| 13 Sep 2025 |
 emily | yes, pretty significant. | 16:39:49 |
| emily | more cores, but even single-core is significantly better | 16:40:18 |
 @emma:rory.gay | ah | 16:40:26 |
 lennart | what should the uplink look like for ofborg? wie have 1000 down, 200 up at the office, could also drop it in a datacenter with so called embedded hosting and symetric 1000 uplink | 16:42:34 |
| emily | no idea, #ofborg:nixos.org more likely to know. some of the build machines were in Graham's basement for the longest time, so I'm guessing there's no stringent bandwidth requirements. | 16:43:32 |
| emily | not sure if they even push things to a shared cache or just use the local store (we don't offer their build outputs publicly since they're untrusted) | 16:43:48 |
| lennart | top | 16:43:55 |
| emily | data centre with remote hands etc. may be very nice if it's convenient though | 16:44:09 |
| emily | but I'm guessing upload throughput won't be a big consideration | 16:44:43 |
| lennart | I meant this offer by a German foundation, media.ccc.de, codeberg.org etc also hosts there. but well Speedbone not like a Digital Realty data center :b no remote hands as well, but its nearby https://in-berlin.de/provider/serverhousing/#/ | 16:46:35 |
 [0x4A6F] | Yeah, know these. There are good people. | 16:49:46 |
| 14 Sep 2025 |
 Bot_wxt1221 | Are we solving the access problem on wiki.nixos.org? | 04:04:43 |
| Bot_wxt1221 | 
Download 图像.png | 04:04:46 |
 Vladimír Čunát | If I read the repo right, those with (ssh) access are (only) Mic92 and lassulus. Given how the reachability check behaves over the last couple of days, I'd guess that it's simply overloaded, possibly due to a surge in scraping. | 06:32:05 |
 Mic92 | I did start with some fastly config. I wish there was more prior art because caching for complex applications can be tricky | 06:33:46 |
| Mic92 | For now we could also scale up the VM as well | 06:37:45 |
| Mic92 | I could maybe adapt https://www.mediawiki.org/wiki/Manual:Varnish_caching | 06:42:07 |
 ners | Hello, as members of the EC, @RossComputerGuy and I would like to request vault access to set up SMTP for the election. Is this the right place to ask? | 14:07:43 |
 Jeremy Fleischman (jfly) | I don't know if the password is stored in Vaultwarden. IMO, it would be better for it not to be stored there and instead for the EC to generate a new password for each election. There are instructions for that here: https://github.com/NixOS/infra/tree/main/non-critical-infra/modules/mailserver#mailing-lists | 14:25:14 |
| Mic92 | Nginx-based fastcgi caching seems to do the trick. | 17:58:07 |
| 15 Sep 2025 |
 raboof | Likely not something that has a quick fix, but perhaps something to take into account next year: people seem to be running into session expiration issues with the survey: https://snug.moe/notes/acnkov8yosyrs62c | 11:23:45 |
| 16 Sep 2025 |
|  jopejoe1 (4094@39c3) joined the room. | 20:33:11 |
|  Ihar Hrachyshka joined the room. | 20:35:54 |
| jopejoe1 (4094@39c3) | Hey release manager for 25.11 here would like to get permissions for hydra needed for Release management as documented in https://github.com/NixOS/infra/wiki/Hydra-Accounts, thanks | 20:39:08 |
| 17 Sep 2025 |
| Vladimír Čunát | done | 04:27:42 |
| jopejoe1 (4094@39c3) | Thanks | 05:44:08 |
| lennart | big oof, it's a 2018 Intel :( | 11:06:36 |
| Mic92 | hexa (signing key rotation when): BMG want's to join the infra meeting tomorrow to show his progress on his gc infra. | 11:13:09 |
 hexa | sure, why not | 11:13:46 |
| hexa | the meeting is open anyway | 11:13:49 |
