- Timestamp:
- 08/21/2026 08:00:05 AM (3 weeks ago)
- Branches:
- main
- Children:
- bb9e54c
- Parents:
- 1294c36
- File:
-
- 1 edited
-
deploy/MULTI-INSTANCE.md (modified) (1 diff)
Legend:
- Unmodified
- Added
- Removed
-
deploy/MULTI-INSTANCE.md
r1294c36 r0f09302 210 210 | `klonkt-update` | update the code, restart all instances | 211 211 | `ls /var/lib/klonkt` | which instances exist | 212 | `cd /opt/klonkt && npm run reset-admin -- --instance <slug>` | reset a forgotten admin password | 213 214 ### Why the password reset needs `--instance` 215 216 Every other command here names its instance, and this one is no different — but 217 the reason is easy to miss, because the app's own help page says only "run 218 `npm run reset-admin` from the project folder". That advice is written for the 219 one-instance layout, where the `.env` sits next to the code. 220 221 Here it does not. The code is shared and read-only at `/opt/klonkt`; the 222 configuration lives at `/var/lib/klonkt/<slug>/.env`, and **systemd** reads it 223 through `EnvironmentFile=`. A script started by hand gets none of that: it looks 224 for an `.env` beside the code, finds nothing, and is left without 225 `DATABASE_PATH`. `--instance` reads the same file systemd does. 226 227 Without it the script now stops and says which path it wanted, instead of 228 creating an empty database inside the shared code directory. If your data lives 229 somewhere other than `/var/lib/klonkt`, set `KLONKT_DATA_ROOT`.
Note:
See TracChangeset
for help on using the changeset viewer.
![(please configure the [header_logo] section in trac.ini)](/chrome/site/your_project_logo.png)