[Linux-cluster] GFS i/o Request size

[Mahmoud Hanafi]

This seem to be an issue with the el5 . On el4 I can set the 
/sys/block/sd*/queue/max_sector_kb for the block device up 4096KB. Using 
xdd writing directly to the device I can achive I/O request sizes of 
4096KB. But doing the same thing on el5 the max I can get is 512kb. I know 
there has been changes to the qla2400 drivers from el4 to el5. I am not 
sure if this limit is at the driver level or not. One other thing to note 
on el5 the default value for max_sector_kb is always set to a max of 
512KB. But on el4 it gets set to 4096KB which is the block size of the 
luns.

-Mahmoud


--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
This is a PRIVATE message. If you are not the intended recipient, please 
delete without copying and kindly advise us by e-mail of the mistake in 
delivery. NOTE: Regardless of content, this e-mail shall not operate to 
bind CSC to any order or other contract unless pursuant to explicit 
written agreement or government initiative expressly permitting the use of 
e-mail for such purpose.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------




Maciej Bogucki <maciej.bogucki at artegence.com> 
07/31/2007 11:08 AM

To
linux clustering <linux-cluster at redhat.com>, Mahmoud Hanafi/DEF/CSC at CSC
cc

Subject
Re: [Linux-cluster] GFS i/o Request size






Mahmoud Hanafi napisał(a):
> 
> 
> It seem that GFS likes to do all I/O in 512KB sizes. Is there a way to
> increase this? I like to do I/O sizes of 1024 or 2048KB.
> 
Hello,

First of all linux kernel breaks IO into pages, and the default page
size is 4K. Second You have to set blocksize for filesystem and default
for GFS is 4096(65536 maximum). Third You have segment size for You SAN
storage - for IBM DS4700 You can set from 8KB to 512KB, and here for
every I/O storage read/write this size od data. If it is clear for You,
which value You want to tune?

Best Regards
Maciej Bogucki



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20070731/66cf1038/attachment.htm>