From: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: log chunking broken with large queries under load |
Date: | 2012-04-02 16:55:30 |
Message-ID: | 4F79DA02.5060201@dunslane.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 04/02/2012 12:44 PM, Tom Lane wrote:
> Andrew Dunstan<andrew(at)dunslane(dot)net> writes:
>> On 04/02/2012 12:00 PM, Tom Lane wrote:
>>> This seems like it isn't actually fixing the problem, only pushing out
>>> the onset of trouble a bit. Should we not replace the fixed-size array
>>> with a dynamic data structure?
>> But maybe your're right. If we do that and stick with my two-dimensional
>> scheme to keep the number of probes per chunk down, we'd need to reorg
>> the array every time we increased it. That might be a bit messy, but
>> might be ok. Or maybe linearly searching an array of several hundred
>> slots for our pid for every log chunk that comes in would be fast enough.
> You could do something like having a list of pending chunks for each
> value of (pid mod 256). The length of each such list ought to be plenty
> short under ordinary circumstances.
>
>
Yeah, ok, that should work. How big would we make each list to start
with? Still 20, or smaller?
cheers
andrew
From | Date | Subject | |
---|---|---|---|
Next Message | Kevin Grittner | 2012-04-02 16:58:23 | Re: measuring lwlock-related latency spikes |
Previous Message | David Johnston | 2012-04-02 16:49:41 | Fwd: [HACKERS] Switching to Homebrew as recommended Mac install? / apology |