ll command take an abnormally long time, to read files from an NFS directory.

[Ahmed ABDOU_1]

Hi
I have problem can you help.

System information

system k570 ( NFS server)
system k460 (NFS client)

Problem: I have SAP/R3 environment and I've noticed that my transports are taking much longer than they have in the past. They used to require 7 hours. Now, they require about 40 minutes. Also, they sometimes abort.

In addition, the ll command takes an abnormally long time or hang, to read files from an NFS directory.

Can you please help me in this issue?

Thanks in advance

Regards,
Omar

[Printaporn_1 · Accepted solution]

the possibility of problem can be

1. Network performance
2. nfs server performance , increase number of nfsd in server


can start from nfsstat more information see manpage

[JACQUET]

Hello,

I've already encountered that kind of problems : usually, such problems is linked with the number of files in the NFS server directory. How many files do you have in /usr/sap/trans ? Are you doing lot's of SAP Transport Every Day ?
First, you should analyse datas in the /usr/sap/trans directory and check what is useful and what can be archive in order to free inodes on NFS Server.
Otherwise, you should modify some parameters such as number of nfsd daemons for the nfs server (/etc/rc.config.d/nfsconf) and the biod daemons on nfs clients servers.
You could find some informations at :
http://docs.hp.com/hpux/onlinedocs/B1031-90048/B1031-90048.html

Regards.

PJA

P.S : If NFS Server use HFS File System, you could eventually set "CacheFs", which can improve performance only in reading data over NFS.

[Mark Vollmers]

Omar-

We had a similar problem with NFS mounted folders on an NT. For us, it turned out that the T1 line we were using was dropping packets like there was no tomorrow (partially broken connection), and this was causing serious problems when we tried to list or move files. I'd do a full check of the network (switches, lines, etc). Assuming that you have not changed anything on the server (settings, etc), it sounds like it might be networking.

Mark

[Deshpande Prashant]

HI
Try increasing nfsd and biod daemon parameters.
NUM_NFSIOD and NUM_NFSD in /etc/rc.config.d/nfsconf file from 4 to 8.

Thanks.
Prashant.

[Brian Hackley]

Omar,

I've posted this answer quite often but you might find it helpful to you.

We have a variety of docs on ITRC to help you out. Look at NETUXKBRC00006283 and KBAN00000261 . Also we have an excellent white paper on NFS performance at http://docs.hp.com -> Networking and Communications -> NFS -> White Papers

Good luck,

-> Brian Hackley