1

I have a RAID system on A server and similar configuration on B server-both with Scientific Linux 5.5. They are connected with infiniband and RAIDS from A are mounted with NFS to ServerB. Now when the users are accessing A-Raids via B-Server, the network getting stacked already with 2 rsyncs. What is the reason? Can I some how bench it and tune the NFS?

kind regarads

Arman.

Arman
  • 555
  • 2
  • 8
  • 18

1 Answers1

1

So you've got mirrored NFS servers and some mechanism for keeping them in sync using rsync? This is a recipe for problems. It also implies that you've got a homegrown script trying to keep it all working. Certainly NFS is the last of your problems.

Ideally, you should consider using a shared storage backend with a cluster filesystem e.g. GFS). Alternatively you might consider something like AFS (which manages its own replcation and tells you when problems arise).

C.

symcbean
  • 19,931
  • 1
  • 29
  • 49
  • Thanks for response, Yes, we have a datacenter, managed with both machines. We are trying to use GlusterFS, but it still under the tests..to this point the access to the mounted filesystem is 10 times slower:( via NFS – Arman Sep 03 '10 at 12:47
  • So have you checked if the problem is with NFS or the other stuff going on here (try comparing local access) – symcbean Sep 03 '10 at 14:29
  • Yes we I did, if I write on one raid via nfs it is 90Mb/s but local one writes with 350Mb/s. I never try GFS, Is GFS faster than NFS4? – Arman Sep 04 '10 at 05:38
  • 1
    Apples and oranges - NFS is a file sharing protocol which runs on top of just about any filesystem. GlusterFS (a replicated filesystem) is NOT the same thing as GFS (a shared filesystem). Suggest you do some reading. – symcbean Sep 06 '10 at 08:49