> random_page_cost = 0.5
Not likely. The lowest this value should ever be is 1, and thats if
you're using something like a ram drive.
If you're drives are doing a ton of extra random IO due to the above
(rather than sequential reads) it would lower the throughput quite a
bit.
Try a value of 2 fo
Simon Riggs wrote:
On the other hand, I was just about to change the wal_debug behaviour to
allow better debugging of PITR features as they're added.
That's a development activity. Enabling the WAL_DEBUG #ifdef by
default during the 7.5 development cycle would be uncontroversial, I
think.
I thin
"Simon Riggs" <[EMAIL PROTECTED]> writes:
> The behaviour I wish to add is:
> Keep wal_debug as a value between 0 and 16.
> If =0 then no debug output (default).
> Use following bitmasks against the value
> Mask 1 = XLOG Checkpoints get logged
> Mask 2 = Archive API calls get logged
> Mask 4 = Tran
Josh Berkus wrote:
Hmmm. I was told that it was this way for 7.4 as well; that's why it's in
the docs that way.
No such statement is made in the docs AFAIK: they merely say "If
nonzero, turn on WAL-related debugging output."
I invented a new #ifdef symbol when making this change in CVS HEAD, s
>Neil Conway
> Simon Riggs wrote:
> >>Josh Berkus wrote
> >>
> >>>Simon Riggs wrote
> >>>Please set WAL_DEBUG to 1 so we can see a bit more info: thanks.
> >>
> >>I'm pretty sure that WAL_DEBUG requires a compile-time option.
> >
> > I'm surprised, but you are right, the manual does SAY this requir
Neil,
> Actually, the manual is correct: in 7.4 and earlier releases, enabling
> wal_debug can be done without also setting a compile-time #ifdef. As
> of current CVS HEAD, the WAL_DEBUG #ifdef must be defined before this
> variable is available.
Hmmm. I was told that it was this way for 7.4
Simon Riggs wrote:
Josh Berkus wrote
Simon Riggs wrote
Please set WAL_DEBUG to 1 so we can see a bit more info: thanks.
I'm pretty sure that WAL_DEBUG requires a compile-time option.
I'm surprised, but you are right, the manual does SAY this requires a
compile time option; it is unfortunately not
>Josh Berkus wrote
> >Simon Riggs wrote
> > Please set WAL_DEBUG to 1 so we can see a bit more info: thanks.
>
> I'm pretty sure that WAL_DEBUG requires a compile-time option.
In my naiveté, I just set and use it. I discovered it in the code, then
set it to take advantage.
I'm surprised, but you
>Tom Lane
> "Simon Riggs" <[EMAIL PROTECTED]> writes:
> > - why checkpoint interval of 300 secs causes them to happen every 10
> > mins in quieter periods; is that an occaisional update occurring?
>
> There is code in there to suppress a checkpoint if no WAL-loggable
> activity has happened since
Simon,
> Please set WAL_DEBUG to 1 so we can see a bit more info: thanks.
I'm pretty sure that WAL_DEBUG requires a compile-time option.
--
-Josh Berkus
Aglio Database Solutions
San Francisco
---(end of broadcast)---
TIP 6: Have you searched o
"Simon Riggs" <[EMAIL PROTECTED]> writes:
> - why checkpoint interval of 300 secs causes them to happen every 10
> mins in quieter periods; is that an occaisional update occurring?
There is code in there to suppress a checkpoint if no WAL-loggable
activity has happened since the last checkpoint.
>Rob Fielding wrote:
> My focus today has been on WAL - I've not looked at WAL before. By
> increasing the settings thus :
>
> wal_buffers = 64 # need to determin WAL usage
> wal_files = 64 # range 0-64
> wal_sync_method = fsync # the default varies across platforms:
> wal_debug = 0
Further update to my WAL experimentation. pg_xlog files have increased
to 81, and checking today up to 84. Currently nothing much going on with
the server save a background process running a select every 30 seconds
with almost no impact (according to IO from vmstats).
This in itself is a good s
Rod Taylor wrote:
random_page_cost = 0.5
Try a value of 2 for a while.
OK thanks Richard and Rod. I've upped this to 2. I think I left this
over from a previous play with setttings on my IDE RAID 0 workstation.
It seemed to have a good effect being set as a low float so it stuck.
I've
Rob
Sir - I have to congratulate you on having the most coherently summarised and
yet complex list query I have ever seen.
I fear that I will be learning from this problem rather than helping, but one
thing did puzzle me - you've set your random_page_cost to 0.5? I'm not sure
this is sensible
Hi,
There alot here, so skip to the middle from my WAL settings if you like.
I'm currently investigating the performance on a large database which
consumes email designated as SPAM for the perusal of customers wishing
to check. This incorporates a number of subprocesses - several delivery
daem
16 matches
Mail list logo