When a reorg happens ord 22.1 is very very slow&stuck. The issue happened twice, once today and once yesterday. Its been stuck for hours and needed to restore from snapshot manually.

I have in parallel an ord19 that worked perfectly fine for the reorgs.

For mainnet ord was run with default params (eg. not overriding max-savepoints and savepoints interval)

ord is run with --index-sats --index-runes

Any ideas on what might cause this?

From the logs it looks like it committed to disk 40 minutes before the reorg

INFO 2025-02-02T10:39:06.853052169Z [2025-02-02T10:39:06Z INFO ord::index::reorg] rolling back database after reorg of depth 2 at height 881954
INFO 2025-02-02T10:39:06.853022763Z [2025-02-02T10:39:06Z INFO ord::index] 2 block deep reorg detected at height 881954
INFO 2025-02-02T10:03:31.582252047Z [2025-02-02T10:03:31Z INFO ord::index::updater] Committing at block height 881954, 6481 outputs traversed, 5483 in map, 999 cached...

Looked through ord 22.2 changelog and it does not seem like any of the changes would fix this.

Update:
It finished rolling back after 2 hours. The machine has 70gb of ram & ssd

[2025-02-02T12:53:37Z INFO  ord::index::reorg] successfully rolled back database to height 881934