RAC Cluster - Hyperfabric : Oracle shutdown error

[Jean Mesquida_1]

Hi all,

I have one RAC (Oracle 9.2 / HP-UX 11.0 64b) cluster that behaves normally except when I want to stop one node : the node halts successfully, but, when I want to stop Oracle instance on the other node, I get the following error message (Hyperfabric linking both nodes is node redundant) :

ERROR
ORA-27504: Message 27504 not found...
ORA-27300: OS System dependant operation : HMP is not UP failed with status: 4
ORA-27301: OS failure message: Interrupted system call
ORA-27302: failure occurred at: skcsicini

As the unique Hyperfabric link is down (the other node has been stopped...), HMP protocol, used for dialog between SGAs on the two nodes, is also down and I agree that Hyperfabric link cannot be activated.

I have installed last Hyperfabric patch (PHNE_26551) which did not help.

My question is : why does the remaining instance refuse to stop in this case ? Is there a way to force it down ?

Thanks a lot for your help.
J

[Jean Mesquida_1]

I found the solution : just relink oracle binary after installing the patch, as HMP libraries are modified by PHNE_26551.

After doing that, everything works fine !