This is the version.

VRTSvcs-4.1.00.10-GA_RHEL4
VRTSvxfs-common-4.1.00.10-GA_RHEL4




On 3/1/07, Rebanta Mitra <[EMAIL PROTECTED]> wrote:

Hi,
Are you seeing a sudden jump of the cpu usage ?


After file size is grown more than 2GB or 4GB, we can see constant high cpu
rate.

Actually, our application is calling fdatasync() periodically to ensure that
the data is written to disk.

Also, does the application
create lots of dirty pages?


Not lot. Application is writing ~1.5KB size data chunk per millisecond. (i.e.
about 700-1000  transactions per second, one transaction is 1.5K)

If that's the case, then in most of the o/s the
cpu usage can increase because it has to go through the dirty list  to
check and lock each of the pages until the data is flushed.


Does VxFS is using Linux's Disk Caching Mechanism or Its own caching
mechanism.

BTW, what's the veritas-ha stack being used ?


VRTSvcs-4.1.00.10-GA_RHEL4
VRTSvxfs-common-4.1.00.10-GA_RHEL4


One thing more,
Where can i download trial version of above software??


regards,
-Rebanta


From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Tharindu
Rukshan Bamunuarachchi
Sent: Tuesday, February 27, 2007 9:35 PM
To: Veritas-ha@mailman.eng.auburn.edu
Cc: veritas-vx@mailman.eng.auburn.edu
Subject: [Veritas-ha] fdatasync makes process 100% cpu


We are using Veritas SFCFS for Linux.

When our program is trying to call fdatasync() to sync file core data
with disk, process cpu was increased to 100%.

Any idea on behavior of fdatasync() when used with Veritas cfs.

I hope Veritas behavior would be differ from typical Linux behavior.

Pls help


Thankx
Tharindu


--
Tharindu Rukshan Bamunuarachchi
all fabrications are subject to decay

Content-Description: ATT535579.txt
_______________________________________________
Veritas-ha maillist  -  Veritas-ha@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha


----- End forwarded message -----
_______________________________________________
Veritas-ha maillist  -  Veritas-ha@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha




--
Tharindu Rukshan Bamunuarachchi
all fabrications are subject to decay
_______________________________________________
Veritas-vx maillist  -  Veritas-vx@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-vx

Reply via email to