Placing paging file on clustered disk

  • Thread starter Thread starter Paul
  • Start date Start date
P

Paul

Can this be done? I am having Error 51 problems on my
clustered boxes and I think it is due to the fact that my
page files are not big enough. I am running 4 gb of ram
and need 6 gb in paging files. Since you can't have more
than 4 gb per disk, I have to use another disk. But my C
drive on one of the boxes is not going to be able to
handle this increase in used space and all I have left is
to either use one of the clustered drives or install a new
drive.

Please let me know.

Thanks,
Paul
 
I doubt that it is caused by the paging file size (you would be getting
other event log entries relating to Virtual Memory problems).

Have you decoded the error (for instructions see:
http://support.microsoft.com/default.aspx?scid=kb;en-us;244780
)?

Error 51's can be caused by lots of things (SCSI bus resets in the middle of
a write for example) and are not necessarily indicative of a total failure
(e.g. they may be logged if a write fails, but will be re-tried).

You don't necessarily need 1.5x RAM for a page file size. If you have less,
then it means that you won't get a full memory dump in the event of a
failure, but it won't necessarily impact normal operations (you would need
to monitor the server for the memory actually in use).

As to cluster use, a paging file can be on a clustered disk. That said, I
am not sure about the implications of having a paging file that spans both a
clustered and non-clustered disk.

Pat
 
Thanks for the info. My error is show data of 10, 20, 30,
and 40. Given the number of these errors I am getting
and the fact that the cluster eventually fails over leads
me to believe maybe I have a controller problem. Not
really sure. The fact that I am having the same behavor
on both clustered servers serving the same shared disk
array also leads me to beleive that maybe I have a hard
disk problem. Basically, the one array of 12 disks is
spilt into 2 logical drives and then one clustered server
hosts one drive and the other server hosts the other
drive. One other thing I get alot of is network errors
when this type of thing occurs. I have run chkdsk /r on
both logical drives and I am running Diskkeeper 8.0 which
smart defrags the drives and that is reporting no
problems after I have run the chkdsk /r. I am running
out of options here. One problem that I am having on one
of the clustered servers is I am running a Dell server
that has a memory riser board that seems to be flakey
with a couple of the memory slots. I don't know if that
is possibly causing my problem or not. Any suggestions
would be appreciated.

Thanks,
Paul
 
Hard to say where to start, given the number of suspects. I would probably
try getting with Dell and seeing if there are any driver, firmware, and/or
BIOS updates to help out. They may also have a more specialized diagnostic
than ChkDsk.

One of the problems w/running ChkDsk is that it relies on the underlying
drivers/hardware to tell it if things are OK. If there is a problem on the
hardware such that it reports false negatives (i.e. a block or read/write is
OK when it really isn't) then there isn't really anything ChkDsk can do.

Pat
 
Unfortunately, Dell has not been much help. I have run
all their diags and also loaded the latest drivers and
firmware up. I went ahead and added a new hard drive and
more page file space. Maybe this will help. It is a long
shot, but it is all I have left from my point of view.

Thanks,
Paul
 
Back
Top