Home » RDBMS Server » Networking and Gateways » WARNING: Subscription for node down event still pending (Oracle 10g 10.2.0.3, Solaris 5.10)
WARNING: Subscription for node down event still pending [message #342904] Mon, 25 August 2008 08:18
deepmachine
Messages: 80
Registered: August 2008
Location: United States
Member
Hi, My system is a 2-node 10g 10.2.0.3 cluster and I have been having issues with serviced instances by a service name in my TNSNAMES file.

Problem 1:
-------------
This is what I used to see by doing a LSNRCTL STATUS:

(truncated results)

Service "prod_x.nipa.org" has 2 instance(s).
Instance "prodx1", status READY, has 2 handler(s) for this service...
Instance "prodx2", status READY, has 1 handler(s) for this service...

Service "prod_y" has 2 instance(s).
Instance "prody1", status READY, has 2 handler(s) for this service...
Instance "prody2", status READY, has 1 handler(s) for this service...

AND NOW, I only see this...

Service "prod_x.nipa.org" has 1 instance(s).
Instance "prodx1", status READY, has 1 handler(s) for this service...


Service "prod_y" has 2 instance(s).
Instance "prody1", status READY, has 2 handler(s) for this service...
Instance "prody2", status READY, has 1 handler(s) for this service...

On the second node, "prod_x.nipa.org" is the same except the instance it services has a differnt name i.e. prodx2.

I can paste the whole summary of LISTENER as is if needed for any further analysis.

Above situation occured after we restarted both nodes. Before restart configuration of files is mentioned below:

- LISTENER file was listening from two host addresses, one of them was a VIP, and second one was a Public IP. ---> I changed the public IP to be also a VIP whereas pointing to second node. For example: HOST = ora1-vip, HOST = ora2-vip

- TNSNAME file only contained VIPs pointing to both nodes, and I added IPs for Management Server as well pointing to both nodes, as we thought that everyone should connect through it except the application. (<-- was this a good approach?)


Actions taken:
- I tried to revert back to the original setting that I had, but its still the same.


Problem 2:
-------------
- Until yesterday we only had a problem of "prod_x.nipa.org" not servicing both instances.

- Today, I noticed that "prod_x_taf.nipa.org" has also disappeared, and previously it was also serving only one isntance, whereas it should have both (this was also after the problem 1)

Could someone give an idea as to what's going on?

Thanks in Advance
Pinky.
Previous Topic: Problem Oracle Listner.log
Next Topic: listener getting hanged
Goto Forum:
  


Current Time: Fri Apr 19 07:17:02 CDT 2024