network polling intervel

[Vijeesh CTK]

what is maximum network polling intervel (NETWORK_POLLING_INTERVEL) we can give in cluster configuration file

Vijeesh CTK

[G. Vrijhoeven]

Hi,

The defauld is 2 secs.. Thus every 2 seconds, the cluster manager polls each network interface to make sure it can still send and receive information. Changing this value can affect how quickly a network failure is detected. Why would you like to change this parameter?

Gideon

[Vijeesh CTK]

for some experiment i would like to change the value to the max possible..can u tellme the maximum value possible for this parameter

Vijeesh CTK

[G. Vrijhoeven]

Hi again:

NETWORK_POLLING_INTERVAL


The frequency at which the networks configured for MC/ServiceGuard are checked. In the ASCII cluster configuration file, this parameter is NETWORK_POLLING_INTERVAL.

Default is 2,000,000 microseconds in the ASCII file (2 seconds in SAM). Thus every 2 seconds, the network manager polls each network interface to make sure it can still send and receive information. Changing this value can affect how quickly a network failure is detected. The minimum value is 1,000,000 (1 second). The maximum value recommended is 15 seconds, and the maximum value supported is 30 seconds.


look at :

http://docs.hp.com/hpux/onlinedocs/B3936-90065/B3936-90065.html


Gideon

[Uday_S_Ankolekar]

Hello,

This is the frequency at which the cluster identified LAN cards are checked by the cluster manager Daemon (cmcld) The default is 2 sec. Every 2 sec. The daemon checks the Lan card.
This can be changed ( Parameters would be in Micro seconds)
But why do you wants this to be changed..?

-USA..

[A. Clay Stephenson]

I would assume that it is a 32-bit signed integer and thus the maximum would be 2^31 - 1 or ~ 2.147 x 10^9 microseconds (2147 seconds). This would be far beyond any value that could possibly be considered reasonable.

[Helen French]

CTK =))

NETWORK_POLLING_INTERVEL:

The minimum value - 1 Sec
The maximum value supported - 30 Sec
The default value - 2 Sec
The maximum recommended value - 15 Sec

HTH,
Shiju

[Helen French]

CT =))

Again, check this for confirmation:

http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/hpux/onlinedocs/B3936-90065/B3936-90065_top.html&con=/hpux/onlinedocs/B3936-90065/00/00/30-con.html&toc=/hpux/onlinedocs/B3936-90065/00/00/30-toc.html&searchterms=network_polling_interval&queryid=20020419-074751

Kaise ho ? no mails ?

HTH,
Shiju

[Vijeesh CTK]

HI all



Thanks guys for the immediate replies...keep it up.. Gideon,ACS,Uday

Shiju...enthokkeyundu visheshangall..sugam aano..


I kept the parameter to 5 min and removed the network connection..then the system started behaving erratically..once it switched the package in 90 sec twice it switched in 220 sec..then the kept it back to my original configuration..

Thanks again Guys.

Regards,

CTK

[Stephen Doud · Accepted solution]

Hello Vijeesh,

Why do you want to alter the NETWORK_POLLING_INTERVAL? I recommend leaving it set to the default value.

I ask because I often talk with customers that want to do something to prevent a premature package halt/failover due to a flakey network. Since they are willing to live with temporary network hangs, they don't want the package to failover for a network outage unless they force it.

If this is your situation, I suggest you consider disabling ServiceGuard's ability to monitor SUBNET health in each package.

This is accomplished by commenting out the SUBNET reference in the package configuration file, and running cmapplyconf on it (with the package halted).

This in no way prevents the package from using the relocatable IP feature nor the cluster from operating the standby LAN feature. It merely causes ServiceGuard to ignore a SUBNET outage (temporary or otherwise).

Those customers who implement this feature are willing to be alerted to an extended network outage by some other means (users!).

See this document for more details: UXSGKBQA00000049
-s.