Re: backend crashes with mnogosearch - Mailing list pgsql-bugs

From Tom Lane
Subject Re: backend crashes with mnogosearch
Date
Msg-id 4256.987178291@sss.pgh.pa.us
Whole thread Raw
In response to backend crashes with mnogosearch  (Nick Wellnhofer <nick@paradigmashift.com>)
List pgsql-bugs
Nick Wellnhofer <nick@paradigmashift.com> writes:
> further investigatoin showed that a simple "SELECT * FROM url" query
> resulted in

> Backend message type 0x44 arrived while idle
> pqReadData() -- backend closed the channel unexpectedly.

Are you sure this is a backend crash?  (Real proof would be a report
of abnormal child termination in the postmaster's log, or appearance
of a core dump file in the database's subdirectory.)  The "arrived while
idle" bit looks suspiciously like the sort of misbehavior that libpq
tends to engage in if it runs out of memory for a query result.

> then i tried several "SELECT * FROM url LIMIT <some number>" queries.
> up to a certain number the queries were successful, but from that
> number on i got the error message.

How big is the "certain number", and how close might that number of
rows come to exhausting your frontend application's available memory?

            regards, tom lane

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: runtest bug and bigcheck problems
Next
From: Tom Lane
Date:
Subject: Re: aliasing table name in update