| RAC on Windows Problem after one server got stuck [message #74961] |
Thu, 13 January 2005 02:52  |
rudi
Messages: 8
Registered: March 2003
|
Junior Member |
|
|
Hello,
yesterday at 10 pm one of our 2 windows 2000 cluster nodes got stuck and didn't answer anymore.
usually the instance on the 2nd node should get all the connects but the database was not accessible.
I couldn't connect via net8 to any of the instances. Locally i could connect only to the instance
on the 2nd node. in the morning i rebootet the 1rst node and everything was ok.
It seems that the first node was the 'master' of the oracle cluster file system. in the alertlog of
the second instance i found an ora-600 when the system tried to get access to the controlfile on the ocfs.
Status from lsnrctl on the second server told status "undefined" for the 1st server.
But this is not what we build this rac up for. the second instance should have been working and
should have got all the connection requests and access to the database or we have again a single point of failure.
did someone make the same experience?
greetings from frankfurt
rudi
|
|
|
|
|
|
| Re: RAC on Windows Problem after one server got stuck [message #74969 is a reply to message #74961] |
Tue, 01 February 2005 05:09  |
Klaus Debus
Messages: 1
Registered: February 2005
|
Junior Member |
|
|
Hello,
you're not alone.
We had the same problem yesterday.
Server one got different values from the SAN
than Server two and crashed. After reboot server one
was OK but Server two was not accessible via listener.
We believe it was a OCFS-failure on Server one.
We use 10g-RAC on 2 Win2003-Servers.
What is your platform exactly?
Hope you are not in production.
Greetings Klaus
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
