I have just released MCTB 2.4.0. This release uses a new algorithm which defers the TCP connection test, happening only when all other criteria for CFOD have already been met. This will speed cycle time and reduce the number of TCP sockets used during each polling cycle (which Windows systems can have difficulty with). Also included is an improved method for avoiding false CFODs during normal giveback operations. Improvements to logging and other areas have also been made. This release has been thoroughly tested in the lab. As always, please post any questions or issues to this forum.
You can upgrade in place from any previous release of MCTB
Stop the MCTB Windows service or Linux daemon prior to the upgrade (no need to stop DFM)
Install the release according to the instructions in the Admin Guide. There is no need to delete the existing plugin, and you can upgrade from any previous release of MCTB. The only difference from the instructions in the Admin Guide is that you will be asked to confirm an "Upgrade" instead of an "Install".
On Windows only, run the "uninstall" and then the "install" script commands after the upgrade. This can help avoid service start issues.
On Windows only, confirm that the service startup type is set to "Automatic (Delayed Start)"
Dear Charlotte, very thanks. I upgrade the tiebreaker form 2.2.0 to 2.4.0. But I have Another question. In test environment, use Tiebreaker 2.2.0 the “POLLINGINTERVAL set 5s, CONNECTTIMEOUT set 5s CONNECTRETRIES set 1” When happened CFOD, the test result: ping down site filer IP address lost：50packet, service hung:63s Then update tiebreaker version to 2.4.0,use “POLLINGINTERVAL set 5s, CONNECTTIMEOUT set 5s CONNECTRETRIES set 1” When happened CFOD, the test result: ping down site filer IP address lost：91packet, service hung:107s I understand, the result is very approach. But this result is very big different. This result is normal?