!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

386 Members
Next Infra call: 2024-07-11, 18:00 CEST (UTC+2) | Infra operational issues backlog: https://github.com/orgs/NixOS/projects/52 | See #infra-alerts:nixos.org for real time alerts from Prometheus.120 Servers

Load older messages


SenderMessageTime
31 May 2025
@nocturnalnerd:matrix.orgKyle RobinsonI'm getting 403 on hydra but my user agent is normal: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.3619:43:43
@hexa:lossy.networkhexaclaims to be Chrome/130.019:44:12
@hexa:lossy.networkhexaamong other things19:44:23
@hexa:lossy.networkhexawhich is from 2024-1019:44:52
@nocturnalnerd:matrix.orgKyle RobinsonThat's my browsers default useragent19:44:58
@nocturnalnerd:matrix.orgKyle Robinson2024 isn't even that old19:45:09
@hexa:lossy.networkhexafor browser that is ancient19:45:17
@hexa:lossy.networkhexabut you're probably using Safari, not Chrome, right?19:45:30
@nocturnalnerd:matrix.orgKyle Robinsonit's really not...19:45:31
@nocturnalnerd:matrix.orgKyle Robinsonvivaldi19:45:40
@hexa:lossy.networkhexaso what vivaldi version is this?19:46:52
@nocturnalnerd:matrix.orgKyle Robinson7.0.3495.18 (Stable channel) stable (64-bit) 19:47:38
@nocturnalnerd:matrix.orgKyle Robinsonim updating it now. waiting for homemanager to build19:47:50
@hexa:lossy.networkhexastable is 7.4.3684.43 from 3 days ago19:48:15
@nocturnalnerd:matrix.orgKyle RobinsonNow 7.3.3635.12 (Stable channel) stable (64-bit) • Updated input 'nixpkgs-vivaldi': 'github:nixos/nixpkgs/1807c2b91223227ad5599d7067a61665c52d1295?narHash=sha256-Pzyb%2BYNG5u3zP79zoi8HXYMs15Q5dfjDgwCdUI5B0nY%3D' (2024-12-22) → 'github:nixos/nixpkgs/78add7b7abb61689e34fc23070a8f55e1d26185b?narHash=sha256-XXILOc80tvlvEQgYpYFnze8MkQQmp3eQxFbTzb3m/R0%3D' (2025-05-28)19:50:43
@nocturnalnerd:matrix.orgKyle Robinsonhydra's good now19:51:46
@emilazy:matrix.orgemilybrowsers get bad CVEs on a timescale of weeks fwiw23:54:47
@emilazy:matrix.orgemilythere's a reason we backport their updates to stable even when they're backwards-incompatible23:54:58
1 Jun 2025
@arianvp:matrix.orgArianI was wondering. Could we reduce our AWS egress bill by having Hydra directly warm the fastly cache? Then we don't need to pay money for fastly fetching from S3 for the first time.10:39:02
@arianvp:matrix.orgArianWe can still keep S3 for archival. As ingress is free. But it would remove the egress fees AWS currently charges10:40:03
@arianvp:matrix.orgArianThis could literally be a fastly script that first fetches from the store hosted by the Hydra coördinator and if the file is not on disk falls back to S3 10:41:35
@arianvp:matrix.orgArianIt seems too simple to not have been tried before. Is there something I'm missing here?10:42:07
@vcunat:matrix.orgVladimír Čunát It surprises me that Fastly supports pushing into their cache. 10:43:14
@arianvp:matrix.orgArianI'm not saying we should push. I'm saying fastly should fetch from hydra instead of s3. And only fetch to s3 when hydra returns a 40410:43:44
@emilazy:matrix.orgemilyI guess the question is just whether Hydra can handle that load. and also how big its pipe to Fastly is?11:34:16
@edef1c:matrix.orgedef
In reply to @arianvp:matrix.org
I'm not saying we should push. I'm saying fastly should fetch from hydra instead of s3. And only fetch to s3 when hydra returns a 404
we could alternatively finish the Tigris story / move more towards s3 as pure archival storage
11:35:31
@edef1c:matrix.orgedefboth tigris and r2 are free egress, the norm is changing gradually11:36:06
@arianvp:matrix.orgArian
In reply to @emilazy:matrix.org
I guess the question is just whether Hydra can handle that load. and also how big its pipe to Fastly is?
It is already handling that load though? It's uploading to S3 after all.
11:41:11
@toonn:matrix.orgtoonn That load would be almost doubling, no? 11:48:45
@edef1c:matrix.orgedefno, we could run the pipeline through fastly, if we wanted to11:49:13

Show newer messages


Back to Room ListRoom Version: 6