prune_blockchain cause entire node to hang on

[SyntheticBird45]

Monero v0.18.3.1 syncing 99%, Linux 6.7 x86_64

prune_blockchain rpc methods cause node to hang on and don't react to SIGTERM. Killing process is the only solution.
The bug is consistent

[SyntheticBird45]

command used for rpc:

curl http://127.0.0.1:18081/json_rpc -d '{"jsonrpc":"2.0","id":"0","method":"prune_blockchain","params":{"check":true}}' -H 'Content-Type: application/json'

[selsta]

How long did you wait? The prune process can take multiple hours.

[SyntheticBird45]

I waited 30 minute. node is already pruned and since i set check: true It shouldn't even trigger the pruning process

[SyntheticBird45]

Issue is still happening when node is fully synced

[selsta]

does entering check_blockchain_pruning in monerod also not work?

[SyntheticBird45]

Yes

[SyntheticBird45]

Just tested if it wasn't alacritty or terminal emulator causing issue but it is consistent across 3 other terminals.

[SyntheticBird45]

Disk read I/O is increasing while hanging so it might be stuck at interacting with the database. I maybe stupid but is the prune_blockchain rpc command check the pruning state over all the database ?

[selsta]

Do you have an SSD or HDD? Running check_blockchain_pruning took about 5 minutes on my system with SSD.

moneromooo said it reads all tx data so it can be lengthy

[SyntheticBird45]

I've an SSD, however the filesystem is almost at capacity.

[SyntheticBird45]

Looking at my I/O Read rate monerod will have finish to check in 1 hour. I believed that monerod didn't checked the all database for pruning state. If the hanging of monerod is something known during the check of the prune worker, I believe we can close this issue.