Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


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

Re: NFS performance question



 I doubt that there is a problem with the switch since I just upgraded a 
server inside the room and now have the same problem. I have not had 
the time to analyze it yet in detail. RHEL4 by default, AFAIK uses NFS 
V4 that does default to TCP.  What I am thinking is possibly an NFS 
lock issue.  There are some nfs debugging tools that come with RHEL, in 
addition, one test I want to do is to serve a directory from one of the 
U6 servers, and see what the situation is.  Since we are gpoing to 
upgrade the NFS server to U6 as soon as we receive some drives we 
ordered. 
Here is our current situation: 
NFS Server running RHEL4 U3 Intel 4x2 Xeon w/8GB memory. 
NFS Client running RHEL4 U3 NFS performance is good. (In the test, 
saving the test file was nearly instantaneous. 
NFS Clients (4) RHEL4U6 Saving the same data takes over 2 minutes. This 
indicates there is some incompatibility between NFS on Update 6 and 
Update 3. 
The first thing I'm going to try is to export from one of the Update 6 
servers and see what the performance is U6 to U6 and U3 to U6. If it 
appears that using U6 as a server I should be able to physically move 
the drives from the U3 to the U6 server, and just change the hostname 
and IP address. If the problem exists between U6 and U6, I'll need to 
really analyze the problem. 

On Fri, 11 Jul 2008 17:23:50 -0400 
"Ben Holland" <[hidden email]> wrote: 

> Perhaps you have a bad port on the switch? Try plugging the computer that 
> you get 2 seconds into the other jack and vis a versa?  Have you looked at 
> packets getting dropped? I may also be that you are flooding your server 
> with packets and I believe NFSv3 is not tpc by default so it's basically 
> best try. Are you running these with little else on the server? ~Ben 
> 
> On Fri, Jul 11, 2008 at 3:49 PM, Jerry Feldman <[hidden email]> wrote: 
> 
> > On Fri, 11 Jul 2008 13:25:28 -0400 
> > Derek Atkins <[hidden email]> wrote: 
> > 
> > > Jerry Feldman <[hidden email]> writes: 
> > > 
> > > > 1. NFS client in server room RHEL4U3, connected directly to the switch. 
> > > > 2. NFS client in office RHEL4U6 connected to a 1GB switch that is 
> > > > connected via an office patch panel to the switch in the server room. 
> > > 
> > > Are you sure that the switch-to-switch link is at 1Gb and not 100Mb? 
> > > 
> > > Also, what else is traversing this switch-to-switch link? 
> > 
> > I've eliminated the switch-to-switch link from the mix. I just 
> > installed RHEL4U6 on one of the servers in the server room. The save 
> > time for an 8MB file is over 2 minutes where the save time on an 
> > adjacent server is under 2 seconds. (clock time). The NFS server and 
> > the 2 clients are sittinng next to eachother and are connected to the 
> > same switch. I have not had a time to read Mat's message yet, and I 
> > have a conference call in 5 minutes. 
> > 
> > 
> > -- 
> > -- 
> > Jerry Feldman <[hidden email]> 
> > Boston Linux and Unix 
> > PGP key id: 537C5846 
> > PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB  CA3B 4607 4319 537C 5846 
> > 
> > _______________________________________________ 
> > Discuss mailing list 
> > [hidden email] 
> > http://lists.blu.org/mailman/listinfo/discuss
> > 
> > 
> 
> -- 
> This message has been scanned for viruses and 
> dangerous content by MailScanner, and is 
> believed to be clean. 
> 


BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org