| 3 Jul 2025 |
 Christian Theune | hexa (clat on linux when): i'm debugging some scalability issues with the ACME unit structure we're having. after a couple of interesting bits I now face https://github.com/NixOS/nixpkgs/pull/81371#issuecomment-605526099 and would like to consider alternatives to adding the renew services to multi-user.target. Specifically, you mention that you'd like changes to be applied, however, in our situation this always causes service restarts and massive (local) lego calls that add up in unfortunate ways - even if nothing has changed. aside from bolting on an option whether to do this or not, I'm wondering whether we can change it in a way that will leverage the regular "this service has changed, lets restart it". I'm currently playing around with remainafter exit, but that causes other things to explode. It's a bit of a minefield ... I'm available to chat basically the whole day if you'd like to give some input on this ... | 05:09:22 |
 @emma:rory.gay | this might be offtopic for here, but what's the mostefficient way to mount a remote filesystem? | 09:56:50 |
| @emma:rory.gay | currently using sshfs and its slow as molasses (and doesnt keep its parallel connections alive it seems) | 09:57:11 |
 K900 | NFSv4 is decently fast | 09:58:18 |
| @emma:rory.gay | (latency is probably the biggest blocker here) | 09:58:20 |
| @emma:rory.gay | does it support clientside caching? | 10:01:28 |
| K900 | Yeah | 10:02:32 |
| @emma:rory.gay | ooh neat, guess ill still be dealing with running a few recursion commands to preload data, but sounds like it might be less painfully slow than sshfs | 10:03:11 |
| @emma:rory.gay | okay, so if i undertand it right, the format of /etc/exports is like this, right?
/path/to/exported/dir client_ip(flags)? | 10:11:31 |
| K900 | Don't we have properly shaped options for that | 10:12:25 |
| @emma:rory.gay | the server isnt on nixos ^^' | 10:12:34 |
| K900 | Oh | 10:12:38 |
| K900 | Then yeah something like that | 10:12:52 |
| @emma:rory.gay | alright, thanks <3 | 10:13:03 |
| @emma:rory.gay | mount: /mnt: fsconfig() failed: NFS: mount program didn't pass remote address., not sure what im doing wrong | 10:21:51 |
| @emma:rory.gay | here's the command im trying: sudo mount full-server-dns-name:/exported/path /mnt | 10:22:27 |
| @emma:rory.gay | using mount.nfs4 however just seems to hang | 10:30:26 |
| K900 | Firewall? | 10:31:02 |
| @emma:rory.gay | already forwarded the port on the server end | 10:31:14 |
| @emma:rory.gay | at least assuming i didnt fail horribly at reading the docs... TCP/2049 | 10:31:46 |
| @emma:rory.gay | oh, didnt start the nfs4-server service... i completely missed that given you need to run a separate command to update the exports list | 10:33:00 |
| @emma:rory.gay | argh, kernel compile time, turns out the server's kernel doesnt have nsfd | 10:33:58 |
| K900 | Maybe also try ksmbd | 10:34:33 |
| K900 | If you have that | 10:34:38 |
| @emma:rory.gay | oh it needs a reboot... lol | 10:34:50 |
| @emma:rory.gay | ... assuming i didnt brick the system, at least | 10:35:44 |
| @emma:rory.gay | cool, it lives, except the database | 10:37:04 |
| @emma:rory.gay | okay, unless i did something wrong, nfs4 seems to end up slower than sshfs? | 10:44:51 |
 @saiko:knifepoint.net | afaik that is to be expected, it does a lot more stuff like locking and whatnot | 10:47:02 |
| @saiko:knifepoint.net | every serious network file system that tries to avoid file corruption is going to be slower than sshfs | 10:47:27 |
