Thread: [pgsql] How bad is this full vacuum error?
Hello.
How bad is this?
2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 ERROR: missing chunk number 0 for toast value 1086399 in pg_toast_987417
2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 STATEMENT: VACUUM (FULL);
Does this mean I have db corruption, or just that the vacuum quit?
Thanks.
How bad is this?
2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 ERROR: missing chunk number 0 for toast value 1086399 in pg_toast_987417
2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 STATEMENT: VACUUM (FULL);
Does this mean I have db corruption, or just that the vacuum quit?
Thanks.
Gary Webster <webster@lexmark.com> writes: > How bad is this? > 2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 ERROR: missing > chunk number 0 for toast value 1086399 in pg_toast_987417 > 2012-08-01 06:30:03 PDT 15961 [local] cp_repository_na2 STATEMENT: VACUUM > (FULL); If it's repeatable, it's corrupted data :-(. If it just happened once, and doesn't recur on the next try, it might have been related to a system catalog race condition that we fixed last fall. (What PG version are you running?) I'm unsure about that though, because the toast table name seems to indicate that it's for a user table not a system catalog, and I'm not very sure why VACUUM would be doing toast dereferences into a user table. regards, tom lane