!sBfrWMVsLoSyFTCkNv:nixos.org

OfBorg

172 Members
Number of builds and evals in queue: https://ofborg.org/prometheus/graph?g0.expr=ofborg_queue_evaluator_waiting&g0.tab=1&g0.stacked=0&g0.show_exemplars=0&g0.range_input=2h&g1.expr=ofborg_queue_builder_waiting%7Barch!~%22.*-lowprior%22%7D&g1.tab=1&g1.stacked=0&g1.show_exemplars=0&g1.range_input=2h62 Servers

You have reached the beginning of time (for this room).


SenderMessageTime
19 Jun 2023
@cole-h:matrix.orgcole-hOK, should start shrinking now. Sigh.13:59:28
20 Jun 2023
@ryantm:matrix.orgryantmr-ryantm has been waiting 4 hrs for ofborg to evaluator.messages to get low enough to make a PR.23:32:28
@hexa:lossy.networkhexahave evaluations become much more costly and/or has the pr volume grown so much23:34:21
@cole-h:matrix.orgcole-hMaybe related? https://github.com/NixOS/nixpkgs/issues/22794523:35:24
@cole-h:matrix.orgcole-hI'll see about scaling up the number of machines tomorrow23:36:35
21 Jun 2023
@vcunat:matrix.orgVladimír Čunát Yesterday I reverted some changes that even lead to crashes due to too expensive evaluation, but I don't know if that's really related. (commit 205ee073b05) 12:35:38
@wegank:matrix.orgWeijia joined the room.13:16:45
@cole-h:matrix.orgcole-hAdding one machine shortly, may add another if it doesn't shrink the queue as much as I'm hoping.15:56:30
@raitobezarius:matrix.orgraitobezariusI don't think ofborg uses nix-eval-jobs, right?16:10:41
@raitobezarius:matrix.orgraitobezariusI feel like it has better controls for memory consumption of each workers16:10:49
@cole-h:matrix.orgcole-h Correct, it uses plain ole nix-eval 16:13:02
@cole-h:matrix.orgcole-hSheesh, even with 9 total evaluators, it's still not shrinking.....17:27:36
@cole-h:matrix.orgcole-hGuys, I'm announcing a temporary pause on Nixpkgs PRs please stop 😭 (Kidding, it'll settle down eventually...)17:37:40
@sandro:supersandro.deSandro 🐧hold my beer17:43:29
@cole-h:matrix.orgcole-h * Guys, I'm announcing a temporary pause on Nixpkgs PRs please stop 😭 (Kidding, it'll settle down eventually...) (I hope...............)17:43:43
@raitobezarius:matrix.orgraitobezariusI think trying nix-eval-jobs can mitigate the problem in a serious fashion18:16:26
@raitobezarius:matrix.orgraitobezariusI cannot hack a PR right now though18:16:35
@raitobezarius:matrix.orgraitobezariusIt's probably time to do a serious look on cgroup accounting for nix workers18:17:14
@hexa:lossy.networkhexaopened a PR23:56:16
@hexa:lossy.networkhexaofborg was the first check I saw23:56:27
@hexa:lossy.networkhexacloning23:56:31
@hexa:lossy.networkhexa😭23:56:43
@hexa:lossy.networkhexaI want this speed always! 😄23:56:56
@hexa:lossy.networkhexaheh, did I misread? Now it says pending agin23:57:29
@hexa:lossy.networkhexa * heh, did I misread? Now it says pending again 😲23:57:32

Show newer messages


Back to Room ListRoom Version: 6