16:58:24 #startmeeting RELENG (2018-05-24) 16:58:24 Meeting started Thu May 24 16:58:24 2018 UTC. 16:58:24 This meeting is logged and archived in a public location. 16:58:24 The chair is mboddu. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:58:24 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:58:24 The meeting name has been set to 'releng_(2018-05-24)' 16:58:24 #meetingname releng 16:58:24 The meeting name has been set to 'releng' 16:58:24 #chair nirik tyll sharkcz masta pbrobinson pingou puiterwijk maxamillion mboddu Kellin dustymabe 16:58:24 #topic init process 16:58:24 Current chairs: Kellin dustymabe masta maxamillion mboddu nirik pbrobinson pingou puiterwijk sharkcz tyll 16:59:04 * sharkcz is here 16:59:36 * mboddu waves at sharkcz 16:59:48 * sharkcz waves back :-) 17:00:32 Hello. Sorta here 17:02:02 * puiterwijk waves at sharkcz and mboddu 17:02:13 thx :-) 17:02:16 Anyone who comes in needs to continue the cycle 17:02:18 puiterwijk: Hey, long time no see, hopefully you got some free time 17:02:29 mboddu: haha.... 17:03:06 I dont have any tickets to discuss, but I have couple of things on open floor 17:03:14 So, lets start with alt arch updates 17:03:21 #topic Alternative Architectures updates 17:03:34 sharkcz: Few more days for the s390x koji instance 17:03:57 last ~5 days :-) 17:04:06 Hehe 17:04:09 Nope, I refuse to accept that. The cert is still valid for about 80 more days, so it'll stay up for that time 17:04:17 I didn't renew the cert last week to be useful for just two weeks! 17:05:04 puiterwijk: Hehe, but I would like to get that resources and add them to builders 17:05:27 mboddu: I'm fine with the builders moving. But the web definition will stay up until the cert expires, even if the backends disappear! 17:05:31 yes, that's the plan, I'll shutdown the existing builder 17:05:37 builders 17:05:47 * linuxmodder sneaks in 17:06:05 sharkcz: the plan is to add those directly to the primary hub, or to ask IT to reuse those resources for new VMs? 17:06:25 puiterwijk: reallocate the mem + cpu to the primary lpar 17:06:32 sharkcz++ 17:06:32 puiterwijk: Karma for sharkcz changed to 1 (for the f28 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:06:35 otehrwise we would the special network setup, etc 17:06:39 Yeah, indeed 17:06:44 That's why I hoped for that answer 17:06:51 ok :-) 17:07:00 s390x is special enough as it is :) 17:07:11 puiterwijk: True 17:07:15 should be just changing few numbers in the machine config 17:08:07 Cool 17:08:21 sharkcz: Awesome 17:08:22 then also the secondary sigul can go offilne and more 17:08:35 ^ puiterwijk will love that 17:08:39 Oh yeah. Down to a single instance of all the setups 17:08:39 I know ;-) 17:08:51 mboddu: well, I would've been fine with keeping it up if it were updated :) 17:09:02 Right now, secondary sigul is kinda outdated, so missing all the stability fixes :( 17:09:14 (I'd dare say that primary sigul is really quite stable these days) 17:09:36 puiterwijk: :) 17:10:12 The end of an era though 17:10:36 sharkcz: thank you for having taken care of it 17:10:36 right, but new, better, era has been already started :-) 17:10:46 Yes, fortunately or unfortunately I was not much involved in that era :) 17:10:47 puiterwijk: np 17:10:50 sharkcz++ 17:10:50 mboddu: Karma for sharkcz changed to 2 (for the f28 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:11:35 sharkcz: After next Tue I guess you will be working with infra in moving those builders to primary koji(or one koji at that point) 17:12:06 mboddu: ack, I plan an internal ticket to get the configs changed 17:12:14 sharkcz: Okay, thanks 17:12:22 sharkcz: Any other updates? 17:12:57 nope, everything else works, except some of the s390 stuff as you know 17:13:20 but having more memory and cpu could help the primary builders 17:13:31 to be more stable, faster 17:13:43 sharkcz: Awesome 17:14:44 #info F26 will go EOL next Tue and that marks end of s390 koji instance. sharkcz will work with infra teams in moving the resources to primary koji 17:14:45 we might reduce the old lpar to a bare minimum for experimenting with kvm based builder 17:15:00 instead of removing it fully 17:15:13 but will talk with Arlinton about it 17:15:51 sharkcz: Okay 17:16:32 puiterwijk: This reminds me, do you know whats going on with nodejs module builds? You said unavailable builder, but when I looked at koji web, that particular builder was active and ready 17:16:49 I didn't check it by ssh'ing into it though 17:17:17 mboddu: Which one? The one it was on when I moved it, or the one it got moved ot? 17:17:26 i.e. did you look before or after the ticket was filed? 17:17:51 I don't know what's going on, but it just stopped processing at all. Last time, sharkcz had to reset it via the console I think. 17:18:00 puiterwijk: I looked at it before it got reassigned 17:18:13 So -14? 17:18:37 puiterwijk: Yep, thats right 17:18:53 I think the whole lpar might have been overloaded, in such situation the vms stop responding for some periods and koj gets confused 17:18:59 Okay. I didn't look too long. ssh took over a minute to connect, then I decided it was dead, and just freed the task 17:20:12 puiterwijk: Okay, as I said, I didn't ssh into it, I just looked at the status of the builder 17:20:15 sharkcz: Oh okay 17:20:44 sharkcz: And then we have the base container problem on s390x 17:21:11 right 17:21:20 I will talk to Kevin about it later 17:21:42 is it always the base container or is it "some" random image? 17:22:53 when the hypervisor starts swapping, strange things can happen ... 17:23:27 sharkcz: Since the last two days its been failing due to the same reason 17:23:55 And, its on the base container image only as far as I can tell 17:25:04 ok 17:26:24 Anyway, moving on 17:26:35 #topic Open Floor 17:26:42 Anybody has anything? 17:26:51 Hi, since we are talking about architectures 17:27:08 Is there any news related to 32 bit support? 17:27:38 I know the SIG lives here: https://fedoraproject.org/wiki/Architectures/x86 17:28:30 And the F28 have it at least workstation 17:28:38 Not only the netinstall like F27 17:28:49 x3mboy: that's a fesco decision I think. 17:28:53 Mmmmm 17:28:56 Ok 17:28:58 Sorry then 17:29:12 Right now, the decision is to let things up to the i686 sig and if they don't pick things up enough, fesco will re-assess. 17:29:16 At least, that's what I think 17:30:40 Anything else? 17:32:56 Okay guys, thanks for joining 17:33:05 #endmeeting