| 20 Jun 2026 |
 Mic92 | Okay that's hopefully an easy fix | 06:00:42 |
| Mic92 | I published my fast-nix-gc fixes I came up when I had to battle with the 1TB nix store / 58GB sqlite nix db: https://github.com/Mic92/fast-nix-gc/pull/45 | 06:04:46 |
| Mic92 | Deploying to staging now | 06:30:33 |
 Vladimír Čunát | _I'll be away from a computer for most of today._ | 06:51:05 |
| Vladimír Čunát | * I'll be away from a computer for most of today. | 06:51:20 |
| Mic92 | QA passed for the fixes. I will now proceed production hydra. | 07:08:28 |
| Mic92 | * QA passed for the fixes. I will now proceed with production hydra. | 07:09:46 |
 K900 | Maybe don't | 07:09:47 |
| K900 | It's been building a kernel for 7 hours | 07:09:56 |
| K900 | We might want to let it finish | 07:09:57 |
| Mic92 | K900: build number? | 07:10:20 |
| K900 | There's a bunch | 07:11:49 |
| K900 | Just search for "linux-" on /status | 07:12:03 |
| Mic92 | https://hydra.nixos.org/build/331695810/nixlog/1 matches the description | 07:13:12 |
| Mic92 | If it's not finished in 90 min, I will deploy | 07:58:18 |
| Mic92 | I also figured out why kernels are not build on non-big-parallel machines | 07:59:05 |
| Mic92 | The json parser for derivations wasn't taking structured attrs into account | 07:59:37 |
| Mic92 | * I also figured out why kernels are build on non-big-parallel machines | 07:59:48 |
 hexa | ugh, kernel builds on 2 cores are no fun | 08:08:19 |
| Mic92 | Classic bug. I did report a similar issue a month ago to nixbuild.net | 08:09:40 |
| hexa | Redacted or Malformed Event | 08:10:47 |
| hexa | I don't think we want to under these circumstances | 08:11:08 |
| hexa | if we restart it gets a big-parallel build slot and will be done much quicker | 08:12:01 |
| Mic92 | Okay, than lets roll | 08:17:29 |
| hexa | do you need any help deploying? | 08:17:49 |
| Mic92 | No. I have all closures already cached | 08:18:05 |
| hexa | cool | 08:18:08 |
| Mic92 | Fix works: https://hydra.nixos.org/machines | 08:40:22 |
| Mic92 | hopeful-rivest now gets the linux build instead | 08:40:34 |
| hexa | also looks like jobs stay much shorter in receiving outputs now | 08:51:08 |
