On Wednesday April 27 2005 4:40 pm, Tom Lane wrote:
> "Ed L." <pgsql@xxxxxxxxxxxxx> writes:
> > On Wednesday April 27 2005 3:55 pm, Tom Lane wrote:
> >> Have you tried REINDEXing the toast table in question?
> >
> > Not yet. Any way to repair it without blocking concurrent
> > access?
>
> Unlikely. But the lock will only affect operations that need
> to touch toasted field values.
Does this shed any light? pg_toast_6221538 is the relevant toast
table...
$ psql -c "set enable_indexscan=off; select chunk_seq,
length(chunk_data) from pg_toast.pg_toast_6221538 where chunk_id
= 19319495 order by chunk_seq;"
chunk_seq | length
-----------+--------
(0 rows)
$ psql -c "select chunk_seq, length(chunk_data) from
pg_toast.pg_toast_6221538 where chunk_id = 19319495 order by
chunk_seq;"
chunk_seq | length
-----------+--------
(0 rows)
Ed
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
