"srvctl start instance -d some_db -i some_db2" executes too long

Question

Hi,
We got [2 node Oracle 10g RAC (10.2.0.4) on hpux 11.31 v3].
While starting second instance of some databases through "srvctl start instance -d some_db -i some_db2" it takes almost 5-10 minutes. While this "instance startup" process, some clients experience "disconnect"/"service loss".
TNSNAMES.ORA for a database looks like:
some_db =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.1.136)(PORT = 1525))
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.1.137)(PORT = 1525))
(LOAD_BALANCE=yes)
(CONNECT_DATA =
(server=dedicated)(service_name=some_db)
(failover_mode=(type=select)(method=basic)(retries=180)(delay=5))
The questions are:
- what could be wrong and where?
- what additional info would be helpful that I can provide?
Best Regards,

Answer 1

Hi,
The space on the disk volumes sould not be the cause for such a message, probably.
Real cause seems the bug mentioned in posts (by Ivan Kartik):
- Not able to start the RAC database service
- http://ivan.kartik.sk/static/42/
Suspicion also on datetime matters in the second node, as log files in that node show different timestamp than that of the OS. (alert log timestamp difference in a RAC node
_Digging further yet.