Re: 2PC-induced lockup

From: Chris Browne <cbbrowne(at)acm(dot)org>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: 2PC-induced lockup
Date: 2007-07-11 14:43:23
Message-ID: 604pkbxapg.fsf@dba2.int.libertyrms.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

tomas(at)tuxteam(dot)de writes:
> On Wed, Jul 11, 2007 at 12:38:09AM -0400, Tom Lane wrote:
>> tomas(at)tuxteam(dot)de writes:
> [...]
>> > It might make sense then to clear the pg_twophase directory on DB
>> > startup.
>>
>> <blink> I fear you have 100% misunderstood the point. The *only*
>> reason for that feature is to survive DB crashes.
>
> Ah -- so it is intentional that it keeps the DB from starting again.
> OK, then I misunderstood. Sorry for the noise.

I don't think that is so much "intentional" as it is an "emergent
property."

The usual point to 2PC is that once transactions are PREPAREd, they
*need* to be stored robustly enough to survive even a DB crash.

If one locks certain vital system resources, as part of that PREPAREd
transaction, that evidently causes some problems, alas...

The right resolution to this is not, a priori, evident yet.
--
(format nil "~S(at)~S" "cbbrowne" "linuxdatabases.info")
http://cbbrowne.com/info/linuxdistributions.html
Rules of the Evil Overlord #38. "If an enemy I have just killed has a
younger sibling or offspring anywhere, I will find them and have them
killed immediately, instead of waiting for them to grow up harboring
feelings of vengeance towards me in my old age."
<http://www.eviloverlord.com/>

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2007-07-11 15:16:32 Re: Bgwriter strategies
Previous Message Alvaro Herrera 2007-07-11 13:08:11 Re: Bgwriter strategies