"Joe Maldonado" <jmaldonado(at)webehosting(dot)biz> writes:
> I have a seemingly corrupt row in a table and wanted to look at it's
> when I try to query it I get the following...
> db=# select * from some_table offset 411069 limit 1;
> ERROR: invalid memory alloc request size 4294967293
> but when I select individual fields within the record I get data.
That's odd ... I'd certainly expect one or the other field of the table
to show that failure.
> Is there a way to read this row from the datafile to examine it closer?
Select "ctid" from the troublesome row to determine its block and item
number, then dump out that block with pg_filedump. If there is data
corruption it'll usually be possible to see it in the pg_filedump dump.
Another line of attack is to attach to the backend process with gdb and
set a breakpoint at errfinish (or elog if a pre-7.4 backend), and then
get a stack trace back from the error report. This will help narrow
down exactly where the bogus allocation request is coming from.
regards, tom lane
In response to
pgsql-general by date
|Next:||From: Tom Lane||Date: 2004-06-30 03:31:07|
|Subject: Re: dup(0) failed after 3195 successes: Bad file descriptor |
|Previous:||From: Martijn van Oosterhout||Date: 2004-06-30 01:13:32|
|Subject: FULL JOIN and mergjoinable conditions...|