Daemon keeps corrupting during sync

[flpDiskJ]

I have been continuously plagued with my daemon corrupting. I have it stored on an internal SSD and run the following command:

./monerod --data-dir /path/to/xmr-daemon --db-sync-mode safe --prune-blockchain 

I get the following error:

2026-02-26 03:05:26.163	E Error adding block with hash: <6a534048a58835365de9c280ca7a94acfae555bd91040643bfffc2496ed5f758> to blockchain, what = Error adding spent key image to db transaction: MDB_CURSOR_FULL: Internal error - cursor stack limit reached

I'm running Debian 12.

It happens randomly. No power losses, shutdowns, suspends or any of that.
Also doesn't correlate to any specific block height as I have been making a backup copy that I revert to when it crashes and then it will keep going for a while.

I know the solution is to re-sync but the frequency in which this keeps happening is my concern. I am unable to fully sync without it crashing hence the backups.

[nahuhh]

Is the ssd failing?

[flpDiskJ]

I don't believe so as I haven't had any issues with any other files corrupting.

[selsta]

Which file system are you using?