Description

This article provides steps needed to increase LLT peerinact value for Veritas Cluster Server (VCS) on Windows platform.  Increasing LLT peerinact will allow more tolerance for network outages in VMWare environments or slow networks to help prevent split brain.

The following Windows System Event log error may be reported:

LLT INFO V-14-1-10205 link X (AdapterX) node X in trouble

LLT INFO V-14-1-10032 link X (Adapter0) node X inactive X sec 

LLT INFO V-14-1-10509 link X (AdapterX) node X expired

Note: The value set for peerinact will impact the amount of time VCS waits before declaring a node as FAULTED for system down or node reboots. The peerinact value should be lower than the time it takes for a server to reboot or groups will not failover on reboot.

1. Stop HAD with force leaving applications online. Run the command below on any node.
# hastop -all -force

2. Stop Veritas heartbeat services. Run the command below on ALL nodes.
# net stop llt /y

3. Edit llttab.txt on ALL nodes to add new peerinact value.
- Go to \Program Files\Veritas\comms\llt\llttab.txt
- Add line "set-timer peerinact:3200" at the end of the file. Do not include quotes

This example will set LLT timeout to 32 seconds. Values are seconds with two 00s at the end. For example, to set peerinact to 40 seconds, use value 4000 in llttab.

4. Start VCS services. Run the command below on any node.
# hastart -all

5. Confirm the new value with the command below.
# lltconfig -T query | findstr peerinact