NETWORK_POLLING_INTERVAL

[Charles McCary]

Group, need your thoughts on something. The subnet that my cluster is on has had some problems in the past of going down for roughly a minute or so. My Network_Polling_Interval is still at the default of 2 seconds, and I'm thinking of changing this to 1 minute. The only downside to this that I can see is that if there is a lan card failure and not a subnet failure, I could conceivably keep the lan card from failing to it's backup for 1 minute. I don't see this as too big of a deal. Right now we are more likely to have a subnet outage than a card failure.

Am I missing anything on this?

tx,
c

[John Poff]

Hello,

We went through a similar problem here. After a production package crashed two or three times due to a subnet problem, we were asked by our management about monitoring the LAN in MC/SG and the timeout. Our reply was, "MC/SG is doing what it is supposed to do. Fix the network!" They fixed the network.

JP

[Helen French · Accepted solution]

Hi Charles:

The HP supported maximum value of this parameter is 30 seconds and the recommended is 15 seconds. The option you are trying may work, but read this thread and check Stephen's suggestion:

http://forums.itrc.hp.com/cm/QuestionAnswer/0,,0x302718276953d61190040090279cd0f9,00.html

HTH,
Shiju

[Charles McCary]

Guys,

thanks for the info. Stephen's post helped a lot. I will probably remove subnet monitoring altogether.

thanks
C