Re: OEM - Node Discovery Failed

Hi Howard ! :)

         We meet again ...thanx for your post. I was finally able to resovle the problem. It was a stale dns record problem. Our windows network's DNS server had stale records of the database host and therefore the name of the machine (database host) was pointing to a old name that no longer exists. Anyhow all i had to do is delete these stale records and the nodes were discoverd fine. Once again many thanx for your reply.

Kind Regards,

Jigar

"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<4027dc1b$0$5224$afc38c87_at_news.optusnet.com.au>...
> "jiggaman" <jiggamanrocks_at_yahoo.com> wrote in message
> news:5374d389.0402090935.3735dafa_at_posting.google.com...
> > Hi,
> >
> > I am trying to use OEM 2.2 to manage a few database that we have.
> > however every time I run the Node Discovery wizard it fails giving
> > the following Error
> > "Discovery Failed : VNI-4044 Cannot contact agent. Node may be down
> > or, the network may be down or slow"
> >
> > In the windows control panel under sevices I see that the
> > Oracle<ORACLE_HOME>Agent services is running. however when i go the
> > console into LSNRCTL and give the dbsnmp_status command it says "The
> > db subagent is not started". Since I get that message i run
> > dbsnmp_start and it just give the the LSNRCTL command prompt. Note
> > that when i run the command dbsnmp_start is doesnt give me any error
> > or success conformation messages. once again i check the status of the
> > dbsnmp and it says the "the db subagent is not started". I also tried
> > deleting all the files int eh <orahome>/network/agent directory and
> > then started the service using windows control panel . I get no errors
> > however when i try to discover the node it keeps failing.
> >
> > Since this is a initial testing scenario I have two database instances
> > running on the samebox and one of them contains the repository. I
> > dunno if this is something that is creating the problem . do i need a
> > seperate instance just for the repository? or is it something else
> > that i need to fix.
> >
> > Your help will be apprieciated
> >
> > TIA
> >
> > Jigga
>
> No, you are supposed to have two separate instances and databases, and put
> the repository in one (usually called OEMREP) and use that to manage the
> other (your "real" database). In a production setting, one would expect
> OEMREP to be running on an entirely seaparate machine, but it's not
> compulsory to do so, and I run two-on-one-node all the time. So you're doing
> OK there.
>
> A number of things spring to mind regarding your problem, but all of them
> revolve around networking configuration. Have you changed your machine's
> name at all? Installing Oracle when the machine has one name, and then
> changing it, can produce odd effects (such as the Management Server refusing
> to start, which isn't your problem, but sounds similar). Does your
> listener.ora have a static declaration of a SID_LIST? Can you connect to
> your instances via SQL*Plus using a tnsnames alias? Is there an
> ORACLE_HOME\network\agent\log file we can see the contents of? Are you doing
> odd things like terminal services?
>
> Not sure you should have deleted *all* the files in O_H\network\agent: just
> the services.ora and the .q files would have been sufficient. If you've done
> more damage than that, you could be in trouble.
>
> Regards
> HJR
Received on Tue Feb 10 2004 - 07:28:16 CST