Andrew Hammond <andrew.george.hammond@xxxxxxxxx> writes:Hmm ... pg_toast_2619 is pg_statistic's toast table. Is this 9.0.x,
> Oct 24 04:05:57 db-app02 postgres[24640]: [2-1]
> user=nutricate,db=admin_production ERROR: missing chunk number 0 for
> toast value
> 2411466 in pg_toast_2619
and are you by any chance in the habit of running CLUSTER or VACUUM FULL
on your system catalogs? Could one have been running in the background
when this happened?
This is 9.0.4.
I am not aware of any VACUUM FULL or CLUSTER operations that are scheduled. To the best of my knowledge this DB has never been either clustered or vacuum full'd. The following occur in the log file immediately before the error message above.
Oct 24 04:05:57 db-app02 postgres[24639]: [2-1] user=,db= LOG: automatic vacuum of table "admin_production.pg_catalog.pg_statistic": index scans: 0
Oct 24 04:05:57 db-app02 postgres[24639]: [2-2] #011pages: 0 removed, 150 remain
Oct 24 04:05:57 db-app02 postgres[24639]: [2-3] #011tuples: 254 removed, 925 remain
Oct 24 04:05:57 db-app02 postgres[24639]: [2-4] #011system usage: CPU 0.01s/0.00u sec elapsed 0.02 sec
Oct 24 04:05:57 db-app02 postgres[24639]: [3-1] user=,db= LOG: automatic vacuum of table "admin_production.public.devices": index scans: 0
Oct 24 04:05:57 db-app02 postgres[24639]: [3-2] #011pages: 0 removed, 353 remain
Oct 24 04:05:57 db-app02 postgres[24639]: [3-3] #011tuples: 824 removed, 2261 remain
Oct 24 04:05:57 db-app02 postgres[24639]: [3-4] #011system usage: CPU 0.00s/0.00u sec elapsed 0.08 sec
Oct 24 04:05:57 db-app02 postgres[24639]: [4-1] user=,db= LOG: automatic analyze of table "admin_production.public.devices" system usage: CPU 0.00s/0.08u sec elapsed 0.17 sec