[CentOS] NFS v4

Ross S. W. Walker rwalker at medallion.com
Wed Jun 6 16:29:42 UTC 2007

> -----Original Message-----
> From: centos-bounces at centos.org 
> [mailto:centos-bounces at centos.org] On Behalf Of Jean Figarella
> Sent: Wednesday, June 06, 2007 12:14 PM
> To: CentOS mailing list
> Subject: Re: [CentOS] NFS v4
> Les Mikesell wrote:
> > Jean Figarella wrote:
> > 
> >>>> Under real user load though it just fell on its arse very
> >>>> quickly and a quick fall back to v3 was implemented. Maybe the
> >>>> EL5 implementation behaves better?
> >>>
> >>> I use one of my CentOS 5 boxen for low-load fileserver with
> >>> NFS4. I've been noticing that although MP3s plays fine from
> >>> that box, it can take a minute or two to write a dozen
> >>> tags--something that should happen in the blink of an eye. I've
> >>> been wondering what the problem is. Maybe it's NFS4!
> >>>
> >>
> >> I've been experiencing 5 to 20 seconds freezes ever since 
> I upgraded 
> >> my nfs server. This serves about 30 people. Anytime 
> someone writes a 
> >> big file to their home directory, the freezes happen. You can 
> >> experience the freezes when browsing the internet, I guess because 
> >> firefox is constantly writing data (cache, cookies, etc...). Other 
> >> apps freeze up too.
> >>
> >> Now the question on my side is: can I force centos4 for 
> serve nfs3 as 
> >> opposed to nfs4. If I am not mistaken this is dependent on 
> the client?
> >>
> >> The reason Im asking is that if I can force the server to 
> serve nfs3 
> >> then the clients will default to nfs3 themselves. right?
> > 
> > Are these mounted with the sync or async option?  If it is 
> sync, the 
> > client waits for the physical flush to disk to happen on every 
> > operation, something that is horribly slow and that you 
> never do for 
> > local disks except for transaction boundaries in databases 
> and other 
> > unusual situations.
> > 
> At first the sync option was being used, and the freezes 
> happened very 
> frequently. About a month and a half ago I started using 
> async, and now 
> I get far fewer freezes, like maybe 3 in an 8hr span.

You can also couple the async with the anticipatory scheduler
so reads will get higher priority then writes which can help
readers during a page-cache flush.


This e-mail, and any attachments thereto, is intended only for use by
the addressee(s) named herein and may contain legally privileged
and/or confidential information. If you are not the intended recipient
of this e-mail, you are hereby notified that any dissemination,
distribution or copying of this e-mail, and any attachments thereto,
is strictly prohibited. If you have received this e-mail in error,
please immediately notify the sender and permanently delete the
original and any copy or printout thereof.

More information about the CentOS mailing list