[ClusterLabs] Q: corosync windows_size: 256000 / netmtu

Ulrich Windl Ulrich.Windl at rz.uni-regensburg.de
Tue May 12 13:08:40 UTC 2015


Hi!

Can anybody explain where that magic "256000" comes from for the formula "256000 / netmtu" described for window_size if corosync.conf?
I'd wish more magic numbers (or the rationale behind) were explained. Example:

"The default is 2500 failures to receive a message." How would you call a communication protocol that needs 2500 retries to deliver a message successfully?

Regards,
Ulrich






More information about the Users mailing list