Production servers: RAC Oracle 10g 10.2.0.3, AIX 5.3 ML3, 2 Nodes IBM
pSeries 550.

Symptoms:

CRS Status "Can not communicate with CRS" After the reboot the first
Node, the Node was
not able to rejoin the cluster RAC reported CRS communication error,
after that the error
transferred to the next Node, so the nodes were not able to
interconnect to each other

#CRS-Stat
CRS - 0184: can not communicate with the CRS daemon.

In fact the CRS is configured to start after reboot, and it ran as
root.
We tried many times to ran healthy check on the 2xNodes,
by doing "$CRS-Home/bin/crsctl check crs as away to determine
if the daemons are running, combined with monitoring "ps" for the
processes,
with no success!

Any input would be really appreciated

Thanks

madunix wrote:
Run the cluster verify tool.
Have you verified the connection by pinging across it?
--
Daniel A. Morgan
Oracle Ace Director & Instructor
University of Washington
damorgan@x.washington.edu (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org

DA Morgan wrote:
Daniel,

Have you been successful in using multiple cross-over interfaces for
cluster interconnect? On Solaris, I have used IPMP (active/passive),
but would like this to be active/active.

Michael.

Michael Austin wrote:
What is required is a couple of GigE NIC cards with a cat 5E or cat 6
piece of copper connecting them through a gig switch ... nothing more.
Get the basics working first.

Then go for multipathing as a separate step.

I've personally not had experience with Solaris's implementation
but I know people that have. Check here:
http://www.oracleracsig.org
In fact there is an online meeting Wednesday so join us. Erik
Peterson will be there and he's the head of the RAC Pack at Oracle.
--
Daniel A. Morgan
Oracle Ace Director & Instructor
University of Washington
damorgan@x.washington.edu (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org