05-23-2014 05:44 AM
Hi,
I have a problem with Windows 2012 netbackup 7.6.0.1 client (fileshare cluster) backup via media Win 2012 7.6.0.1 server.
Client and media server are Hyper-V machines running on the same hypervisor. No firewall between.
Policy type: MS-Windows
Storage: Basic disk - on media server - actually NAS attached and mounted on media server
Take check points every 5 min
Jobs per policy: 1
We treid both with Perform snapshot backups and without...
Each time we tried to make a full backup it ended up with status 13 after 2h.
bpbkar log on client side show this:
An Exception of type [SocketWriteException] has occured at:
Module: @(#) $Source: src/ncf/tfi/lib/TransporterRemote.cpp,v $ $Revision: 1.55 $ , Function: TransporterRemote::write[2](), Line: 338
Module: @(#) $Source: src/ncf/tfi/lib/Packer.cpp,v $ $Revision: 1.91 $ , Function: Packer::getBuffer(), Line: 652
Module: tar_tfi::getBuffer, Function: D:\NB\NB_7.6.0.1\src\cl\clientpc\util\tar_tfi.cpp, Line: 311
Local Address: [::]:0
Remote Address: [::]:0
OS Error: 10054 (An existing connection was forcibly closed by the remote host.
Solutions that we tried to use in this case are based on recommendations from:
http://www.symantec.com/business/support/index?page=content&id=TECH7963
http://www.symantec.com/business/support/index?page=content&id=TECH6127
We tried: running backups in different time, turned offload settings on NICs and hyper-v switch, changed timeout settings on client/server side, compared NIC drivers verions on other Win2012 hyperv servers (the same version), checked event logs on server (no record), restart client server,media server, switched to different hyper-v node, swithced client to different node (fileshare cluster)
Finally we found a way around by creating two separate policys - that take less than 2h each to finish up. That’s working.
What timeout setting could cause network timeout after 2h?
What could possibly cause this issue?
05-24-2014 05:48 AM
Does the event logs on the client show any errors at that time ? there have been issues with dll versions before which showed up as eventid 100 or 1000
Think the Windows tcp have a default timeout of 2 hours for idle connections
Regards
Michael
05-24-2014 10:40 AM
05-25-2014 11:54 AM
hmmm - check for firewalls. They typically have a two hour timeout.
06-11-2014 07:05 AM
Hm... Something is causing timeout. Right now I've found a way-around by splitting this backup into 2 policies - so they finish before that timeout.
Problem still exists though.