[NTLUG:Discuss] NFS Server Timeout

[agoats at compuserve.com]

Since you had a power outage, that may have corrupted your hard drive 
enough to cause a long search time during look up. The length of time is 
very similar to a disk drive delay, ranging up to 30 seconds.

I recently (past week) had  lightning strike nearby that cut off power 
and cause some corruption of my own system. Weird things like having to 
continually move my mouse in order for downloads to function, internet 
connections with firefox having latchup up issues (won't move to the 
item clicked, hard drive light on hard and bright).

Re-seating the power AND data cables on the SATA drives helped. I run 
ext2 file systems (my preference). Slackware uses fsck as the default 
for checking the file system after a crash, but this doesn't seem to be 
sufficient, like it does a file system check but does minimal work to 
repair anything. Running e2fsck -v -y -D on the drive found many errors, 
cross linked files and many other issues. On my home directory 
partition, e2fsck found so many errors that it had to restart itself 
twice. Once complete, everything is working better. Some files that were 
damaged were some of the initiation and configuration files, which could 
be your problem, i.e. the nfs configuration files may be damaged.

Only solution for fixing/replacing the damaged files is to re-install 
the software and update to current rev on the software.

Alvin




On 10/22/2015 09:50 AM, George Lass wrote:
>        From: Eric Schnoebelen<eric at cirr.com>
>   To: NTLUG Discussion List<discuss at ntlug.org>
>   Sent: Wednesday, October 21, 2015 6:27 PM
>   Subject: Re: [NTLUG:Discuss] NFS Server Timeout
>
> George Lass writes:
> - Following a recent power outage that took down my NFS server I am
> - no longer able to mount anything on that server, even from a client
> - on the same local network.  On the client side the mount returns:
> -Â  mount: RPC: Timed out
> -
> - Running tcpdump on the server, I discovered that the mount response
> - is not being sent until some 38 seconds after the mount request
> - was received.  The client is timing out after 20 seconds.
>
> [...]
>
> - Any ideas?
>
> It sounds like you've got a networking issue, possibly address
> lookup, etc.
>
> Tracing the execution of mountd or rpcbind might provide some
> insight.
>
> Do you have hostnames in your exports file?
>
> If you are using hostnames, do you have working reverse address
> maps?
>
> Do you have working hostname to IP address maps?
>
> --
> Eric Schnoebelen        eric at cirr.com        http://www.cirr.com
> Â Â Â  Â Â Â  Seen on a T-shirt in Pasadena:
> Â Â Â  Â Â Â  Â Â Â  FAILURE IS NOT AN OPTION -
> Â Â Â  Â Â Â  Â Â Â  it comes bundled with the software.
>
>
>  From a network standpoint, I believe all is well. No firewalls, I can ping the server from the client and vice versa. /etc/exports only contains IPv4 addresses (no hostnames), and running arp on either client or server shows both IP address and hostname for the other. Â What I am trying to focus on right now is the fact that there is a 10 second delay running rpcifo -p and showmount -d times out after 20 seconds even when running them on the server itself. Â This pretty much rules out any network / name resolution issues.
> George
> _______________________________________________
> http://www.ntlug.org/mailman/listinfo/discuss
>
>
>
> _______________________________________________
> http://www.ntlug.org/mailman/listinfo/discuss