FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Ubuntu > Ubuntu User

 
 
LinkBack Thread Tools
 
Old 07-26-2011, 10:00 PM
William King
 
Default Server becomes io-bound on no disk writes.

This issue is being documented here:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/815540


tl;dr:
Two servers nearly identical hardware, one becomes io-bound the other
one does not. Both are only processing boinc work units for
rosetta@home. No good reason for the hard drives to be io-bound found yet.


--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 12:00 AM
"compdoc"
 
Default Server becomes io-bound on no disk writes.

>Two servers nearly identical hardware, one becomes io-bound the other
>one does not.

That's some monster hardware you have there. Wish I had even one of those.

SO what's different between the two, then? Start by describing the drive
hardware...






--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 12:41 AM
NoOp
 
Default Server becomes io-bound on no disk writes.

On 07/26/2011 05:00 PM, compdoc wrote:
>>Two servers nearly identical hardware, one becomes io-bound the other
>>one does not.
>
> That's some monster hardware you have there. Wish I had even one of those.
>
> SO what's different between the two, then? Start by describing the drive
> hardware...

Ummm... comments #7 & #8 in the bug report provides that information.




--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 01:23 AM
William King
 
Default Server becomes io-bound on no disk writes.

The one that works has 3 of these drives:

*-disk:2
description: ATA Disk
product: WDC WD1001FALS-0
vendor: Western Digital
physical id: 0.0.0
bus info: scsi@4:0.0.0
logical name: /dev/sdc
version: 05.0
serial: WD-WMATV9394647
size: 931GiB (1TB)
capabilities: partitioned partitioned:dos
configuration: ansiversion=5 signature=000794e4

Where the one that does not work as intended has:

*-disk:2
description: ATA Disk
product: ST31000528AS
vendor: Seagate
physical id: 0.0.0
bus info: scsi@4:0.0.0
logical name: /dev/sdc
version: CC46
serial: 5VP7RA7M
size: 931GiB (1TB)
capabilities: partitioned partitioned:dos
configuration: ansiversion=5 signature=000d40a1


And the file system is ext4.


On 07/26/2011 05:41 PM, NoOp wrote:

On 07/26/2011 05:00 PM, compdoc wrote:

Two servers nearly identical hardware, one becomes io-bound the other
one does not.


That's some monster hardware you have there. Wish I had even one of those.

SO what's different between the two, then? Start by describing the drive
hardware...


Ummm... comments #7& #8 in the bug report provides that information.






--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 01:32 AM
"compdoc"
 
Default Server becomes io-bound on no disk writes.

In my own testing I do know that the speed/throughput is better for the WD
Black than the Barracudas I've tested.

Possible to get the output of this on each?

alpha:

sudo hdparm -t /dev/sdc2


bravo:

sudo hdparm -t /dev/sdb2


If I have it right, those are the large partitions on disk #2 of each
server. Hopefully, they aren't busy at the time of testing...




--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 04:25 AM
William King
 
Default Server becomes io-bound on no disk writes.

compdoc,

I am trying to find support to track down the cause of the spurious disk
writes that force the server into an io-bound state. Now is not the time
to be doing hardware benchmarks. Ask me again when the issue is resolved
and I would glad to provide such benchmarks.


-William King

On 07/26/2011 06:32 PM, compdoc wrote:

In my own testing I do know that the speed/throughput is better for the WD
Black than the Barracudas I've tested.

Possible to get the output of this on each?

alpha:

sudo hdparm -t /dev/sdc2


bravo:

sudo hdparm -t /dev/sdb2


If I have it right, those are the large partitions on disk #2 of each
server. Hopefully, they aren't busy at the time of testing...






--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 04:57 AM
Ric Moore
 
Default Server becomes io-bound on no disk writes.

On Tue, 2011-07-26 at 21:25 -0700, William King wrote:
> compdoc,
>
> I am trying to find support to track down the cause of the spurious disk
> writes that force the server into an io-bound state. Now is not the time
> to be doing hardware benchmarks. Ask me again when the issue is resolved
> and I would glad to provide such benchmarks.

Maybe he had a line of thinking towards your problem, and checking
through-put was a part of his winnowing through to the problem. Ric
--
My father, Victor Moore (Vic) used to say:
"There are two Great Sins in the world...
..the Sin of Ignorance, and the Sin of Stupidity.
Only the former may be overcome." R.I.P. Dad.
Linux user# 44256


--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 05:16 AM
Lucio M Nicolosi
 
Default Server becomes io-bound on no disk writes.

On Wed, Jul 27, 2011 at 1:57 AM, Ric Moore <wayward4now@gmail.com> wrote:
> On Tue, 2011-07-26 at 21:25 -0700, William King wrote:
>> compdoc,
>>
>> I am trying to find support to track down the cause of the spurious disk
>> writes that force the server into an io-bound state. Now is not the time
>> to be doing hardware benchmarks. Ask me again when the issue is resolved
>> and I would glad to provide such benchmarks.
>
> Maybe he had a line of thinking towards your problem, and checking
> through-put was a part of his winnowing through to the problem. Ric
> --

If the systems are nearly identical, wouldn't it be possible to make
them identical by running the Seagate array on the working machine and
trying to reproduce the bug? Perhaps they are not so identical as it
seems...

L.

--
Lucio M. Nicolosi, Eng.
Open Source Implementation
System and Applications
GNU/Linux

--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 05:22 AM
William King
 
Default Server becomes io-bound on no disk writes.

Lucio that is a good idea and is on my list of things to test.
Unfortunately the servers are racked in a datacenter and it will be a
few days before I can go and swap the drives.


-William King

On 07/26/2011 10:16 PM, Lucio M Nicolosi wrote:

On Wed, Jul 27, 2011 at 1:57 AM, Ric Moore<wayward4now@gmail.com> wrote:

On Tue, 2011-07-26 at 21:25 -0700, William King wrote:

compdoc,

I am trying to find support to track down the cause of the spurious disk
writes that force the server into an io-bound state. Now is not the time
to be doing hardware benchmarks. Ask me again when the issue is resolved
and I would glad to provide such benchmarks.


Maybe he had a line of thinking towards your problem, and checking
through-put was a part of his winnowing through to the problem. Ric
--


If the systems are nearly identical, wouldn't it be possible to make
them identical by running the Seagate array on the working machine and
trying to reproduce the bug? Perhaps they are not so identical as it
seems...

L.



--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 
Old 07-27-2011, 08:37 AM
Jordon Bedwell
 
Default Server becomes io-bound on no disk writes.

On 26/07/11 20:32, compdoc wrote:
> In my own testing I do know that the speed/throughput is better for the WD
> Black than the Barracudas I've tested.

Is this a serious statement that you want us to take seriously or would
you like to recall it now before we start refuting your tests because
they are so biased? When you "tested" these drives, did you bother to
check out the specs when you put them toe to toe? Hint hint: Hard
drives aren't just about how fast they spin when they have equal RPMs.

--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users
 

Thread Tools




All times are GMT. The time now is 09:57 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org