07-17-2013 09:12 AM
Hello
We are getting some
A timeout (30000 milliseconds) was reached while waiting for a transaction response from the NetBackup Job Manager service.
in system event log on our Netbackup 7.5.0.5 master server running Windows 2008 R2 SP1 MSCS which correlates with get_string() failed - network read error (10053) in the problems report.
All the Netbackup exe files and folder should be excluded from the anti virus
I already have a case on the getstring issue with Symantec, but thought I would ask on the forum too
Regards
Michael
07-17-2013 09:43 AM
Michael, have you tried temporarily turning the antivirus off to double check?
07-19-2013 07:43 AM
try increasing the default timeout value from 3000 milliseconds to 6000 millieseconds.
Steps for increasing it are given in the Microsoft article.
http://support.microsoft.com/kb/922918
07-25-2013 08:17 AM
Yes have had killed the anitvirus processes and was still getting the error
Not sure it is a good idea to raise the timeout for services to a full minute.
07-25-2013 09:02 AM
It is a very busy system?
It could be that the Master is running out of tcpip ports
A netstat -a will show the state of your ports - if a lot are in a TIMED WAIT state then add a registry key:
Under HKLM\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\
add a new DWORD named TcpTimedWaitDelay with a decimal Decimal Value of 30
You can alsoincrese the number of ports available as WIndows 2008 only has 16383 available. You can increase this using:
netsh int ipv4 set dynamicport tcp start=1500 num=63000
from an administrative command prompt - this takes effect immediately, the registry key needs a reboot - set these on both nodes of the cluster
Hope this helps
07-26-2013 06:49 AM
Assume you mean TIME_WAIT state ? Have about 500 right now of about 1150 connections, guess there will be more connections at night when the big backups run
Have already increased the number of dynamic ports on our master and media servers to 50000
07-26-2013 07:00 AM
Yes sorry - the default on Windows is 4 minutes - not far off the NetBackup default timeout of 5 minutes!
The registry entry i gave reduces that to 30 seconds to keep the system cleared down quickly so that you dont run out of ports or get unneccesary timeouts.
If you are using advanced disk or de-dupe then, just in case it is waiting for responses from the Media Severs it is also worth adding the following to the Master and all Media Servers:
Unix:
/usr/openv/netbackup/db/config/DPS_PROXYDEFAULTRECVTMO
Windows:
<install path>\veritas\netbackup\db\config\DPS_PROXYDEFAULTRECVTMO
Open it up and enter a value of 800 into it
Hope this helps
07-28-2013 12:35 PM
08-07-2013 03:47 AM
The problem has persisted over multiple reboots.
Hoping that Symantec Support find the reason for the get_string timeouts