Re: Domain catastrophe?

On 21 Jul 2003 04:13:16 -0700, amani <amanclark_at_yahoo.co.uk> wrote:

> hello,
>
> I'm currently having problems trying to decide whether i have forced a
> database into meltdown or there is merely a startling coincidence. I
> wouldn't describe myself as knowledgable in the ways of oracle as i'm
> more a server monkey to be honest.
>
> I installed a Database on a Windows 2000 Active Directory server and
> gave it the a .world domain.

The installation on to what I still fondly call a Domain Controller is not a good idea. It means there would be a lot of non-Oracle-related I/O, and that can only slow things down for Oracle itself. However, the domain name issue is a complete red herring. Ordinarily, it's the same as your 'real' domain, obviously. But there's no immutable law of the universe that says it must be.

> Despite the fact i now suspect it should
> have been then domain of the server. The next day, one of "my" Oracle
> 8.0i databases bites the dust with connection problems, something
> about package Standard (I've tried catproc.sql, it's still down) and i
> can't tnsping the instance although it's definately running. This
> database is running on a NT server.

I thought you said earlier that the database was running on a Windows 2000 Domain Controller?

Whatever...

You never tnsping the instance, in any case. Tnsping tells you only whether the listener is contactable, and nothing at all about the state of the instance you want to connect to. Therefore, if tnsping isn't working, it suggests that your listener is down, and that most certainly would give you connection problems! You'll need to restart the listener, and see if you can get it working. You may need to switch on tracing and logging for the listener if it refuses to stay up, because that way, when it dies, you'll have some information explaining why it's died... and then you'll be able to post back with complete error messages!

As for package standard: you're right that re-running catproc.sql (as SYS) should re-create this, and thus fix it. You might try re-running both catalog.sql and catproc.sql (as SYS) to see if you can't fix the problem that way. Otherwise, it's difficult to know what to suggest, because "something about package Standard" isn't much of an error message to go on. Write the exact error message down, and post back here with it.

>
> There are still five other databases ranging from 7.3.3 to 8.1.7
> running fine in the .world domain. Can anyone tell me from this
> outline of the events if it was in fact me or a strange twist of fate,
> i don't believe in cooincidence.
>

The domain name has nothing to do with anything.

> Also i'm about to cut the domain names of the server and client to see
> if that will work. Might this have further implications? (It's all
> backed up)

I'd advise against that. Things are obviously sticky enough without you breaking all the other databases! The domain name will have no impact on whether a database can run, or be connected to. On the other hand, if you've got database links set up, or are using advanced replication, suddenly changing the global database name (which is what you're proposing to do here) would likely break those links at a stroke. Leave the other databases alone.

>
> F.Y.I I've had six databases dumped on me with a few days oracle
> training and i'm struggling as my company refuses to buy an outright
> oracle support contract. Are there any support groups people know of
> that i could join to talk things over?

You're in one. However, to help others give you better 'support' in the future, always remember to post exact error messages, and an Oracle version. Clarity about whether a given database is running on an NT server or a 2000 Active Directory server would also help.

~QM

>
> I feel better already actually. Cheers for any RE:'s.
>
Received on Mon Jul 21 2003 - 15:29:12 CDT