Hello,
I don't know your use case, and it might be confidential. Several problems here:
- If you need 19GB of RAM, you must have ++10 millions of timeseries, or series with huge labels/attributes. There might be a data modelisation problem, we can help you (support contract + NDA needed).
- Deleting data won't release RAM immediately (java will keep allocated RAM, up to the WARP10_HEAP limit you defined in your Warp 10 startup script.
- Deleting data won't release disk immediately (you need to wait for leveldb compaction cycles)
- api/v0/delete?selector=~.*{}&deleteall => should be OK, try to url encode the {} : api/v0/delete?deleteall&selector=~.*%7B%7D
Note : On a standalone version, if your purpose is just to restart from scratch, you can just stop warp10, delete all the /leveldb/* files, restart warp10. It won't affect tokens nor change anything.