From 32c77f7f1a6cb13b9ad984e14e0620da4e837039 Mon Sep 17 00:00:00 2001 From: NicholaiVogel Date: Wed, 1 Oct 2025 20:04:42 -0600 Subject: [PATCH] Truenas Maintenance logs 10-1-2025 - setting up nextcloud preferences after the operating system deleted itself again --- README.md | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 59aeb39..56f7965 100644 --- a/README.md +++ b/README.md @@ -2,4 +2,21 @@ ### A REPO OF FUCKING LOGS -(FUCK YOU) \ No newline at end of file +(FUCK YOU) + +--- + +## Events + +### **10-1-2025** - Truenas OS went down during S.M.A.R.T. tests, reinstalled and configured again. + +1. Installed new boot SSD +2. Set up users, samba, snapshots, backups, etc. +3. **Went through a full tuning pass:** + - Got Redis working by pulling the right password from `nextcloud-redis-creds`. + - Fixed persistence and eviction policies by setting env vars (`REDIS_APPENDONLY`, `REDIS_MAXMEMORY`, etc.) after discovering extra flags in args were breaking the Bitnami entrypoint. + - Once Redis confirmed `8GB allkeys-lru` with AOF enabled, I pivoted to Postgres. + - found performance limited by a `-c shared_buffers` override injected at container startup, which ignored both `postgresql.conf` and `ALTER SYSTEM`. + - Worked around this by testing direct SQL changes, editing configs, restarting the deployment, and finally confirming the cluster was still using command-line sources. + +**Results:** Redis is tuned and persistent across rollouts, Postgres is partially tuned but still needs the command-line overrides stripped from the deployment spec to honor the larger memory settings. \ No newline at end of file