15:02:10 <pwhalen> #startmeeting Fedora ARM & AArch64 Status Meeting
15:02:10 <zodbot> Meeting started Tue Jul  7 15:02:10 2015 UTC.  The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:10 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:10 <pwhalen> #chair pwhalen bconoboy pbrobinson dgilmore hrw jsmith kyle dmarlin
15:02:10 <zodbot> Current chairs: bconoboy dgilmore dmarlin hrw jsmith kyle pbrobinson pwhalen
15:02:20 <pwhalen> sorry i'm late
15:02:26 <pwhalen> whos here today?
15:02:45 * pbrobinson is here
15:05:25 <pwhalen> hmm, shall we wait for others?
15:05:39 <bconoboy> I'm hereish
15:05:45 <pbrobinson> may as well for a minute
15:06:17 <pbrobinson> damn that means we can't assign everything to blc.... jcm it is then!
15:06:31 <bconoboy> :-P
15:07:19 <pwhalen> :)
15:07:25 <pwhalen> lets start
15:07:32 <pwhalen> #topic 1) ==== Package Status & Issues  ====
15:07:43 <pwhalen> lets start with the good news
15:09:10 <pbrobinson> so from last week have qemu fixed, thankfully not a wider toolchain issue in the end, and openssl has come and was fixed this morning
15:09:39 <pbrobinson> so now koji-shadow is slowly winding itself up to get full on into the mass rebuild#
15:10:02 <pbrobinson> I think we're around 4-5 weeks behind as there's still some of the perl rebuild in there too
15:10:22 <pbrobinson> I had to bash a bunch of perl packages into submission over the last few days
15:10:39 <pwhalen> #info QEMU is now fixed. Openssl issue also resolved. Koji-shadow is now continuing. Roughly 4-5 weeks behind Primary.
15:11:10 <pbrobinson> but as it currently stands there's no known blocker packages
15:11:22 <pwhalen> nice.
15:12:44 <pbrobinson> also there might be golang landing shortly
15:13:03 <pbrobinson> I'm working with those maintainers to ensure we should have aarch64 (and power) from the outset
15:13:21 <masta> cool
15:13:37 <pbrobinson> and when the new nodejs and friends land we should also end up with v8 and all of his friends too
15:13:57 <pbrobinson> that should all happen between now and the end of the month..... in theory
15:14:03 <masta> that will be great
15:14:22 <bconoboy> pbrobinson: Any estimate on how long the mass rebuild will take now?
15:14:26 <bconoboy> (IE, if all goes well)
15:14:46 <pbrobinson> ROTFLOL..... no
15:16:08 <bconoboy> Am mostly wondering what the pending downtime means for its completion
15:16:21 <pbrobinson> and given that we're moving hosts and it looks like due to budget someone will have to walk them to PHX it looks like we'll have a reduced infra for a more extended period than planned so as always it'll be me that gets fucked over on that one
15:17:40 <bconoboy> We can probably stand up some temporary builders during their transit
15:17:49 <bconoboy> That's a different topic though- sorry pwhalen
15:18:25 <pwhalen> #topic 2) ==== Kernel Status ====
15:18:46 <pbrobinson> so 4.2rc1 has landed on primary
15:19:16 <pbrobinson> I might manually push a aarch64 later today or else in the next day or so unless k-s gets there first
15:19:29 <pwhalen> #info Latest - kernel-4.2.0-0.rc1.git0.1.fc23
15:19:59 <pwhalen> ive just booted it up on the bpi, no network, no usb . but it boots
15:20:03 <pbrobinson> I've had very little feedback on 4.1 in terms of either ARMv7 or aarch64 so I'm just going to assume that it's perfect for rebase!
15:20:23 * dgilmore has not tested yet
15:20:42 <pbrobinson> pwhalen: I've not had time to look at 4.2 yet
15:21:10 <pbrobinson> dgilmore: welcome, I have some other testing for you.... u-boot related, will sync later though
15:21:15 <pbrobinson> speaking of which.....
15:21:35 <pbrobinson> u-booot 2015.07rc3 is in rawhide
15:22:00 <pwhalen> thats next :)
15:22:02 <pbrobinson> it should now almost be on par featurewise with 2015.01
15:22:08 <pbrobinson> ah
15:22:09 <pwhalen> #topic 3) ==== U-Boot Testing ====
15:22:19 <pbrobinson> u-booot 2015.07rc3 is in rawhide
15:22:26 <pbrobinson> it should now almost be on par featurewise with 2015.01
15:22:46 <pbrobinson> there's know changes on console handling for some devices
15:23:26 <pwhalen> #info uboot-tools-2015.07-0.4.rc3.fc23 in Rawhide. Requires console handling for some devices.
15:23:30 <pbrobinson> which is an improvement on all AllWinner devices, possible regressions on other devices (need to specify console=
15:23:43 <pbrobinson> but please test it all
15:24:19 <pbrobinson> dgilmore: can you test wandboard on your B1/C1 revs and make sure it all auto detects OK, and also test cuboxi
15:24:33 <dgilmore> pbrobinson: sure.
15:24:52 <pwhalen> B1 is working, uses the right dtb
15:25:06 <pbrobinson> yes, I've tested by B1 too
15:25:22 <pbrobinson> dgilmore has both so he's a useful test person for comparison
15:25:37 <pbrobinson> and he has a cubox-i which now had upstream support
15:25:48 <dgilmore> when I am not in two meetings i will test
15:26:34 <pbrobinson> it also looks like Utilite support there is improving but I need to test more and we've had reports of it not fully booting so I need to debug that some more
15:27:07 <pbrobinson> upstrea, 2015.07 is due on the 13th
15:27:41 <pwhalen> #info U-Boot 2015.07 due to be released upstream on July 13th
15:27:56 <pbrobinson> so test and report issues please :)
15:28:17 <pwhalen> #topic 4) == Open Floor ==
15:29:15 <pbrobinson> so bconoboy did you have something about migrations
15:29:36 <bconoboy> yeah, just wondering what we want to do while hosts are en-route
15:29:53 <pbrobinson> dance? what do you suggest?
15:29:55 <bconoboy> if we can stand up some temporaries behind a reverse squid proxy is that useful?
15:30:17 <bconoboy> and if so, should we be doing that right now to help clear out the mass rebuild?
15:30:42 <pbrobinson> why no ship the devices we'd use as temp and then swap out the shipped ones to replace the temps
15:30:51 <pbrobinson> out the non shipped ones
15:31:12 <bconoboy> don't think I can do that- they're in a shared pool
15:31:30 <bconoboy> the goal here is to minimize downtime
15:31:30 <pbrobinson> I'd sooner not have to piss about with building temp ones and issues certs etc
15:31:37 <pbrobinson> it's lost time
15:31:54 <pbrobinson> my goal is to minimise my wasted time
15:32:01 <bconoboy> :-)
15:32:16 <dgilmore> there is a cost to bringing up the temp boxes
15:32:25 <dgilmore> not sure that outweighs the benefit
15:32:31 <bconoboy> we could stand them up at hte same time the builders ship, then use the systems' certs that are in transit
15:32:42 <pbrobinson> me either and I've said that from the outset
15:32:54 <pbrobinson> bconoboy: why can
15:32:57 <pbrobinson> why can
15:33:08 <pbrobinson> why can't we just overnight them?
15:33:37 <bconoboy> pbrobinson: I suppose we could, just have to arrange for a bigger transit budget
15:33:47 <bconoboy> it's still at least a half week of downtime.
15:33:59 <pbrobinson> well what is the person on a push bike cost vs overnight?
15:34:06 <bconoboy> arrival monday means shipment friday means derack and pack thursday
15:34:17 <pbrobinson> yes
15:34:40 <pbrobinson> and what extra time to do need to add to that for road freight #
15:34:41 <bconoboy> and even then there will be time to bring them back up in phx config
15:34:59 <pbrobinson> that's easy because I've already got that in ansible
15:35:02 <bconoboy> so we're really looking at a lot of downtime, even with an overnight shipment
15:35:16 <pwhalen> i can bring up new builders relatively easy
15:35:23 <pbrobinson> I can provision a buider in PHX in around 20 mins
15:35:48 <pbrobinson> bconoboy: we currently have 2 builders
15:35:53 <pbrobinson> in PHX
15:35:59 <bconoboy> also can't assume smooge will do aarch64 the moment he arrives.  So with an overnight shipment you have to assume they're still going to be out of commission for a week
15:36:00 <pwhalen> (I was referring to the temporary ones, if need be)
15:36:14 <pbrobinson> but you've still not answered the question of time for road frielght
15:36:23 <bconoboy> pbrobinson: I don't know.
15:36:30 <bconoboy> I'd guess a week of transit
15:36:43 <bconoboy> Not acquainted with shipping times between westford and phx2
15:36:48 <pbrobinson> so I can handle a Thurs -> Mon downtime
15:37:30 <pwhalen> im trying to clarify when they will be deracked and packed for shipping
15:37:37 <pbrobinson> someone needs to give me cost of overnight and a "X extra days" for the road freight
15:38:16 <bconoboy> sure, we can try and find that out
15:38:23 <pwhalen> suggestion is July 17th for shipping, deracked on the 16th (hot off the presses)
15:38:46 <pbrobinson> I would think to guarantee them we'd likely need to ship them a week before we want them to arrive
15:39:14 <pbrobinson> pwhalen: is that for road?
15:39:33 <pwhalen> not specified in the ticket, so i dont know
15:40:10 <bconoboy> in any case, it sounds like pwhalen is willing to setup some temporaries for the interim
15:40:17 <pbrobinson> OK, either way nearly 2 weeks outage isn't accceptible #
15:40:41 <bconoboy> so do we want to pursue that?  if so we'll line up the systems for it.
15:41:57 <pbrobinson> bconoboy: no, why can't we ship 6 spares from Boston and the two (or how ever many) from the other location (the later can ship when ever, it's only two) and then make the old builders spares in two weeks time
15:44:32 <bconoboy> pbrobinson: Because the systems in question aren't spares and aren't burn-in tested like the ones we are planning on shipping.
15:45:15 <bconoboy> We'll get some more details on the shipping options and revisit next week- that was my only topic for open floor
15:46:13 <bconoboy> anybody else?
15:46:45 <pbrobinson> bconoboy: it seems it needs to be sorted this week if they're being de-racked on the 16th
15:47:00 <bconoboy> agree
15:47:34 <pwhalen> i can coordinate with you both as it progresses
15:47:45 <pwhalen> anything else we want to cover today?
15:48:46 <pbrobinson> bconoboy: and I don't really think "burnt in" is a major issue, it's more can we swap 6 units on loan for 2 weeks so we don't need to take a big outage hit and I don't have to work all evenings and weekends between now and branching to attempt to deliver alpha close to mainline
15:49:23 <pbrobinson> bconoboy: I realise that isn't of a major concern to you but ultimately I care because it'll be me that's fucked over
15:49:49 <bconoboy> pbrobinson: We will endeavor to not waste any of your time
15:51:11 <pwhalen> going once..
15:52:14 <pwhalen> #endmeeting