pg.ddx.io pgsql-performance@postgresql.org mailing list archive
help / color / mirror / Atom feedDisk Fills Up and fsck "Compresses" it
5+ messages / 3 participants
[nested] [flat]
* Disk Fills Up and fsck "Compresses" it
@ 2007-05-15 20:18 Y Sidhu <ysidhu@gmail.com>
2007-05-15 23:41 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
0 siblings, 1 reply; 5+ messages in thread
From: Y Sidhu @ 2007-05-15 20:18 UTC (permalink / raw)
To: pgsql-performance
Anyone seen PG filling up a 66 GB partition from say 40-ish percentage to
60-ish percentage in a manner of minutes. When I run a 'fsck' the disk usage
comes down to 40-ish percentage. That's about 10+ GB's variance.
This is a FreeBSD 6.2 RC2, 4GB memory, Xeon 3.2 GHz '4' of the '8' CPUs in
use - dual cpu, dual core with HTT turned off in the sense that the other 4
cpu's have been masked out. The drive is a Western Digital 70 GB SATA.
--
Yudhvir Singh Sidhu
408 375 3134 cell
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Disk Fills Up and fsck "Compresses" it
2007-05-15 20:18 Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
@ 2007-05-15 23:41 ` Jim C. Nasby <decibel@decibel.org>
2007-05-16 15:21 ` Re: Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
0 siblings, 1 reply; 5+ messages in thread
From: Jim C. Nasby @ 2007-05-15 23:41 UTC (permalink / raw)
To: Y Sidhu <ysidhu@gmail.com>; +Cc: pgsql-performance
I'm guessing you're seeing the affect of softupdates. With those enabled
it can take some time before the space freed by a delete will actually
show up as available.
On Tue, May 15, 2007 at 01:18:42PM -0700, Y Sidhu wrote:
> Anyone seen PG filling up a 66 GB partition from say 40-ish percentage to
> 60-ish percentage in a manner of minutes. When I run a 'fsck' the disk usage
> comes down to 40-ish percentage. That's about 10+ GB's variance.
>
> This is a FreeBSD 6.2 RC2, 4GB memory, Xeon 3.2 GHz '4' of the '8' CPUs in
> use - dual cpu, dual core with HTT turned off in the sense that the other 4
> cpu's have been masked out. The drive is a Western Digital 70 GB SATA.
>
> --
> Yudhvir Singh Sidhu
> 408 375 3134 cell
--
Jim Nasby decibel@decibel.org
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Disk Fills Up and fsck "Compresses" it
2007-05-15 20:18 Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
2007-05-15 23:41 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
@ 2007-05-16 15:21 ` Y Sidhu <ysidhu@gmail.com>
2007-05-16 15:51 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
0 siblings, 1 reply; 5+ messages in thread
From: Y Sidhu @ 2007-05-16 15:21 UTC (permalink / raw)
To: Jim C. Nasby <decibel@decibel.org>; +Cc: pgsql-performance
What do you mean by "softupdates?" Is that a parameter in what I am guessing
is the conf file?
Yudhvir
On 5/15/07, Jim C. Nasby <decibel@decibel.org> wrote:
>
> I'm guessing you're seeing the affect of softupdates. With those enabled
> it can take some time before the space freed by a delete will actually
> show up as available.
>
> On Tue, May 15, 2007 at 01:18:42PM -0700, Y Sidhu wrote:
> > Anyone seen PG filling up a 66 GB partition from say 40-ish percentage
> to
> > 60-ish percentage in a manner of minutes. When I run a 'fsck' the disk
> usage
> > comes down to 40-ish percentage. That's about 10+ GB's variance.
> >
> > This is a FreeBSD 6.2 RC2, 4GB memory, Xeon 3.2 GHz '4' of the '8' CPUs
> in
> > use - dual cpu, dual core with HTT turned off in the sense that the
> other 4
> > cpu's have been masked out. The drive is a Western Digital 70 GB SATA.
> >
> > --
> > Yudhvir Singh Sidhu
> > 408 375 3134 cell
>
> --
> Jim Nasby decibel@decibel.org
> EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
>
--
Yudhvir Singh Sidhu
408 375 3134 cell
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Disk Fills Up and fsck "Compresses" it
2007-05-15 20:18 Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
2007-05-15 23:41 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
2007-05-16 15:21 ` Re: Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
@ 2007-05-16 15:51 ` Jim C. Nasby <decibel@decibel.org>
2007-05-16 23:28 ` Re: Disk Fills Up and fsck "Compresses" it Mark Kirkwood <markir@paradise.net.nz>
0 siblings, 1 reply; 5+ messages in thread
From: Jim C. Nasby @ 2007-05-16 15:51 UTC (permalink / raw)
To: Y Sidhu <ysidhu@gmail.com>; +Cc: pgsql-performance
No, it's part of FreeBSD's UFS. google FreeBSD softupdates and you
should get plenty of info.
As I said, it's probably not worth worrying about.
On Wed, May 16, 2007 at 08:21:23AM -0700, Y Sidhu wrote:
> What do you mean by "softupdates?" Is that a parameter in what I am guessing
> is the conf file?
--
Jim Nasby decibel@decibel.org
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: Disk Fills Up and fsck "Compresses" it
2007-05-15 20:18 Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
2007-05-15 23:41 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
2007-05-16 15:21 ` Re: Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
2007-05-16 15:51 ` Re: Disk Fills Up and fsck "Compresses" it Jim C. Nasby <decibel@decibel.org>
@ 2007-05-16 23:28 ` Mark Kirkwood <markir@paradise.net.nz>
0 siblings, 0 replies; 5+ messages in thread
From: Mark Kirkwood @ 2007-05-16 23:28 UTC (permalink / raw)
To: Jim C. Nasby <decibel@decibel.org>; Y Sidhu <ysidhu@gmail.com>; +Cc: pgsql-performance
Jim C. Nasby wrote:
> No, it's part of FreeBSD's UFS. google FreeBSD softupdates and you
> should get plenty of info.
>
> As I said, it's probably not worth worrying about.
>
> On Wed, May 16, 2007 at 08:21:23AM -0700, Y Sidhu wrote:
>> What do you mean by "softupdates?" Is that a parameter in what I am guessing
>> is the conf file?
Here is quite a good article on the interaction of fsck and softupdates
in FreeBSD:
http://www.usenix.org/publications/library/proceedings/bsdcon02/mckusick/mckusick_html/index.html
Having said that, it seems to talk about space lost by unreferenced
blocks and inodes is the context of panic or power outage, as opposed to
normal softupdate operation (unless I'm missing something...)
Cheers
Mark
^ permalink raw reply [nested|flat] 5+ messages in thread
end of thread, other threads:[~2007-05-16 23:28 UTC | newest]
Thread overview: 5+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2007-05-15 20:18 Disk Fills Up and fsck "Compresses" it Y Sidhu <ysidhu@gmail.com>
2007-05-15 23:41 ` Jim C. Nasby <decibel@decibel.org>
2007-05-16 15:21 ` Y Sidhu <ysidhu@gmail.com>
2007-05-16 15:51 ` Jim C. Nasby <decibel@decibel.org>
2007-05-16 23:28 ` Mark Kirkwood <markir@paradise.net.nz>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox