Date: Tue, 31 Jul 2007 17:46:26 -0400
From: "Jignesh K. Shah" <J.K.Shah@Sun.COM>
Subject: Re: User concurrency thresholding: where do I look?
In-reply-to: <200707311433.20568.josh@agliodbs.com>
To: josh@agliodbs.com
Cc: pgsql-performance@postgresql.org
Message-id: <46AFADB2.5070409@sun.com>
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=UTF-8
Content-transfer-encoding: 7BIT
References: <200707190828.59088.josh@agliodbs.com> <46A90F83.8050401@sun.com>
	<1185527727.4191.22.camel@ebony.site>
	<200707311433.20568.josh@agliodbs.com>
User-Agent: Thunderbird 2.0.0.5 (Windows/20070716)

Yep quite a bit of transactions .. But the piece that's slow is where it 
is clearing it up in CommitTransaction().
I am not sure of how ProcArrayLock is designed to work and hence not 
clear what  we are seeing is what we expect.

Do we have some design doc on ProcArrayLock to understand its purpose?

Thanks.
Regards,
Jignesh


Josh Berkus wrote:
> Simon,
>
>   
>> Well thats pretty weird. That code path clearly only happens once per
>> transaction and ought to be fast. The other code paths that take
>> ProcArrayLock like TransactionIdIsInProgress() and GetSnapshotData()
>> ought to spend more time holding the lock. Presumably you are running
>> with a fair number of SERIALIZABLE transactions?
>>     
>
> Given that this is TPCC-analog, I'd assume that we are.
>
> Jignesh?
>
>