[nycbug-talk] nfs client fbsd 5.4

[max]

On top of this, may be you should consider afs or coda? NFS is pretty slow usually.
migration is a pain, but it would be worth it.


On Thu, Aug 11, 2005 at 01:12:51PM -0700, pete wright wrote:
> On 8/11/05, Steve Rieger <steve.rieger at tbwachiat.com> wrote:
> > 
> > i thought that i have the ability to mount nfs with the hard option
> > in my /etc/fstab but obviously i dont
> > 
> > and so i have a question for you people
> > 
> > am currently mounting all my websites via nfs off a san (about 1TB)
> > with the following options
> > nfs-server:/export/http /http nfs rw,noatime 0 0
> 
> 
> 
> looks like you are using more of a NAS topology instead of a SAN topology ;)
> 
> what can i do to make a (even slight) increase in speed/stat times,
> > am getting too many smarty issues, to be comfortable to sleep at night.
> 
> 
> first thing i'd due is check your read/write buffer size between the nas and 
> your client. Depending on the size/frequence of the data you are moving 
> maybe having 16k blocks or 32k blocks will give you better performance than 
> the default, which i belive is 8k. Also, check your network settings....are 
> you using gig-e between the nfs client and server? if so try using "fat 
> packets", some switches support MTU of 9600. this will improve performace 
> greatly for large data sets.
> 
> i'd also finally check how you have your disks RAID'd, are you happy with 
> your disk I/O etc..
> 
> HTH
> -pete
> 
> 
> 
> -- 
> ~~o0OO0o~~
> Pete Wright
> www.nycbug.org <http://www.nycbug.org>
> NYC's *BSD User Group

> _______________________________________________
> % NYC*BUG talk mailing list
> http://lists.nycbug.org/mailman/listinfo/talk
> %Be sure to check out our Jobs and NYCBUG-announce lists
> %We meet the first Wednesday of the month