Re: I/O operations

["rohit vashist" <rohitvashist2kk3@xxxxxxxxx>]

On Sat, Jan 3, 2009 at 2:51 AM, Greg Freemyer <greg.freemyer@xxxxxxxxx> wrote:

On Fri, Jan 2, 2009 at 3:34 PM, rohit vashist
<rohitvashist2kk3@xxxxxxxxx> wrote:
> Hi
>
> I went through the code of iostat.In the ninth field it gives the count of
> number of I/O,but it goes zero but it can be thought.
> Can u specify the grounds how hard it would to be suspend all the I/O
> currently on the device.
>
> Thanks

I'm confused.  You only need to suspend all write activity per your
first post not all I/O, right?

Are you trying to suspend write activity to the filesystem, or to the
underlying block device (ie. hdd/dm/md devices)?

I just need to stop all the write operations on file system i.e it can be deviated to the request queue for a while though reads can be allowed .
 

In a previous thread you asked about the filesystem block device
freeze support and seemed unhappy that it only suspended writes to the
block device, but left the filesystem writable via temporarily holding
data in a write cache.

Sorry but i think that was not my thread.

But in the above you ask about how to suspend writes to the block
device, and you already know how to do that.  At least writes
initiated by a filesystem.

Please explain with more clarity where you need to suspend writes and
why a block device freeze is not adequate.

See i need to suspend all writes on the file system for a moment.But i can  allow all the reads to continue.That's all i need for now.

Is the problem that whatever filesystem you are working with does not
support block device freeze?

Greg
--
Greg Freemyer
Litigation Triage Solutions Specialist
http://www.linkedin.com/in/gregfreemyer
First 99 Days Litigation White Paper -
http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf

The Norcross Group
The Intersection of Evidence & Technology
http://www.norcrossgroup.com

-- 
Regards