Hi all,

I am using corosync-1.1.2 on CentOS 5.3. I set token 200ms and hold 50ms in
corosync.conf. Then token_retransmit_timeout is auto calculated as 47ms.
Suppose very idle ring with two-node where there is no new message to send.
Think about the following scenario.

1. The node A sends a token to next node will start the reetransmit
timer(47ms).
2. The next node B receives a token and starts hold timer(50ms).
3. When hold timeout, node B will pass the token to node A.

The 50ms is larger than 47ms, so at node A the retransmit timer will expire
before receiving a token, as a result node A will resend the token until it
receives it. Assuming no token lost, then there will be at least one token
retransmission every token pass. Is my assessment correct? Or I am wrong? If
I am correct, then should I reduce the hold timeout?

Thanks
hj
_______________________________________________
Openais mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/openais

Reply via email to