On Wednesday 10 March 2004 04:49, Errol Casey wrote:
On Tue, 9 Mar 2004, Magnar Stav Johanssen wrote:
Hi
I'm currently running OnTAP 6.4.3 on my 760c. I've got 3 servers (Solaris 9) running postfix with high volum of mail. These 3 machines alone uses all free CPU on one of the nodes in the NetApp-cluster.
Are you only running the MTA on the client machines , or are you also possibly using qpopper also on your client machines?
The clients are only running MTA!
This is our mount-options for the clients: rw,nosuid,intr,grpid,proto=udp,xattr,rsize=8192,wsize=8192
(The performance was even lower when mounting: rw,nosuid,intr,grpid,proto=udp,xattr -> nfsv3, rsize = wsize = 32768)
On Tuesday 09 March 2004 23:22, Neil Stichbury wrote:
How are your minra and atime settings ?. What are your rsize and wsize ?. Using tcp or udp ?.
As you can see of the mount-options used by the clients: - rsize and wsize are 8192 bytes - we uses udp.
Setup of the filer:
filer> vol status -v Volume State Status Options vol0 online normal root, diskroot, nosnap=off, nosnapdir=off, minra=off, no_atime_update=off, raidsize=14, nvfail=off, snapmirrored=off, resyncsnaptime=60, create_ucode=off, convert_ucode=off, maxdirsize=10240, fs_size_fixed=off, create_reserved=off, raid_type=RAID4
Plex /vol0/plex0: online, normal, active RAID group /vol0/plex0/rg0: normal RAID group /vol0/plex0/rg1: normal
filer> options nfs. nfs.ifc.xmt.high 16 nfs.ifc.xmt.low 8 nfs.kerberos.file_keytab.enable off nfs.kerberos.principal nfs.kerberos.realm nfs.locking.check_domain on nfs.mount_rootonly on nfs.mountd.trace off nfs.per_client_stats.enable on nfs.require_valid_mapped_uid off nfs.tcp.enable on nfs.udp.enable on nfs.udp.xfersize 32768 nfs.v2.df_2gb_lim off nfs.v3.enable on nfs.v4.enable on nfs.v4.id.domain *****.no nfs.webnfs.enable off nfs.webnfs.rootdir XXX nfs.webnfs.rootdir.set off
Regards, Magnar