Re: pg_dumpall failing from possible corrupted shared memory

[Richard Broersma Jr <rabroersma(at)yahoo(dot)com>]

> > mydb=# select pg_get_indexdef(indexrelid) from pg_index where indrelid = '16737';
> > server closed the connection unexpectedly
> 
> So what do you get from 'select * from pg_index where indrelid = 16737' ?
> If that crashes, which individual columns can you select from the row?

I found the exact row and column that causes the server to crash.

http://archives.postgresql.org/pgsql-general/2006-10/msg01320.php

I believe that corrupted value should be:

CREATE INDEX index_daily ON process USING btree (date_trunc('day'::text, tstamp))

I tried to:

 select * from process;
 select * from process limit 1;
 explain analyze select * from process;

but they also crashes the server.

Regards,

Richard Broersma Jr.