Re: BUG #17245: Index corruption involving deduplicated entries

On 29.10.2021 0:42, Andres Freund wrote:
>> Unfortunately there's no WAL archiving, and no available WAL files from
>> before the state was known to be corrupt. We'll look into enabling
>> archiving in case the problem recurs or can be reproduced.
> 
> Any chance autovacuum logging was enabled and that those logs still exist? Not
> quite as good as still having the WAL, but it still might allow us to
> reconstruct the index / vacuum scans, and the horizons used.

The issue manifested again earlier today *after* a REINDEX followed by 
enabling WAL replica logging on the 24th of October. I saved a snapshot 
of the filesystem holding the data directory. Would that be useful for 
further analysis?

(The daily vacuum runs have not been disabled so there have been at 
least 3 since.)

-- 
K. R.