I am in the process of configuring and testing a new TruCluster. It
consists of 2 nodes running 4.0f and TruCluster 1.6.
Yesterday afternoon, we tested failover by locating all of the services on
node B and then turning off the power. Everything failed over to node A
fine, with the exception that the apps on one of the services didn;t start
because of an error in the script that gets called to start them.
So I made changes to the script and ran just ran the same test. After
powering off node B, the daemon log on node A reported that it couldn't
ping node B, and then nothing else. Nothing failed over. So I went into
asemgr and looked at the member status. It reported both members UP and
RUNNING, even with one of them powered down!
So is there a bug somewhere? Did I have to reset something after
recovering from the first test?
Greg.
Received on Fri Apr 21 2000 - 11:35:08 NZST