ha state transition after partner-down
nhk at snhc-krueger.de
Tue Mar 23 12:46:08 UTC 2010
I'm experimenting with v4.1.1 on Solaris in HA-setup. During this a took
down one ha member numerous times, to make various changes, by using the
"smooth" method setting state=2 in the control object via omshell.
When restarting that ha member, most of the tries it comes up by
transitioning from recover to recover-done to normal directly. But
sometimes I observe that it goes through recover-wait before
recover-done. Can anyone explain what it may depend on whether this happens?
Even more strange, when it goes to recover-wait I would expect that it
stays in that state for exactly mclt seconds, but this is not always
true. I've seen examples where it transitions to recover-done after less
than mctl/2 seconds. Same question, can anyone explain what this depends on?
Thanks for help.
More information about the dhcp-users