SLP and error 84 with large backups

[PaulN]

I have 7.1.0.2 installed.

all my SLP replications work fine except for backups that are large, over 500gb or so... They timeout with error 84.   Im not sure what to try next.

Anyone got any insight to what is going on with the large files?

[PDragon]

These can sometimes be troublesome to sort out. I know because I had these issues when replication data between our Data Domains in different locations. We would normally see these errors when (as you) we would replicate large files across the OC3 connection between sites. The cause was two things for us. 

1: The amount of data going at one time (We had to stagger other traffic connections around the backup replication/duplication)

2: The MTU settings on the receiving Data Domain needed to be increased. 

Once this was done, the errors went away.

The way I figured this out was Enabling the view of the following logs:

Unix: /usr/openv/netbackup/logs/bpdm - Also check syslogs

 Or if your windows I believe its

Windows: <install_path>\NetBackup\logs\bpdm - Also check event viewer

I went looking for disk errors on the media servers logs & master. Along with looking network errors of any kind. This also took the help of my Network Admins to see it they saw errors on the ports, VLANS etc...etc...

Hope this helps as i know it was frustrating for me. 

[Lee_C]

Hi Paul,

I had an issue recently where we had to do a re-seeding task of around 200 GB over a 2Mbit/s link.  Although the error 84 occured every 12 hours, job threads were still running and the "effectiveness" of the data transfer appeared to degrade. 
(http://www.symantec.com/connect/forums/error-84-observed-after-12hrs-slp-duplication)

The release notes for 7.1.0.2 indicated that a similar issue should be resolved by 7.1.0.2, so I was wondering how long your backups ran for before you experienced the error 84.

During the issue we wanted to know the pending duplicates so ran nbstutil report

Thanks

[PaulN]

ive been playing with the network settings with no resolve.  i also have backups running over a WAN that take 20hours to run, backing up to a non dedup pool, and it runs fine. I also have SAN replication running over a WAN and that never fails.

its only the dedup option, with large files.  the last one i tested ran for 9 hours and died, then on every retry it fails after 2 minutes.

right now my only option is to take these large backup jobs, have netbackup back it up to a non-dedup pool, and then have my SAN replicate it over.

(ive open 2 tickets with netbackup support on this issue with no resolve.  so in my point of view, this is a huge bug, considering it wont even resume after it fails the first time....)

ive ran netflow, span, and wireshark on all the port hops and cannot find anything that says the tcp connection times out...

right now im at a loss, dont know what to do from here.  anybody have any suggestions??