Re: Review: Hot standby

From: "Pavan Deolasee" <pavan(dot)deolasee(at)gmail(dot)com>
To: "Simon Riggs" <simon(at)2ndquadrant(dot)com>
Cc: pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Review: Hot standby
Date: 2008-11-25 09:26:11
Message-ID: 2e78013d0811250126k5d8ddb07s61543f42d4dda4f3@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

The following sequence of commands causes server crash.

postgres=# BEGIN TRANSACTION READ WRITE ;
BEGIN
postgres=# SELECT * FROM pg_class FOR UPDATE;
FATAL: cannot assign TransactionIds during recovery
STATEMENT: SELECT * FROM pg_class FOR UPDATE;
FATAL: cannot assign TransactionIds during recovery
server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
The connection to the server was lost. Attempting reset: Succeeded.

I think we should disallow starting a read-write transaction during
recovery. The patch attempts to do that by setting transaction mode to
read-only, but not enough to prevent somebody to explicitly mark the
transaction as read-write.

Thanks,
Pavan

--
Pavan Deolasee
EnterpriseDB http://www.enterprisedb.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2008-11-25 09:52:26 Erroring out on parser conflicts
Previous Message Stephen R. van den Berg 2008-11-25 09:23:35 Re: blatantly a bug in the documentation