19 Sep 2024 |
K900 | Because I restarted it | 13:44:07 |
K900 | And no, not the same eval | 13:44:11 |
emily | (also two people separately reported seeing it before it hit Hydra? but it doesn't really matter if it happening on Hydra is going to regularly stop updates reaching users without manual restarts) | 13:44:24 |
Mic92 | How many restarts did we have so far? | 13:45:15 |
K900 | Three, twice on x86_64-linux and once on aarch64-linux | 13:45:52 |
K900 | But that's at least one per eval | 13:45:56 |
Mic92 | Always .simpleUefiSystemdBoot? | 13:46:11 |
K900 | Yes | 13:46:29 |
K900 | But Dawn and Puck hit it on unrelated expression | 13:46:37 |
K900 | * But Dawn and Puck hit it on unrelated expressions | 13:46:38 |
Mic92 | I don't have Dawn's or Pucks configuration, so I cannot take this into account. | 13:47:10 |
K900 | Also note that we don't have it on unstable-large yet | 13:47:25 |
emily | In reply to @lily:lily.flowers For what it's worth, i'd been experiencing list corruption errors in CI as well since 2.23~2.24 (i'm not sure when exactly) almost daily in https://github.com/lilyinstarlight/foosteros before i finally gave up CI'ing against cppnix at all. I've no clue if it's related, but here if anyone wants to see an example of a CI run that failed with nonsensical list mis-evaluation weeks ago that then succeeded on a subsequent rerun (and this was happening constantly so i can provide multiple examples. i forget how varied the examples were): https://github.com/lilyinstarlight/foosteros/actions/runs/10822449565/job/30026445898#step:6:4654 there was this, which is public and maybe related | 13:48:34 |
emily | In reply to @joerg:thalheim.io Always .simpleUefiSystemdBoot? this is just about -small right? where that boot test is ~the most elaborate NixOS test it's building | 13:49:14 |
emily | I'd expect more on the large channel | 13:49:19 |
Lily Foster | In reply to @emilazy:matrix.org there was this, which is public and maybe related Possibly unrelated mis-evaulation, idk. I didn't dig into it. But i can provide a ton of those gha logs showing misevals 🤷🏻‍♀️ | 13:49:22 |
Mic92 | Lily Foster: which of the hosts? | 13:49:30 |
emily | (ironically the small channel didn't build simpleUefiSystemdBoot until recently) | 13:49:42 |
K900 | Opened https://github.com/NixOS/nixpkgs/pull/343032 | 13:49:49 |
K900 | In reply to@emilazy:matrix.org this is just about -small right? where that boot test is ~the most elaborate NixOS test it's building Notably Hydra itself is still on 2.19 | 13:49:59 |
K900 | So there will not be much evaluating on 2.24 | 13:50:08 |
K900 | The installer tests are special because they run an eval in the VM | 13:50:19 |
K900 | Most tests don't | 13:50:23 |
emily | In reply to @k900:0upti.me Notably Hydra itself is still on 2.19 (this does need fixing as apparently only 2.18 and 2.24 are supported…) | 13:50:29 |
K900 | And that makes me wonder if it's related to memory pressure | 13:50:44 |
Lily Foster | In reply to @joerg:thalheim.io Lily Foster: which of the hosts? the one i linked above was github:lilyinstarlight/foosteros#checks.x86_64-linux.host-bina with nix 2.22.1. i thought it was a newer version though, that specific one may have been fixed by now | 13:50:50 |
K900 | I'm running a nixpkgs-review on the PR because there might be other changes we need to revert | 13:51:20 |
emily | In reply to @emilazy:matrix.org (this does need fixing as apparently only 2.18 and 2.24 are supported…) (cf. https://matrix.to/#/!RROtHmAaQIkiJzJZZE:nixos.org/$VWPyj_hLCQXhDvslicySSM1XCH-Nt8gmysJB73d0UAU?via=nixos.org&via=matrix.org&via=nixos.dev) | 13:51:04 |
K900 | @emily re: the fallback-paths commit, it's just to make the reverts cleaner | 14:08:18 |
Lily Foster | * the one i linked above was github:lilyinstarlight/foosteros/02fc624bf177f479420c587964d3a3d21b61b5bd#checks.x86_64-linux.host-bina with nix 2.22.1. i thought it was a newer version though, that specific one may have been fixed by now | 13:51:34 |