[PostgreSQL-Hackers] Re: ALTER TABLE lock strength reduction patch is unsafe

[PostgreSQL-Hackers] Re: ALTER TABLE lock strength reduction patch is unsafe

	Simon Riggs
	Aug 1, 2013 at 6:04 am

On 1 August 2013 01:53, Noah Misch wrote:

I was able to distill the above hypothesis into an actual crash with
reduce_lock_levels.v13.patch. Test recipe:

Thank you for the report; thank you again for reporting in sufficient
time to allow me to investigate and fix by the next CF.

--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

Search Discussions

Discussion Posts

Noah Misch: I was able to distill the above hypothesis into an actual crash with reduce_lock_levels.v13.patch. Test recipe: 1. Build with --enable-cassert and with -DCATCACHE_FORCE_RELEASE=1. An AcceptInvalidationMessages() will then happen at nearly every syscache lookup, making it far easier to hit a problem of this sort. 2. Run these commands as setup: create table root (c int); create table part (check (c > 0), check (c > 0)) inherits (root); 3. Attach a debugger to your session and set a breakpoint at

Related Discussions

Discussion Navigation

view	thread \| post
posts	‹ prev \| 7 of 7 \| next ›

Discussion Overview

group	pgsql-hackers @
categories	postgresql
posted	Jul 7, '13 at 1:24p
active	Aug 1, '13 at 6:04a
posts	7
users	4
website	postgresql.org...
irc	#postgresql

font	variable	Hotkey: f
user style	avatars	Hotkey: a

[PostgreSQL-Hackers] Re: ALTER TABLE lock strength reduction patch is unsafe

Search Discussions

Discussion Posts

Related Discussions

4 users in discussion