ndd settings 1

[RenC]

I just had a backup of the Master server fail with a "network timeout" error. As I was troubleshooting, I ran a netstat -a|grep bpcd and received pages of TIME_WAIT messages. Does anyone know of any recommended ndd settings I should have in place for netbackup for time_wait intervals...or any other ndd settings. We are just getting started with NetBackup and we have not tuned any of these parameters yet.

Thanks for your help.

Renee

 

[PGPhantom]

Check out my FAQ - faq776-3124, this will list the settings etc. If you are getting time out errors - Your DNS may not be 100% for resolution, use the bpclntcmd utility to vrify connectivity.

 

[RenC]

PGPHantom,

Thanks for your response, but this FAQ seems to just be about Buffer sizes. Those have been set. What I am looking for is recommended ndd parameter settings. I have adjusted the tcp_time_wait_interval from the default of 24000 down to 1000 alerady. I am not sure if this is the best setting or not, and I was wondering if there were any others I should look at.

As far as dns, nslookups, reverse lookups, and the bpclient command all seem to indicate there aren't any dns problems.

Renee

 

[PGPhantom]

The only one I change is in Windows and is the TCP_Timed_Wait_Interval - I change this to a dword of 64.

 

[rdapilmoto]

We are primarily a Sun shop in my company and we are running Solaris 8 on our systems. By default, the ndd tcp_time_wait_interval setting is 240000. We have changed this setting to either 30 or 60 seconds to help out with the TIME_WAIT messages that you have been seeing....

ndd -set /dev/tcp tcp_time_wait_interval 30000

If you want to make this a permanent change on your Unix client, you can vi the /etc/init.d/inetinit file and place this command in it.

Hope this helps!