hi lew

to find out which machine is the one holding up
your network... ( the one timing out )
        - open up 2 xterms
-->>    - do "df" in one of um ... it should hang if you have 
        a network problem...

-->>    - in the other xterm, "cat /proc/mounts"

        - the listing in both xterms should agree
        and the next line after the last server listed in df
        is the one that is hanging... the one your machine
        cannot find

- since you have it hard mounted... you cannot ctrl-C that hung df
  process....
        - and you'd probably have tons of hung processes ???
        ( df, ls, ..etc...
                ( see what is listed with "ps axuw"

        - you probably have a bunch of stale mounts and zombies ???

- you should be able to kill all those hung process..maybe not

        - easiest way to clean up ...
                - send email to all those logged  - or -
                - shutdown -rf "clean up hung nfs"  time+10
                        - ctrl-C the job .. its just to let them know
                        the machine is being rebooted ...
                - init 1  before the system reboots
                - init 3  to bet back online 


c ya
alvin
http://www.Linux-1U.net .... 1.6TeraByte 1U Raid5[tm] ... 8x 200Gb each 


On Thu, 13 Dec 2001, Lew E. Lefton wrote:

> Hi,
> 
> I have been having some problems recently with our NFS setup.  My Linux
> clients seem to hang for several minutes and I can't figure out what is
> causing the problem. I know the "can't get a request slot" and "server not
> responding"  errors which are appearing in our logs indicate a network
> problem due to RPC timeouts, but, the NFS server does not seem to be
> overloaded and the NFS mounts to Solaris boxes on the same network are not
> having trouble. Our NFS server is running Solaris 8 and the clients are RH
> Linux 7.1.  I am running 2.4.9-12 kernel and nfs-utils-0.3.1-13, both
> from Redhat on the client.
> 
> I have "direct" nfs mounts in /etc/fstab with options
> rw,nodev,rsize=32768,wsize=32768,hard,intr,actimeo=10
> 
> I also have an autofs entry (/home) with options
> rw,nodev,rsize=32768,wsize=32768,actimeo=10
> 
> How can I tell what mount is causing the hang?  Any other 
> troubleshooting suggestions?  I am still not able to reproduce the problem 
> on demand but it is occurring several times a day.
> 
> Thanks in advance for any help.  Please reply to me as well as the list.
> 
> Cheers,
> Lew Lefton
> 
>  -----------------------------------------------------------------------
> | Lew Lefton, IT Director         | Phone:     (404) 385-0052           |
> | School of Mathematics           | FAX:       (404) 894-4409           |
> | Georgia Institute of Technology | e-mail:    [EMAIL PROTECTED]  |
> | Atlanta, GA  30332-0160         | http://www.math.gatech.edu/~llefton |
>  -----------------------------------------------------------------------
> 
> 
> 
> 
> 

Reply via email to