Details on incompatibilities
If forced switch (hvswitch -f) of the userApplication is issued when RMS on all necessary cluster hosts is not online, there is a possibility of data corruption. To prevent this from happening, the nodes on which RMS is not running will be panicked when hvswitch -f is issued.
Changes
If forced switch (hvswitch -f) of the userApplication is issued, the userApplication is brought to online state regardless of whether RMS on all necessary cluster hosts is online or not, hence there is a possibility of data corruption.
If forced switch (hvswitch -f) of the userApplication is issued when RMS on all necessary cluster hosts is not online, the nodes on which RMS is not running will be panicked to reduce the risk of data corruption. In case of failure of panic, the userApplication will not be brought to online state.
Note
Please refer to "7.2.2.1 Notes on forcibly switching an application" in "PRIMECLUSTER Reliant Monitor Services (RMS) with Wizard Tools Configuration and Administration Guide" before using hvswitch -f.
