| 20 May 2026 |
 K900 | They are currently copied to the coordinator for signing | 12:20:52 |
 Cobalt | Oh, well, that does sound really expensive for bandwidth (and compute with compression). | 12:21:27 |
| Cobalt | * Oh, that does sound really expensive for bandwidth (and compute with compression). | 12:21:33 |
| K900 | It is | 12:21:47 |
 vcunat | And S3 is over the Atlantic. | 12:21:53 |
| vcunat | There are lots of "design problems". | 12:22:09 |
| Cobalt | Interesting, I will have to take a closer look at it then (especially if my knowledge of the architecture is apparently inaccurate).
I will still probably stick with scheduling for now due to other constraints on the topic but thank you for the extra information. | 12:24:03 |
| Cobalt | * Interesting, I will have to take a closer look at it then (especially if my knowledge of the architecture is inaccurate).
I will still probably stick with scheduling for now due to other constraints on the topic but thank you for the extra information. | 12:25:18 |
 Mic92 | In reply to @c0ba1t:matrix.org
Maybe I misunderstood something there but aren't they uploaded directly from the builder to the S3 bucket? The new queue runner has this as a feature flag. | 13:46:36 |
| vcunat | For reference, some jobs were canceled temporarily to avoid advancing channels which follow nixpkgs master. https://matrix.to/#/!SsuQHVzf59iBgk6YIm:0upti.me?via=nixos.org&via=matrix.org&via=tchncs.de are working on a workaround to a regression. | 16:25:17 |
| 21 May 2026 |
| K900 | Are the channel scripts OK? | 10:19:07 |
| K900 | The job has been running for like 15 minutes now | 10:19:13 |
| vcunat | Looks OK to me. | 10:21:16 |
| vcunat | Doing debuginfo now. | 10:21:25 |
| vcunat | (uploads of redirects to S3) | 10:21:40 |
| vcunat | Maybe this step isn't parallelized enough, as S3 has latency 200-250 ms per operation. | 10:22:57 |
| vcunat | * Maybe this step isn't parallelized enough, as S3 has latency 200-250 ms per operation (goes overseas). | 10:23:17 |
| vcunat | * Maybe this step isn't parallelized enough, as S3 has latency 200-250 ms per operation (.de -> USA). | 10:23:32 |
| vcunat | nixos-unstable-small advanced now. | 10:26:27 |
| vcunat | For reference, we're still plagued by repeated S3 abortions:
https://hydra.nixos.org/build/329544227#tabs-buildsteps | 11:48:16 |
| vcunat | * For reference, we're still plagued by repeated S3 upload abortions:
https://hydra.nixos.org/build/329544227#tabs-buildsteps | 11:48:26 |
 Sergei Zimmerman (xokdvium) | Eh, shall the old queue runner be resurrected on some other hydra branch so that we could fix this by adding more retries in nix? | 12:06:32 |
| Sergei Zimmerman (xokdvium) | Since hydra master doesn’t seem to be ready at all and there’s need to fix this in the meantime. Probably the wrong channel for this though | 12:08:21 |
| vcunat | We're at
github:NixOS/hydra/a40d42862da88cce78a27dd594e1484a034aac4d
| 12:13:32 |
| vcunat | coupled with github:NixOS/nix/2.34-maintenance | 12:14:45 |
| vcunat | effectively we have github:NixOS/hydra/infra | 12:16:15 |
| Sergei Zimmerman (xokdvium) | Hm ok so if we add more retries to 2.34 branch it would help | 12:16:37 |
| vcunat | You don't need to push it into that branch. It suffices to create a commit on that branch (to avoid too many changes) and we can switch to it easily. | 12:17:29 |
| vcunat | * You don't need to push it into that branch. It suffices to create a commit on that branch (to avoid too many changes) and we can switch to it easily. | 12:17:45 |
| Sergei Zimmerman (xokdvium) | Ok, I’ll try to get to that this/next week | 12:19:20 |
