2013-01-17 11:51 AM
I have two major problems that I need your help to solve on a customer stretched cluster .
1. After implementing the MCTB on the oncommand (5.0.2) all went well, MCTB monitored the site and initiate cf forcetakeover -d when we pulled off the power from site A.
I came back to the customer site several days afterwards, MCTB service was started but I didn't see any activity on the mctb log. Only after several restarts the MCTB came back to live?
2. I configured MCTB on the customer seconed cluster (separated environments) same configurations as i did on the firs cluster (identical).
The problem there is that after pulling off the power from one of the sites, cf forcetakeover -d is NOT initiated. quick review in the log shoed that MCTB recognized that IC is down and that the aggregates are degraded, but the site is still reachable for some reason, beats me I don't know why....any suggestions?
2013-01-18 09:16 AM
Sorry for the delay in responding.
Regarding #1, I have never seen that behavior. Do you have the log files? Although you said there was no activity, it would be nice to see what was going on before the logging stopped, and what was logged during your restart attempts.
Regarding #2, as long as MCTB can connect to the IP address in the config file, it will not issue a CFOD. The logging should tell which port was successfully connected to. Perhaps there was a typo in the IP address in the configuration file, and the address was actually accessible even though the controller was not. Perhaps you are using the same address for both controllers, or an address of a controller in the first MC (I often make this type of mistake when doing copy/paste, or copying config files :-) ).