[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: NFS options



On Sat, Feb 04, 2012 at 10:40:22AM +0100, Petter Reinholdtsen wrote:
> [Giorgio Pioda]
> > Probably my i3 test machine is not the perfect setup to have an
> > idea about performance, but...
> > 
> > I see that Ubuntu as workstation client is WAY faster than DebianEdu
> > workstation. Probably tis is more related to Gnome2 vs KDE4, but in any case
> > my Ubuntu setup is running without NFS3 daemon.
> 
> Well, unless you can document that the Debian Edu workstation get
> significantly faster by disabling rpc.statd, I believe it is best to
> leave this change to Wheezy.  Ubuntu/Gnome is faster than Debian Edu/KDE
> is not really telling much about how rpc.statd affect performance.
> 
> If you believe rpc.statd is started by mistake by the package when only
> NFSv4 is used, is it not better to report a bug to BTS and ask for the
> Debian package to be changed?

Digging further into it ...

AFAIK rpc.statd isn't needed and doesn't run on tjener even if LTSP is 
used. For testing, I've set NEED_STATD=no and LTSP seems to work quite 
well. Maybe that's due to the fact that nfsv4 supports nfsv3 mounts.

(logged in as root / LTSP terminal): 
modinfo nfs

filename:       /lib/modules/2.6.32-5-486/kernel/fs/nfs/nfs.ko
license:        GPL
author:         Olaf Kirch <okir@monad.swb.de>
depends:        sunrpc,fscache,lockd,auth_rpcgss,nfs_acl
vermagic:       2.6.32-5-486 mod_unload modversions 486 
parm:           callback_tcpport:portnr
parm:           cache_getent:Path to the client cache upcall program (string)
parm:           cache_getent_timeout:Timeout (in seconds) after which the cache upcall is assumed to have failed (ulong)
parm:           enable_ino64:bool

 
Wolfgang

Attachment: signature.asc
Description: Digital signature


Reply to: