Linux
3.1.12
I have installed a Joplin server (latest version) on my synology Nas and as soon as it is enabled in docker the server writes on the hard-drive at ~1MB/s therefore I constantly hear my hard-drives writing noise. This happens even though I am not syncing with the app (even with the network cable of my nas disconnected).
Is there a way to reduce or stop this permanent writing ?
Thank you for your help
If you look at the log, what is the server doing?
Here is a log extract. But the taskservice #11 occure every 10s and for less than a second. Whereas the disk activity while the Joplin server is up is continuous. I does not seems to explain the constant writing on disk.
20:51:08 0|app | 2024-09-16 20:51:08: TaskService: Completed #11 (Process shared items) in 645ms
20:51:17 0|app | 2024-09-16 20:51:17: TaskService: Running #11 (Process shared items) (scheduled)...
20:51:18 0|app | 2024-09-16 20:51:18: TaskService: Completed #11 (Process shared items) in 448ms
20:51:27 0|app | 2024-09-16 20:51:27: TaskService: Running #11 (Process shared items) (scheduled)...
20:51:28 0|app | 2024-09-16 20:51:28: TaskService: Completed #11 (Process shared items) in 480ms
20:51:37 0|app | 2024-09-16 20:51:37: TaskService: Running #11 (Process shared items) (scheduled)...
20:51:38 0|app | 2024-09-16 20:51:38: TaskService: Completed #11 (Process shared items) in 463ms
20:51:47 0|app | 2024-09-16 20:51:47: TaskService: Running #11 (Process shared items) (scheduled)...
20:51:48 0|app | 2024-09-16 20:51:48: TaskService: Completed #11 (Process shared items) in 1006ms
20:51:57 0|app | 2024-09-16 20:51:57: TaskService: Running #11 (Process shared items) (scheduled)...
20:51:58 0|app | 2024-09-16 20:51:58: TaskService: Completed #11 (Process shared items) in 432ms
Hmm, but that wouldn't produce 1MB/s. It's every 10 seconds, and no data should be generated except for this log statement
You confirm that it is not an expected behavior (the constant writing on disk) ?
I will try to find out what is written on disk.
No it's not expected, but I don't know what could be causing it
I have a file in the the joplin/pg_wal directory that is updated every 10s but it's size remains 16Mo.
File name is 00000001000000000000007A.
Could WAL be the culprit?
16Mo every 10s is coherent with the writing rates I get from the nas performance monitoring tool.
This is a file that's managed by Postgres and it's indeed going to be written to every time there's a log statement. I don't know how your server is setup but generally you shouldn't mirror this file since it's to manage current transactions, not for data.
I am not sure it's the raid mirroring in fact just the 16Mo every 10s is coherent with the writing rates I get.
Is there a way to prevent log to activate the update of that WAL file?
Or to disable log
Not currently and I'm not sure it's something I'd want to add support for at this point.
Is there a way to select a log level to limit log to errors?
Not currently, no. The list of available env variable is there: joplin/packages/server/src/env.ts at dev · laurent22/joplin · GitHub
But anyway the server logs to stdout so you could just send everything to do /dev/null if you wanted, but I don't think this is causing any disk activity. More likely it's the service to handle the shared notebooks that is saving its internal state, and that is necessary for it to work so it can't be configured.
OK thank you for your help anyway
Possibly related? High disk read of joplin-server - #3 by jade12