Re: Oracle parallel server

From: <markp7832_at_my-deja.com>
Date: 1999/12/10
Message-ID: <82plr0$vie$1_at_nnrp1.deja.com>#1/1


In article <384FBAE0.F50AD9C_at_de.abb.com>,   Marcus Korte <marcus.korte_at_de.abb.com> wrote:
> Hello,
>
> has somebody experienced the following problem:
>
> We are using oracle parallel server 8.0.4 on two Digital alphaserver
> 1200 with memory channel and raid and Digital/Compaq UNIX 4.0D.
> The servers are connected via two FDDI interfaces over two hubs to the
> network.
> If one hub fails due to a power failure Oracle can not be reached
> anymore by sqlplus,..
> Only a stop and restart of Oracle and the restauration of the network
> device/hub helps.
>
> Is there a configuration problem with tnsnames.ora, sqlnet.ora?
> Is there a need for configuration of oracle for dual lan use ?
>
> Regards,
> Marcus
>
A few questions
1) Which instance is your tnsnames pointing to, the one behind the failed hub or the one behind the still powered hub? 2) Can you get to the back-end machine(s) where Oracle is still running when you have a hub failure? If not, then you need to make changes to your network connectivity
3) If you can logon to the back-end server can you create a session? If so you have a network related issue. If not then under the assumption that both nodes are still running then you should query the lock manager for its status.
4) With Dec (Compact) how are the two nodes linked, that is does the private interface between the two lock managers use the failed hub?

I hope these questions help you find the problem.

--
Mark D. Powell  -- The only advice that counts is the advice that
 you follow so follow your own advice --


Sent via Deja.com http://www.deja.com/
Before you buy.
Received on Fri Dec 10 1999 - 00:00:00 CET

Original text of this message