I'm running Aurora Postgres and noticed that metric Maximum Used Transaction IDs is increasing a lot. Trying to manually vacuum the tables that's was not automatically vacuumed lead me to the following errors:
sql> VACUUM sales.hist_status [2019-01-06 15:22:44] [58P01] ERROR: could not access status of transaction 247756185 [2019-01-06 15:22:44] Detail: CLOG segment 29 does not exist: No such file or directory. sql> VACUUM sales.hist_items [2019-01-06 15:22:55] [XX001] ERROR: found xmin 235941673 from before relfrozenxid 256796328
Googling these erros points me to database corruption. What should I do?