We ran into this recently with a newly implemented load balancer.  A time out 
setting was disconnecting after a certain period of time.  The time was long 
enough for incrementals and some files systems to complete.  It was not long 
enough for the larger savesets to complete so we only noticed this on full 
backups.

Check and see if the runtimes on your saves are always terminating at after the 
same amount of time. 

 

-----Original Message-----
From: EMC NetWorker discussion [mailto:NETWORKER AT LISTSERV.TEMPLE DOT EDU] On 
Behalf Of Andy Fahy
Sent: Sunday, May 27, 2012 10:15 AM
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] RPC send operation failed; errno = Broken pipe

I agree that the keep alive time setting should be added. Try that and restart. 
I can't remember if a reboot is required for the os settings to take effect.
Hope it works for you.
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.

_____________________________________________
From: Yaron Zabary <yaron AT ARISTO.TAU.AC DOT IL>
Sent: Sun May 27 02:35:54 GMT+01:00 2012
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] RPC send operation failed; errno = Broken pipe


Check logs on the client for hardware issues (dmesg, /var/log/messages, 
/nsr/logs/*). Then, if all seems well, run a tar on the file system in question 
and make sure it is able to walk the entire file system.

tar cvf - / | cat > /dev/null

Another issue might be a firewall between the client and server with a 
connection table time out too short for your session. The solution is to 
increase the timeout (if possible) or enable TCP keep alive on your server. 
This is my setup:

# more /etc/rc2.d/S93nettwicks
#!/bin/sh

# Helps to solve RPC errors under heavy load.

/usr/sbin/ndd -set /dev/tcp tcp_conn_req_max_q 1024 /usr/sbin/ndd -set /dev/tcp 
tcp_conn_req_max_q0 4096

# Should help with backups thru firewall

/usr/sbin/ndd -set /dev/tcp tcp_keepalive_interval 600000



On 05/26/2012 09:41 PM, ruhailmushtaqvakil wrote:
> HI,
>
> My Networker Server is Solaris and my client is linux.
>
> I am trying to run a backup on this client however it fails everytime.
>
> The backup starts nice, and backs up some 110 GB data and then says lost 
> connection to server.
> it just retries and then fails, tried running manual, same result.
> I m getting connection lost message evry now and then. the total data 
> on system is 275 gb, one saveset only, backup speed is good 
> throughout. almost 25-50 MBps
>
> RPC send operation failed; errno = Broken pipe
>
> this is another error i got from daemon.og
>
>
> Any help is appreciated.
>
> +_____________________________________________

> |This was sent by mytechnicalknowledgebase AT gmail DOT com via Backup 
> Central.
> |Forward SPAM to abuse AT backupcentral DOT com.
> +_____________________________________________



-- 

-- Yaron.