi in the lab?
[09:09] <rvba> bigjools: I was precisely in the process of testing this when I came across bug 1294516.
[09:09] <rvba> bigjools: now that it's fixed, I'm testing the package again.
[09:10] <bigjools> rvba: cool.  One more thing, unrelated, but when there's no boot images the architecture drop down is empty and there's no message saying why
[09:11] <bigjools> remember the work we did for the power types?
[09:11] <bigjools> that is missing in the arch dropdown
[09:11] <bigjools> I added a card as this is a little more serious
[09:11] <bigjools> as it will trap many people
[09:11] <bigjools> I have to travel now, I will be late for the call later
[09:11] <rvba> Okay.
[09:12] <bigjools> probably at the top of the hour instead
[09:12] <bigjools> hope you don;t mind waiting for me
[09:12] <bigjools> :)
[09:12] <rvba> :)
[09:53] <rvba> bigjools: the int. tests passed with the new package.  Starting a new test run just for safety…
[10:28] <bigjools> \o/
[10:28] <jtv> Why is "maas-region-admin dbshell" broken again?  I thought I'd fixed that.
[10:28] <bigjools> jtv: yeah me too, I saw that earlier :/
[10:29] <rvba> bigjools: Second run is not over yet but it's passed the point where it failed before.  I think we can consider the bug fixed.
[10:29] <rvba> gmb: I'm going to work on the boot images stuff… so you're can grab the "Architecture drop-down empty and no message why (see description)" task if you want :)
[10:30] <gmb> rvba: WFM.
[10:30] <jtv> bigjools: Ahh, it's the --installed option.
[10:32] <bigjools> jtv: whaaaa?
[10:33] <jtv> bigjools: run dbshell with the --installed option.
[10:33] <jtv> We still need a user-friendly error, don't we?
[10:37] <rvba> Do you guys think it makes sense to deal with boot images in the API globally on only on a cluster per cluster basis?
[10:37] <rvba> I mean, is it useful to have a way to list all the boot images?  Or do we want the API to operate on a nodegroup object?
[10:39] <rvba> In one case, listing bootimages is done with /api/boot-images/op=list
[10:39] <rvba> In the other case, it's /api/nodegroups/<ng_uuid>/boot-images/op=list
[10:42] <jtv> Think in batches.
[10:42] <jtv> I'd just make nodegroup a constraint on a global nodegroups listing.
[10:42] <jtv> Although... this isn't very scale-sensitive.
[10:42] <jtv> Never mind me.
[10:43] <rvba> Not really sure which option is best.
[10:44] <rvba> Nodes objects are defined "globally" (although they are related to nodegroups) because it was important to be able to manage them as a set.
[10:44] <jtv> In terms of the functionality we have in mind, having a separate list per nodegroup seems the most appropriate.
[10:44] <rvba> Interfaces are defined on nodegroups.
[10:46] <jtv> I think it all depends on how likely we are to need a cross-nodegroup listing.
[10:47] <jtv> For the uses we have, we plan to move away from that.
[10:47] <rvba> Yeah, defining this globally only gives us that: a cross-nodegroup listing.
[10:48] <rvba> Or rather "gives us only that"
[10:48] <jtv> If we need it in rare cases, I don't think we need to worry about a client having a few extra round trips to make.
[10:48] <jtv> Either "only" is correct.
[10:50] <rvba> All right, I'll define this only the clusters objects.  It's definitely a bit cleaner to do it this way.
[14:24] <zchander> HI, is anyone familiar with preseed settings for MaaS? I want to partition a disk while commissioning
[15:18] <perrito666> hello fine people, is anyone available to review a maas related juju-core change? https://codereview.appspot.com/77850043 there are a couple of lines with doubts and you might have the answer
[15:22] <mgz> rvba: ^ pls :)
[15:23] <rvba> perrito666: sure, I'll have a look in a sec.
[15:23] <perrito666> rvba: thank a lot
[15:23] <perrito666> *thanks
[15:51] <rvba> perrito666: I think I replied to your question, see the MP.
[15:52] <perrito666> rvba: thank you,checking
[16:16]  * gmb -> afk for a while
[16:47] <perrito666> rvba: your changes got applied, thanks for the input
[16:48] <rvba> np
[16:58] <tych0> hi rvba, is there a way to add and remove tags via the UI?
[17:01] <rvba> tych0: no, you have to use the API/CLI.
[17:01] <tych0> ok, cool