21:00:17 #startmeeting scientific-sig 21:00:18 Meeting started Tue Oct 1 21:00:17 2019 UTC and is due to finish in 60 minutes. The chair is oneswig. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:22 The meeting name has been set to 'scientific_sig' 21:00:30 Afternoon all 21:00:33 O/ 21:00:43 hello everybody :) 21:00:50 #chair martial 21:00:51 Current chairs: martial oneswig 21:01:22 What's new? 21:02:20 I've been working on some follow-up for this project 21:02:34 #link federated Ceph for the Euclid space telescope https://indico.cern.ch/event/765214/contributions/3517141/attachments/1908894/3153554/2019-09-17-TelferGarbutt-Ad-hoc-Filesystems.pdf 21:03:00 Turns out our Ceph deployment is not at all balanced in its data distribution 21:03:15 I'm kinda here for the meeting but might get interrupted...so my one question is for martial. Is everything sorted for the SC19 BoF? I'll be there on a full technical pass this year. 21:03:18 had OSD occupancy ranging ~1% - ~94% 21:03:57 trandles: thanks for the reminder, had forgotten to buy a ticket, better put that on the list... 21:04:23 you know, I need to do that too :) 21:06:57 trandles: do you get your money's worth from the full pass? 21:07:08 martial: we don't need to submit a revised proposal or anything? 21:07:16 trandles: I ned to confirm attendance but beyond that I believe we have everything needed 21:07:19 oneswig: yeah, I'll be able to attend all the talks I like 21:07:46 Usually I'm on exhibitor and can't do anything but attend meetings and stay in the exhibit hall :( 21:08:19 trandles: I only went to SC16 previously and I think that's what I had too. 21:11:23 oneswig: so what about that OSD? :) 21:12:20 ha. I'm working hard to get this cluster reweighted to the point where it resembles balance. Still nowhere near. 21:13:05 I wonder if there's a way in Ceph to establish how big each PG is, some of these ones do seem quite chunky. 21:13:58 Got another interesting problem, in that the MDS has halted processing CephFS, possibly due to nearfull OSDs (hopefully - because that's something I can fix) 21:14:19 The MDS claims to be "behind on trimming" and ballooning in memory 21:15:57 still working on that bit. I was hoping jomlowe would have some wise words on it... 21:16:35 was hoping to ask Cloud Federation question with jmlowe as well 21:17:53 martial: what's up there? 21:18:46 mostly about metadata exchange solutions 21:19:02 and understand how xseed does it 21:21:42 Mike and Bob Budden gave a talk once on a way of using IIRC Murano to share application images between clouds... 21:24:23 might ask that of them next time I see them 21:24:40 OK, I think I need to shut down for the night... anything more to raise? 21:24:54 all good on my end 21:24:58 rbudden: PING! You're being talked about 21:25:12 ha, IRC ears are burning 21:25:41 OK I'm going to head off before he notices and corrects my slanders! :-) 21:25:42 I have nothing. Been working on Ironic stuff. We should have our new cloud engineer starting very soon. Hope to have some new developments to discuss in Denver. 21:25:59 trandles: good to hear you filled the post 21:26:08 Should be an exciting role, no doubt! 21:26:15 cool :) 21:26:20 Fingers crossed it goes to plan 21:27:02 btw trandles: am somewhat proud of this recent work on Ironic https://www.stackhpc.com/bespoke-bare-metal.html 21:27:37 Thx for the link, I'll read up 21:28:06 excellent. Until next time y'all 14:02:58 #endmeeting