Meaning of log entry 'WARN ipfs_sqlite_block_store::db: delete_orphaned took 1107380us and failed'

Hi there,

we keep getting the following entries in the logs of our nodes:

Dec 02 08:11:40.639  INFO ipfs_embed::db: going for gc!
Dec 02 08:11:41.719  INFO ipfs_sqlite_block_store::db: determine_orphaned took 1079940us
Dec 02 08:11:41.747  WARN ipfs_sqlite_block_store::db: delete_orphaned took 1107380us and failed

What could be the reason for this and is there something we should/could do?

Kind regards,
Max

Hi Max,

thanks for reporting! There’s a fix for this in 2.8.1 — in short, there was an inconsequential error in the code that led to the WARN line. That version will also properly clean up temporary storage (which could in some cases sum up to a few GB! — not sure what sqlite is doing there) during Actyx startup, so it is generally advisable to update.

(more info: the sqlite call to checkpoint the write-ahead log was actually performed as it should, but using the wrong library function, which then reported an error because there was a result row)

Regards,

Roland

Hi Roland,

thanks for the reply and the technical background info!

According to the log of the node, we are currently using Actyx 2.8.1-5ac08c40ada3b996ea76463462a83f9b35f89993-linux-x86_64-release.
Was the fix already shipped in that version?

Kind regards
Max

My bad, I should’ve looked more closely: the fix is in 2.8.2. Sorry about the mixup!