[00:00:16] <MirahezeRelay>	 <mariosinmobile> nvm It's working now
[00:06:42] <MirahezeRelay>	 <posix_memalign, replying to posix_memalign> Still no updates for KockaAdmiralac's PR, though...
[02:20:32] <MirahezeRelay>	 <posix_memalign, replying to abaddriverlol> AFAIK Bucket doesn't have the option to use a separate DB. Isolation is achieved by doing most DB operations with a separate user (`$wgBucketDBuser`). Permissions are granted to the bucket user from the main DB connection.
[02:34:28] <MirahezeRelay>	 <abaddriverlol, replying to posix_memalign> My concern wasn't isolation but rather replication, since it is common practice in most extensions to use a replica (which is the primary DB if you don't have replication set up) for reading data, and only use the primary DB if data is being written
[02:34:36] <MirahezeRelay>	 <abaddriverlol> <https://mediawiki.org/wiki/Database_optimization#Typical_read_queries_should_query_replicas>
[02:43:21] <MirahezeRelay>	 <posix_memalign, replying to abaddriverlol> Ah I see. Yeah it seems to only connect to `DB_PRIMARY` and never mentions `DB_REPLICA`.
[02:44:45] <MirahezeRelay>	 <cosmicalpha> That is a semi-blocker (probably not necessarily a hard blocker though) for us to install as well since we intend to go to some replicas next year (unless something happens and our plan to do so fails)
[06:47:51] <MirahezeRelay>	 <cosmicalpha> @Technology Team I have a lot running on mwtask181 right now, perhaps it would be best if any scripts you need to run be done on another mwtask server as to not overload mwtask181. Thanks!
[06:48:23] <MirahezeRelay>	 <cosmicalpha> CC MacFan4000 IRC-side ^
[09:49:42] <MirahezeRelay>	 <rodejong> Anyone else experiencing a very slow loading Miraheze?
[09:52:31] <MirahezeRelay>	 <rodejong> [1/2] ```Error 503 Backend fetch failed via cp37 at Mon, 08 Sep 2025 09:51:36 GMT
[09:52:31] <MirahezeRelay>	 <rodejong> [2/2] Varnish XID 308117755```
[09:55:40] <MirahezeRelay>	 <rodejong> @paladox could you have a look?
[09:57:12] <MirahezeRelay>	 <paladox> works for me
[09:57:20] <MirahezeRelay>	 <paladox> and i see no problems on mw*
[09:57:38] <MirahezeRelay>	 <rodejong> hmm. I had it 3 times now
[09:57:52] <MirahezeRelay>	 <rodejong> weird
[09:58:22] <MirahezeRelay>	 <rodejong> [1/3] ```Web server is returning an unknown error
[09:58:22] <MirahezeRelay>	 <rodejong> [2/3] There is an unknown connection issue between Cloudflare and the origin web server. As a result, the web page can not be displayed.
[09:58:23] <MirahezeRelay>	 <rodejong> [3/3] Ray ID: 97bd8f4dad29daa7```
[09:58:43] <MirahezeRelay>	 <rodejong> Could this be a regional issue?
[09:59:02] <MirahezeRelay>	 <rodejong> (I'm in Denmark)
[10:18:04] <MirahezeRelay>	 <rodejong> It seems solved now
[12:23:26] <MirahezeRelay>	 <abaddriverlol, replying to posix_memalign> shouldn't be hard to fix in theory
[12:26:02] <MirahezeRelay>	 <posix_memalign, replying to abaddriverlol> Just need to identify read-only workloads and then use the replica database for these?
[12:53:36] <MirahezeRelay>	 <abaddriverlol> yep
[14:04:29] <MirahezeRelay>	 <snowfleaf> > An error has occurred while searching: We could not complete your search due to a temporary problem. Please try again later.
[14:04:34] <MirahezeRelay>	 <snowfleaf> adrenwiki, cirrussearch enabled
[14:59:04] <MirahezeRelay>	 <abaddriverlol, replying to snowfleaf> fixed
[23:57:53] <MirahezeRelay>	 <posix_memalign> Maybe it's a fluke during the update, but I'm seeing inconsistent preview behavior for DPL4.