adsj@novozymes.com (Adam =?utf-8?Q?Sj=C3=B8gren?=) writes:
> Also, the error we are getting is now: "unexpected chunk number 2
> (expected 3) for toast value 1498303849 in pg_toast_10919630", where
> previously we've only seen "unexpected chunk number 0 (expected 1)".
> We are kind of at a loss, so any suggestions on what we could try are
> welcome.
The basic thrust of these messages is "I'm reading what should be
sequentially numbered data chunks for this toast OID, and the sequence
numbers are wrong". Both of these instances could be explained by
duplicate toast rows (or duplicate index entries pointing at one row),
though of course that would just move to the next question of how it
got that way. Anyway, you could move the investigation along with
some manual checking into what's in that toast table. For instance
select chunk_id, chunk_seq, ctid, xmin, xmax, length(chunk_data)
from pg_toast.pg_toast_10919630
where chunk_id = 1498303849
order by 1,2;
might be informative. If you do see what seem to be duplicate
chunk_seq values, checking whether they're still there in a
seqscan would be good.
regards, tom lane