| 27 Oct 2025 |
 emily | someone participating in a public torrent swarm is asking a public database (torrent tracker) to offer their IP to the general public in order for people to access services (downloading chunks / peer exchange / …) from it | 17:58:21 |
| emily | (of course, whether people realize that clicking download is essentially volunteering to act as a public internet server is another matter, which I can imagine the GDPR caring about…) | 17:58:51 |
 tgerbet | Google is not a natural person and even if you can consider that consent was given to the torrent tracker, it does not extend to the other parties that might process the information (collection is considered a processing)
And yes I agree the whole thing is weird for P2P context | 18:21:36 |
| emily | then I guess we should hydraPlatforms = [ ]; it? though that won't stop people using it in packages. | 18:31:52 |
| emily | AFAICT it would require turning down the log level all the way to not showing any warnings or info messages at all to get rid of the IP logging. | 18:32:12 |
 toonn | IP addresses can be personal data, but if they cannot be traced back to a specific natural person then it's not personal data. | 18:55:01 |
| toonn | If it's only timestamps and IPs I doubt that would count as personal data. | 18:55:19 |
| toonn | Other factors that come into this are whether the treatment is reasonable. Logging IPs every hour for years doesn't seem reasonable. A log every couple days for a month or so seems pretty reasonable. | 18:56:18 |
| emily | I'll leave the decision about how to handle it to others I think | 19:41:47 |
| emily | since it's not totally obvious to me if we want hydraPlatforms = [ ] or a config.inHydra throw or a postFetch making the FOD not big or making sure we don't have IPs in the logs or what | 19:42:37 |
| toonn | My vote is for none of those, it doesn't matter. At least IP logging-wise. Caching Sintel over and over on Hydra is different consideration. | 19:44:16 |
| emily | yeah I meant in terms of cache use too | 20:48:58 |
|  ➡️@amadaluzia:unredacted.org joined the room. | 23:56:29 |
| 28 Oct 2025 |
 Mic92 | hexa (signing key rotation when): I would like to test those earlier commits on Wednesday: https://github.com/NixOS/hydra/issues/1535#issuecomment-3438017869 | 05:06:17 |
| Mic92 | * hexa (signing key rotation when): I would like to test those earlier hydra commits on Wednesday: https://github.com/NixOS/hydra/issues/1535#issuecomment-3438017869 | 05:08:06 |
|  @hsngrmpf:matrix.org left the room. | 06:32:42 |
 vcunat | Go ahead, I'd say. Is there branch for the infra repo for this? | 07:27:21 |
| vcunat | * Go ahead, I'd say. Is there a branch for the infra repo for this? | 07:27:29 |
 hexa (signing key rotation when) | 
Download image.png | 20:22:35 |
| hexa (signing key rotation when) | i can at least confirm that my hydra has the same problem and this is how it looks | 20:22:45 |
| 29 Oct 2025 |
| Mic92 | @hexa:lossy.network: just to confirm, this is happening on latest git ref or the one John suggested? | 06:13:00 |
| Mic92 | In reply to @vcunat:matrix.org
Go ahead, I'd say. Is there a branch for the infra repo for this? Not yet. | 06:13:33 |
| hexa (signing key rotation when) | Yeah, latest git | 10:28:51 |
| Mic92 | hydra rebooting... | 19:32:29 |
| Mic92 | back | 19:35:01 |
| 30 Oct 2025 |
| Mic92 | hydra still seems fine on https://github.com/NixOS/hydra/commit/94eaad22bc68c199ac1f6939b5dae65a618e550b. Would you like to test https://github.com/NixOS/hydra/commit/58846b0a1c896c55f3897f45b363ad2ba8a53ed5 in your hydra? Otherwise this would be the next commit I would test. | 11:47:07 |
| hexa (signing key rotation when) | can do | 11:47:24 |
| hexa (signing key rotation when) | Mic92: 58846b0 looks fine | 12:32:08 |
| Mic92 | hexa (signing key rotation when): what was the meeting url again? | 17:03:46 |
 dgrig | https://meet.cccda.de/nix-osin-fra | 17:04:06 |
