Re: listener intermittently down

Ujang, you are seeing clear symptoms. What alex is describing, is investigating the symptoms.

A clear symptom is the listener appearing to be down. One investigation you did is looking at crs_stat -t output (hopefully while the listener was not functioning, otherwise the investigation doesn't make sense), which doesn't appear to show any problems. Well, get one step further: are there any messages in the listener's logfile which indicate it is being brought down or started up?

The next thing you are mentioning: multiple lines with listener -inherit. That would indicate the cluster daemon trying to start the listener, because it thinks the listener is down. (provided you are the only one on the system, and not a colleague trying to start the listener because he found the listener to be not functioning too, just like you)
Evidence of this event will be in the cluster log file.

The next thing you are mentioning: vip problems. VIP problems (probably you mean the vip moving to another node). This also is something which will be in the cluster logfiles. A VIP doesn't move without any notice anywhere. It will be logged.

High CPU utilization is caused by processes doing something. Using basic operating system commands like top it is easy to determine which processes are the cause of this. Again, this gives a lead to investigate further.

Good luck!

frits

On Nov 11, 2007 4:38 AM, Ujang Jaenudin <ujang.jaenudin_at_gmail.com> wrote:
> crs_stat -t shows all are fine.
>
> but when I check with ps -ef | grep lsn
> sometimes listener -inherit comes 2 or 3 lines
> and back again to 1 line output.
>
> is this due to vip problem? or something else?
>
> from the log, i could not find it...
>
> turn on tracing...hm... it so hard due to listener down
> intermittently, even it can be down after 3 days.
>
> btw, it seem that when listener down , the cpu utilization is on high
> percentage of user processes.
>
> regards
> ujang
>
>
> On Nov 11, 2007 2:25 AM, Alex Gorbachev <ag_at_oracloid.com> wrote:
> > What does "crs_stat -t" show?
> >
> > You should check logs
> > $ORACLE_HOME/log/<node>/racg/<listener_resource_name>.log (not in
> > ORA_CRS_HOME).
> > If details there are not enough - you can enable debugging
> > (_USR_ORA_DEBUG in racgvip or USR_ORA_DEBUG on resource attributes).
> >
> > Also, check that you VIP doesn't fail and not moved to another node.
> > You can also enable listener trace if it dies to try to capture more details.
> >
> >
> > On Nov 8, 2007 6:55 AM, Ujang Jaenudin <ujang.jaenudin_at_gmail.com> wrote:
> > > all,
> > >
> > > 2 nodes linux rhel4.4 IA 64
> > > node 1 has 4 processors
> > > node 2 has 8 processors
> > >
> > > oracle RAC 10.2.0.3
> > >
> > > have the problem that intermittently listener down.
> > > could not found some valuable info from crsd.log / cssd.log or even on
> > > lsn.log or vip.log under client crs log.
> > >
> > > my application comes from IIS6 with ODAC (oracle data access
> > > connection???) with connection pooling disable.
> > >
> > > have been set on listerner.ora
> > > SUBSCRIBE_FOR_NODE_DOWN_EVENT_LSNR_HQRAC1=OFF
> > > INBOUND_CONNECT_TIMEOUT_LSNR_HQRAC1 = 120
> > >
> > > on sqlnet.ora
> > > SQLNET.INBOUND_CONNECT_TIMEOUT = 120
> > >
> > > but same problem still raised.
> > >
> > > please help....and thyanks for the share
> > >
> > > --
> > > regards
> > > ujang
> > > jakarta - indonesia
> > > http://ujang-id.blogspot.com
> > > --
> > > http://www.freelists.org/webpage/oracle-l
> > >
> > >
> > >
> >
> >
> >
> > --
> > Alex Gorbachev, Oracle DBA Brewer, The Pythian Group
> > http://www.pythian.com/blogs/author/alex http://www.oracloid.com
> > BAAG party - www.BattleAgainstAnyGuess.com
> >
>
>
>
> --
> regards
> ujang
> jakarta - indonesia
> http://ujang-id.blogspot.com
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
http://www.freelists.org/webpage/oracle-l