how to tell whether failover Oracle Net address is in use
- From: John Clarke <jclarke@xxxxxxxxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx
- Date: Thu, 11 Aug 2005 14:03:59 -0400
I've got an environment that uses Data Guard over a slow network, and have
implemented SSH port forwarding w/ compression, with very good success, to
speed up archivelog transmission times. On the production node, my
log_archive_dest_2 goes to an Oracle Net address PROD-dg-ssh.
In the production tnsnames.ora file, the PROD-dg-ssh entry looks like this
(domain names changed to protect the innocent):
PROD-dg-ssh = (DESCRIPTION=
(ADDRESS=(PROTOCOL=tcp)(HOST=localhost)(PORT=9000))
(ADDRESS=(PROTOCOL=tcp)(HOST=x22.domain.com)(PORT=1521))
(CONNECT_DATA=(SERVICE_NAME=PROD) (SERVER=DEDICATED))
)
I've got an SSH tunnel established locally and on x22.domain.com that basically
does the FAL server/SSH stuff.
My question is this - from time to time, I notice things slow-down while
archiving, as evident by ARCH spending may too much time waiting on "ARCH wait
on SENDREQ" events. When troubleshooting, I'll do tnspings to PROD-dg-ssh
service name is get very very slow responses. If I kill and re-establish my
SSH tunnel, tnsping times return to the realm of normal. My suspicion all
along is that something wrong has happened on the SSH side of things and the
Oracle Net communciations had begun failing over to the second address.
Is there any way, short of turning on tracing, that I can tell which address
things are actually resolving to? I'd like to be able to monitor these
conditions and restart my SSH tunnel if I see a condition arise.
( I saw "short of tracing" b/c I really don't want to deal w/ tons of Oracle
Net trace files in a production environment).
Of course, I obviously want to figure out what makes my ssh tunnel "break" and
resolve the root cause, but that's a different topic ...
Thanks,
John
--
//www.freelists.org/webpage/oracle-l
Other related posts:
- » how to tell whether failover Oracle Net address is in use
