Re: Still getting NFS lockups

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

 



On Tue, 20 Jan 2004, Robert Mohrmann wrote:
> 
> The other 3 servers can mount /home and /data just fine and
> they all have: timeo=14,rsize=8192,wsize=8192,intr
> as thier mount options for those 2 directories.
> 
> This morning I came in and shells on all of the 3 servers that
> are mounting to /home and /data would lockup hard if
> I accessed /home or /data

Your NFS server had died.. or become unreachable

> I did a fuser -m /home and a fuser -m /data and both had
> quite a large list of processes.

These are processes that had been trying to access the filesystem but 
couldnt as the filesystem was not responding and not telling the processes 
that there was an IO error.

> I had to restart the server running the /home and /data
> directories and it hung up shutting down the nfs services.
> So I gritted my teeth and hard cycled the power.
> Fortunately everything came back up.

if it is mounting the NFS exports locally then you may have had an issue 
with NFS dieing and this would also cause lockups on the export server.

> But I cant continue to have NFS lockups like this.

You may wish to investigate using soft mounts instead of hard mounts as 
this should aleviate some of the problems - however, this also seems to be 
a common problem when using NFS - simple answer, dont use NFS, it is evil 
and the work of the devil ! (pass the keyword "soft" as one of the mount 
options)

also, 'man nfs' and investigate some of the other options there - you may 
find the some of the defaults are a bad idea (the retry= keyword that 
causes around a weeks worth of retrys (by default) before a timeout is 
generated seems like a bad thing [tm] as well)

> Ive done alot of googling and other searches and Ive
> read lots of documents but I can not seem to find anything
> to point me to the root of this problem.

The root of the problem is that your NFS server went away (I am guessing, 
probably the NFS server process died) - NFS tends to assume a 100% 
reliable environment and is not very forgiving if this is not the case.



-- 
Steve.


-- 
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list

[Index of Archives]     [CentOS]     [Kernel Development]     [PAM]     [Fedora Users]     [Red Hat Development]     [Big List of Linux Books]     [Linux Admin]     [Gimp]     [Asterisk PBX]     [Yosemite News]     [Red Hat Crash Utility]


  Powered by Linux