how to ensure the cluster is still running if the heart beat link is down.

[kenken_3]

Hi,

I have a two nodes cluster with primary lan and heart beat link. If I am not able to have the 2nd heart beat link due to the lack of the lan card.

How can I ensure my cluster is still running if my heart beat link is down?

thanks

[Anshumali]

I hope you can use RS232...

else ..have a look at below threads..

Anshu

[Srikanth Arunachalam]

Hi,

If you do not have alternate Heartbeat link You are risking the availablity of cluster in the event of loss of primary heartbeat link.

You cannot ensure by any means that the cluster is still running, if the heartbeat link is down (it defeats the purpose of heartbeat channel).

If heartbeat link fails and NODE_TIMEOUT limit is reached, one of the node is determined to have failed and a cluster reformation commences.

Thanks,
Srikanth

Thanks,
Srikanth

[Srikanth Arunachalam]

Hi ,

Do make sure that you increase the NODE_TIMEOUT to a higher value. Otherwise, it will be difficult to identify the infinite loop of cluster reforming assuming the other node of the 2-node cluster is down.

If you set the NODE_TIMEOUT to a larger value, it gives you an option of manual intervention after initial failures.

Thanks,
Srikanth

[kenken_3]

Hi Skikanth,

Thanks for the information. If I changed the node_timeout value. Do I need to re-combile the cluster binary?

I have 2 options:

1) if I have a standby lan connected. Does it going to take over if the primary heart beat is down?

2) if I add in the 2nd heartbeat link now, do I need to re-compile the cluster binary again?

thanks

[Srikanth Arunachalam · Accepted solution]

Hi Kenken,

Yes you need to re-combile the cluster binary?

Of the 2 options:

1) Yes. But, it will be worthy to have 2nd heartbeat link for shifting the LAN network in case of failure of Primary LAN.

If you have one heartbeat configured, with the required standby LAN, and if NODE_TIMEOUT value is less than 4 seconds, you can reduce failover time if you configure multiple heartbeats instead. Since
heartbeat messages are sent over all heartbeat subnets concurrently, there will be no wait for network witching if a primary LAN fails. To avoid delays from busy networks, configure at least one private dedicated network for heartbeat.


2) You need to simply re-run the cmquerycl that will automatically detect the environment changes.

Thanks,
Srikanth

[kenken_3]

it is closed.