CLOSE_WAIT question

[gallows]

I'm running Solaris 9 with latest Recommended Patches. We are using NIS+. I am seeing a lot of CLOSE_WAITs (netstat) on almost all of my servers coming from my NIS+ server. Does anyone have any ideas on what I can look at to correct this? It is my understanding that the only way to get rid of them is to reboot, which is not an option for me.

Thanks

 

[daFranze]

what is the value of tcp_time_wait_interval?

ndd -get /dev/tcp tcp_time_wait_interval

afaik if you are running a http Server or a server with many short connections you should set this to eg 30000 (miliseconds)

ndd -set /dev/tcp tcp_time_wait_interval 30000

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[gallows]

Thanks for the reply Franz!
The server that has about 7 Close_Waits has the interval set to 240000. It has an application on it that uses Oracle. This server is running Solaris 5.8

The Nis+ server has the interval set to 60000. This one is Solaris 5.9. Seems like quite a bit of difference between the two.

 

[daFranze]

I suggest to set it to 60000 on the Solaris 7 host; I think this will not affect the applications or Oracle; even if it affect's it, you know what you have tuned...

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[gallows]

Thanks Franz!
I just thought the big difference with the default settings for Solaris 8 (240000) and Solaris 9 (60000) was interesting. It would seem to me that I would need to increase the 60000 to 300000 to get rid of the WAITS. I will play around both ways and see what happens.

gallows