I have an HP-UX server with a Network Appliance Filer attached over Gigabit Ethernet. I am noticing very slow response time when using "ls -l" on one directory on one of the several NFS mounted filesystems. The "ls" command by itself does not seem to be a problem. Typically I get a response within 30 seconds on this directory but sometimes it has taken as much as 12 minutes to respond!! There are less than 50 files in this directory and it is only nested a few directories deep. I have not noticed this with any directories in any of the other filesystems mounted in the same manner. All other directories will give me a file listing in less than 1 second. If I use the command twice in succession then the second time responds very quickly. I am assuming this is coming from cache.
I even tried mounting the same filesystem over NFS to another similar system and I do not experience this problem.
We do seem to be experiencing some sort of slow down in our database and I am not sure if it is related to this particular symptom or not.
I was not able to find anything on HP's knowledgebase nor Netapp's. Has anyone else experienced anything like this?
i have, though possibly not related to your problem.
i have a webserver which i run and have installed a perl module to stream mp3s. so all my mp3s are on my computer upstairs in my room. instead of moving everything, i set up an NFS export from my desktop computer. now, lazy person that i am i just dont feel like giving my desktop a static IP. when the lease on the ip from my router expires and i get a new, the NFS mount on the server goes stale.
when i try to do a directory listing inside the parent directory of the NFS mount, the shell locks up.
so have you checked to see if the NFS service is still running on the server-end ?
ps- my fix for the dynamic ip problem:
generally the ip address on my desktop only changes from 192.168.1.102 to 192.168.1.103 and back.
so, i have an NFS mount from BOTH ip address to the same place on the NFS client. i didnt know that would work and i actually dont know WHY that works...but it does
Yes... there're many causes for slow NFS performance...
Are your nameservice working well on your client ? Can it resolve hostnames with no problem ? Are you using the hostname or the IP address to mount the share ? Try to use the IP address...
Are the machines on the same network ? Check for errors on both the server and the client... use a tool like tcpdump to check for NFS traffic...
