TruCluster
cancel
Showing results for 
Search instead for 
Did you mean: 

5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Mohit Bhatnagar
Occasional Visitor

5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Please help

Case 1 - (Using RDG for IPC)
Incase of Node2 reboot/shutdown faiover is happening , but when node2 comes up
whole database hang and the solution is only to reboot node 1 also.
But Failover of existing connection, when node2 goes down is ok.

Case 2 - (Using UDP for IPC)
- Incase of Node2 reboot/shutdown , node2 comes up properly , but failover is
not happening (ora-3113,ora-3114), even for ideal sessions (ora-3114) and
running select (ora-3113) .
- need to set multiblock read count to 8 instead of 16, because of second
instance take 30 mins to come up if 16 is used


Both cases are not acceptable by business.
6 REPLIES
Mohit Bhatnagar
Occasional Visitor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

one thing ...we are not using memory channel ....it is 100M for interconnect .

rdg patch from HP and related patch from Oracle has alredy been applied .
Agnes Coat
Advisor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Which PK ?

>> rdg patch from HP

Did you talk about ECO Name: TCRKIT0020500-V51BB24-20031029

New behaviour ? which changes ?
Mohit Bhatnagar
Occasional Visitor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Aqnes,

Thanks for reply ..

yes HP tru64 's
TCRKIT0020500-V51BB24-20031029
(without it 2nd instance was not comming up and was crashing)

and oracle's
Patch : 3144315
URGENT: STARTUP OF A CRASHED NODE MAKES THE SURVIVING NODE COMPLETELY HANG

regards.
Ralf Puchner
Honored Contributor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

does the cluster failover properly without oracle?
Help() { FirstReadManual(urgently); Go_to_it;; }
Mohit Bhatnagar
Occasional Visitor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Cluster failover ?
Cluster joins successfully ...at os level .
also.
Oracle failover works perfect in case of instance shutdown normal/abort .....but problem is if restart the node .

both oracle instance hangs and the solution is restarting the other node also .

Regards,

Norman_20
Advisor

Re: 5.1b tru64 cluster/ RAC 9.2.0.4.0 one node shutdown

Is clu_check_config ok?
Have you set correct kernel parameter in sysconfigtab like white paper of Oracle-HP(Compaq)?
I've a customer that have similar problem and all, is an incorrect setting of kernel parameter.
By Norman