Linux Intelligent Agent - OEM cannot discover node

Hi,
I have a Linux AS 2.1 Server with a Oracle 9.2.0.1 Database running. I have 2 Oracle Management Servers ... one is running on this running on the Linux AS 2.1 Server and another on Solaris 9i.
When start the OEM on the Linux Server and try to discover the Databases on the Linux Server themselves then everything goes fine. When I am trying to discover the databases on the Linux Server from the OEM running in Solaris it does not work.
The Solaris box has connection to the Linux Server and the repository for the Solaris OEM runs on the Linux Machine.
When I go back to the OEM on the Linux Machine then I can discover the Solaris box.
To come to a boddom line here: Whatever I do I am not able to discover the databases on the Linux AS Server besides from the OEM directly running on this box.
Is that a bug in the Linux Intelligent Agent ? If yes is there a Patch to fix that ? Or is there a work around.
Please let me know if you have any idea.
Thanks Marcus

Ok I figured it out .... the problem is/was the entry of the Loopback Interface entry in the etc hosts:
127.0.0.1 dbserver localhost.localdomain localhost
The Intelligent Agent listens to the Loopback Interface now. All incoming requests from and OMS will of course get a timeout because the IP address of the server is 10.1.0.10 .
You have 2 options now:
1. Change the entry in the /etc/hosts below to localhost and add an entry to the /etc/hosts file:
127.0.0.1 localhost localhost.localdomain localhost
10.10.0.1 dbserver
Restart the Intelligent Agent and any OMS in your network should be able to discover the note.
2. In the snmp_rw.ora file you can bind the Intelligent Agent with the following entries to an specific hostname and/or port:
DBSNMP.HOSTNAME = <hostname or ip address>
DBSNMP.CS_BASE_PORT = <port number>
If you start up Agent now you will see in the nmiconf.log file error messages like that:
Warning: No Listener found for SID db01. db01 will be skipped
Warning : No Listener found for SID db02. db02 will be skipped
Warning : No Listener found for SID db03. db03 will be skipped
Warning : No Listener found for SID omsrep. omsrep will be skipped
This is because the Agent looks up in the tnsnames.ora file or whatever you name resolution is and finds the connect descriptors still withe the old name. So all entries in the tnsnames.ora file have to be changed.

Similar Messages

SCOM 2007R2 unix/linux service monitor setup - cannot discover service under /etc/init.d

SCOM 2007R2 unix/linux service monitor setup - cannot discover service under /etc/init.d

Can you give a few more details on what you have tried, and what didn't work for you? Were you stepping through the wizard to monitor a Linux/UNIX service and something went wrong? Or did you successfully create the service monitor in the wizard, but it
didn't trigger if the monitored process terminated?
Michael Kelley, Lead Program Manager, Open Source Technology Center

Intelligent Agents -- OEM

How can I check that Intelligent agent is running on my Oracle machine (For Sun Solaris Intel)
Regds

Nope. Totally different architectures.

8.1.7 Intelligent Agent on Redhat Linux 6.2

Hello,
Can anyone help?
My OEM fails to discover a Redhat Linux node but I can manually configure it (so the node address must be OK). When I check that the agent is running (lsnrctl dbsnmp_agent) it says that the db subagent is already running. If there is something obvious, I've probably missed it because I'm relatively new to both Oracle and Linux/Unix. Any clues, please?

I found this reply on Oracle Metalink and it worked for me (looks like agent on Linux for some reason does not use default ports:
On IA nodes, edit the snmp_rw.ora file and add the following lines :
DBSNMP.ADDRESS = (ADDRESS=(PROTOCOL=tcp)(HOST=<ip-address>)(PORT=1748))
DBSNMP.SPAWNADDRESS = (ADDRESS=(PROTOCOL=tcp)(HOST=<ip-address>)(PORT=
1754))
where <ip-address> is the IP-Address of the node to be discovered.
Note: ports 1748/1754 must be open for the OMS to communicate with the Agent.
The OMS starts communication on ports 1748 and 1754 to get the discovery
information from the Intelligent Agent.
The ports 1748 and 1754 are dedicated for the Intelligent Agent, and should
not be used by any other program.
Ports used by the Intelligent Agent:
++++++++++++++++++++++++++
During the discovery phase, the Communication Daemon OMS
will contact the Intelligent Agent via the ports 1748 and 1754 and retrieve
the discovery information.
Port 1748 is used for the basic communication, while the port 1754 is
used for file transfer. The communication is always initiated from the
Communication Daemon or OMS towards the Intelligent Agent.
The ports are hardcoded in the Intelligent Agent and should not be altered
in any way. The ports are referenced in the snmp_rw.ora file as:
dbsnmp.address=(address=(protocol=tcp)(host=somehostname)(port=1748))
dbsnmp.spawnaddress=(address=(protocol=tcp)(host=somehostname)(port=1754))
The meaning of the dbsnmp.address and dbsnmp.spawnaddress are solely to
determine the resolution path for the Enterprise Manager framework to get
in touch with the Intelligent Agent, and should not be used to specify
other ports.

OEM Repository and Intelligent Agent

How many OEM repository and Intelligent Agent do I need to have to schedule a unique job on three of the 4 databases I have?
Thanks,
Mirabelle

If databases are not discovered, you could have a problem with any of the following files.
listener.ora
oratab ( UNIX/Linux)
Microsoft Registry (Windows)
1. Check the $ORACLE_HOME\network\agent\log\nmiconf.log for warnings or errors.
2. Verify the database sid exists in the listener.ora file. When the Agent parses the listener.ora, it looks for the "sid_list=...." section. If this section is missing, the Agent will not discover the database.
3. Verify the database sid exists in the oratab or Microsoft Registry. (In some cases, the sid is case sensitive.)
4. Stop and Restart the Agent. NOTE: This step is necessary because the Agent will only parse the files during startup.
5. Check the services.ora file and verify the Database services now exists in the file.
6. Refresh the Node through the EM Console.

Dicovering Intelligent Agent on Linux fails with VD-4565

I try to discover an Intelligent Agent 8.1.5 running on a Linux server with an Oracle 8.1.5 DB from an Oracle Enterprise Management Server 2.1. We just upgraded from OEM 2.0.4 to 2.1 and I had no problem with the Linux agent under 2.0.4.
After starting the Agent on the Linux server the SERVICES.ORA and SNMP_RO.ORA files contain the information about the services running on the server. When I start the Discovery Wizard it fails with the error 'Discovery Failed:VD-4565:The agent of the node is already being deleted'.
Then I stoped the agent and all service on the Linux server and did the following. I removed the SID_LIST entry from the LISTENER.ORA file and used the LOCAL_LISTENER init.ora parameter for the database. I know that the agent needs the SID_LIST to discover the database services, but I just wanted to give it a try. After restarting the database, listener and agent the SERVICE.ORA and SNMP_RO.ORA files did not contain any service information as expected. I tried to discover the agent again from the OEM 2.1 console and everything seemed to work fine. No errors this time. The Linux node was added to Nodes folder in the Navigator tree, but for sure no database or listener.
Therefore I put the SID_LIST back into the listener.ora file and restarted the listener and agent. The service information were included in SERVICES.ORA and SNMP_RO.ORA file again. So I tried a Refresh Node from the OEM console and the error VD-4565 showed up again.
Does anyone have any idea what's going on an how I can manage a Linux database from an OEM 2.1 console?
Any help is appreciated.
Thanks
Peter
null

Hello Peter,
I, alas do not have the answer for you. But... your symptom is EXACTLY LIKE MINE!!I was running OEM 2.0.4 and Oracle on a linux box (unfortunately, today I did something wrong and blew the database away!!!, I do not want to tell you how many hours I spend getting the database up, connected and communitcating with Windows NT box running OEM!) I made the switch from 2.0.4 to 2.1, and everything quit. More or less. I fixed my tnsnames.ora file on the NT side. That allowed me to connect. I got the Agent (sortof) running on the linux box and then I could discover the linux host computer name and domain, but it would not find ANY database. Although I could connect from the linux box, the nt box using oem (direct to db), sqlplus, and Access odbc!!! All listeners were there, etc. I did check the snmp*.ora files, and they would not discover anything (linux side) even though the LISTENER was listening, and tnsnames were naming and instances were instancing. My guess is yet another bug.
Sorry for long reply. Now I will RELOAD THE ENTIRE DATABASE AND APP AGAIN!! I so far have been totally unable to run on the liux side, net assist and net easy config, but I could sucessfully bypass those apps by correctly using and configuring initSID.ora, etc and the other *.ora files.
Eric
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Peter C:
I try to discover an Intelligent Agent 8.1.5 running on a Linux server with an Oracle 8.1.5 DB from an Oracle Enterprise Management Server 2.1. We just upgraded from OEM 2.0.4 to 2.1 and I had no problem with the Linux agent under 2.0.4.
After starting the Agent on the Linux server the SERVICES.ORA and SNMP_RO.ORA files contain the information about the services running on the server. When I start the Discovery Wizard it fails with the error 'Discovery Failed:VD-4565:The agent of the node is already being deleted'.
Then I stoped the agent and all service on the Linux server and did the following. I removed the SID_LIST entry from the LISTENER.ORA file and used the LOCAL_LISTENER init.ora parameter for the database. I know that the agent needs the SID_LIST to discover the database services, but I just wanted to give it a try. After restarting the database, listener and agent the SERVICE.ORA and SNMP_RO.ORA files did not contain any service information as expected. I tried to discover the agent again from the OEM 2.1 console and everything seemed to work fine. No errors this time. The Linux node was added to Nodes folder in the Navigator tree, but for sure no database or listener.
Therefore I put the SID_LIST back into the listener.ora file and restarted the listener and agent. The service information were included in SERVICES.ORA and SNMP_RO.ORA file again. So I tried a Refresh Node from the OEM console and the error VD-4565 showed up again.
Does anyone have any idea what's going on an how I can manage a Linux database from an OEM 2.1 console?
Any help is appreciated.
Thanks
Peter<HR></BLOCKQUOTE>
null

Oracle Intelligent Agent cannot be started

I have installed Oracle 9i into RedHat Linux 9. When I start the Oracle Intelligent Agent, I get nothing response, and find the agent cannot be started.
I find the error message in dbsnmp.nohup which is showed below:
/opt/ora9/product/9.2/bin/dbsnmp: error while loading shared libraries: libvppdc.so: cannot open shared object file: No such file or directory
Agent exited at Thu Nov 13 23:20:31 HKT 2003 with return value 127
Agent thrashing. Exiting dbsnmpwd at Thu Nov 13 23:20:31 HKT 2003
I found the library object is existed.
[oracle@localhost lib]$ ll libvpp*
-rw-r--r-- 1 oracle oinstall 552036 Apr 26 2002 libvppdc.so
Can anyone help me to solve the problem?
Thank you.
Regards,
Franklin

Problem of method 1 (for re-install the oem agent / dbsnmp)
When I follow the "work Around Procedure" from
http://www.puschitz.com/InstallingOracle9i.shtml#OracleInstallationErrors
and download the RPMs (glibc-2.3.2-5.i686.rpm glibc-common-2.3.2-5.i386.rpm glibc-devel-2.3.2-5.i386.rpm) for the procedures.
When I run the command (#rpm -Uvh --oldpackage glibc-2.3.2-5.i686.rpm glibc-common-2.3.2-5.i386.rpm glibc-devel-2.3.2-5.i386.rpm),I found the error shown below:
error:glibc-2.3.2-5.i686.rpm:V3 DSA Signature:BAD, key ID db42a60e
error:glibc-2.3.2-5.i686.rpm cannot be installed
Problem of method 2
When I try to unpack glibc-2.3.2-5.i686.rpm and replace libc-2.3.2.so and libc.so.6 by using the following steps
# mkdir /root/tmp
# mv glibc-2.3.2-5.i686.rpm /root/tmp
# cd /root/tmp
# rpm2cpio glibc-2.3.2-5.i686.rpm | cpio -idmv
# cp /lib/libc-2.3.2.so /lib/libc-2.3.2.so.ORIG
# cp lib/libc-2.3.2.so /lib/.
# rm -f /lib/libc.so.6
# ln -s /lib/libc-2.3.2.so /lib/libc.so.6
I find the system is halted, when I run (# cp lib/libc-2.3.2.so /lib/.) When I reboot the system, I find error
init:error while loading shared libraries:libc.so.6:cannot open shared object file:No such file or directory
Kernel panic: Attempted to kill init.
Please help me to fix the problem. Many thanks.
Regards,
Franklin.

Starting the Oracle Intelligent Agent on Linux/390

I have recently installed Oracle Developer's Edition on SuSE Linux/390. When I try to start the Intelligent Agent, I get the following error.
------------------------ Mon Jun 3 11:59:43 EDT 2002 ------------------------
Certain conflicts were resolved during discovery
Please consult /oracle/product/9.0.1/network/log/nmiconf.log for details
Error: Failed to get directory/file info from environment
Failed while initializing Collection Service
Error initializing subsystems
Agent exited at Mon Jun 3 11:59:45 EDT 2002 with return value 55
Could not start agent. Initialization failure
Contents if nmiconf.log:
DBSNMP for Linux 2.4.7-SuSE-SMP: release 9.0.1 - Developer's Release on 03-JUN-02 11:59:43
Discovering services on linux1 configured on host addresses {linux1.dowjones.com 172.26.151.11}
Using oratab file /var/opt/oracle/oratab
Warning : problem getting ops version for /oracle/product/9.0.1 : CIDW
Any assistance would be greatly appreciated.

I don't think the Intelligent Agent is included with the developers release -
check the directory where it's supposed to be, I think you'll find it's not there.
You only need IA to start/stop the database from oem and schedule jobs.
I use OEM to control my instance and run performance packs etc, IA isn't needed for this.

How to config or start Intelligent Agents for OEM

I have install the ORACLE 8.1.7 on redhat 7 successfully. Then I start the "o8ibegin" database instance, start the Oracle Management Server successfully. But, when I start the OEM by using "oemapp console", I could not discover the nodes, it tells me that I maybe do not have Oracle Intelligent Agents. I have chosen to install Oracle Intelligent Agents on install phase, and I have checked that Oracle Intelligent Agents have installed on my system.
I want to know how to start the Oracle Intelligent Agents service. I remember Oracle Intelligent Agents is a service on Windows NT/2000 version. I alos met such problem on Windows NT/2000, and when I start the Oracle Intelligent Agents service throught service applet on control panel, I resolve the problem.

try 'lsnrctl dbsnmp_start' to start
and 'lsnrctl dbsnmp_status' to verify

Cannot discover 9iAS/817db nodes

I have configured two servers each running Windows 2000 Server SP1, Oracle 8.1.7.1.5 EE, and Oracle 9iAS 1.0.2.2.
I followed the Confiuration guide "A Step By Step Guide to install Oracle8i Rel 3 (V8.1.7) and Oracle9iAS Rel1(V1.0.2.2) On Windows NT" from Metalink for this installation.
I want to use OEM to administrer the 8i databases.
I've ensured that the Oracle 8.1.7 Home Agent is running on both machines, but I cannot discover the nodes with OEM.
I can connect to them via SQL*PLUS and DBA STUDIO.
Have checked to make sure tnsnames.ora in all homes include the machines. Both machines reside on the LAN under the same domain as the OEM console and management server.
Have tried manually configuring the nodes.
Thanks in advance!
Dan

I'll answer my own question. The 'Static' IP addresses that the LAN guys gave me were also being served up by DHCP. Strangely, though, 9iAS and the database functioned fine in this situation. Only the OEM discoverer choked on it. Everything works fine now that I've gotten real static IP addresses. OEM Discoverer found the nodes immediately.

Ora8, Intelligent Agent, RH6 & OEM on NT

I've spent most of today trawling through these messages looking
for information on getting the Intelligent Agent up and running
on Linux, RedHat 6.0.
I am running 8.0.5 EE with the glibcpatch installed (hopefully
correctly, as I cannot get the README file of the FTP site).
Everything works, the Agent even starts (2 * dbsnmp processes)
but I cannot discover the services on my box from an NT machine
running the Oracle Enterprise Manager.
By putting tracing on the Agent (level 16) I find that the log
file ends with the following (continually being added to the
file).
nttcnr: exit
nttcon: exit
nsopen: unable to open transport
nsmfr: 1596 bytes at 0x844f2a8
nsmfr: 416 bytes at 0x844f100
nsmal: 416 bytes at 0x844f060
nsmal: 1596 bytes at 0x844f208
nsopen: opening transport...
nttcon: entry
nttcon: toc = 3
nttcnp: entry
ntvlin: entry
ntvlin: exit
nttcnp: Validnode Table IN use; err 0x0
nttcnp: getting sockname
nttcnp: exit
nttcnr: entry
nttcnr: waiting to accept a connection.
ntt2err: entry
ntt2err: exit
nttcnr: exit
nttcon: exit
Any ideas? Any help greatly appreciated. I really don't want to
go back to 5.2 or 5.1 if I can help it.
ps. Can you put the glibcpatch README on the HTTP site and not
the FTP site? I'd love to know what's in it!
Angus
null

Angus (guest) wrote:
: I've spent most of today trawling through these messages
looking
: for information on getting the Intelligent Agent up and
running
: on Linux, RedHat 6.0.
: I am running 8.0.5 EE with the glibcpatch installed (hopefully
: correctly, as I cannot get the README file of the FTP site).
: Everything works, the Agent even starts (2 * dbsnmp processes)
: but I cannot discover the services on my box from an NT
machine
blah-blah
I've reverted to RedHat5.2 and RedHat 5.1's version of TCL.
Everything now runs as expected.
Angus.
null

Oracle Support: SNMP, OEM Console & Intelligent Agent

I have the Intelligent Agent running fine (thanks to the "Does
anyone at Oracle Support read these messages" thread). I can
tnsping the agent and listener from the Win98 OEM Console
machine, I can access the listener when manually discovered.
When I try to submit a job, there is nothing in the Destinations
window. I have SNMP enabled on the 8.0.5.1 server.
Simple question, will it work or not, I really have got better
things to do with my time.
Sick of wasting time.
Paul Smyth
null

Yes it will work. You'de better sign a Technical Support Contract
with Oracle (if you havent) and call them.
They surely can help.
Regards,
Michael
Paul Smyth (guest) wrote:
: I have the Intelligent Agent running fine (thanks to the "Does
: anyone at Oracle Support read these messages" thread). I can
: tnsping the agent and listener from the Win98 OEM Console
: machine, I can access the listener when manually discovered.
: When I try to submit a job, there is nothing in the
Destinations
: window. I have SNMP enabled on the 8.0.5.1 server.
: Simple question, will it work or not, I really have got better
: things to do with my time.
: Sick of wasting time.
: Paul Smyth
null

OEM 2.2 not discovering nodes

Between the space of 2 days, OEM has chosen to not
refresh/drop/discover any nodes. It tries to but stalls out
at "Discovery Wizard, step 2 of 2: Progress" at 96%. It is 96%
for all nodes, remote and local. If allowed to sit long enough,
the OMS service will shutdown. Is it possible to force manual
configuration of the nodes without going thru the Discovery
Wizard first?

Hi
I am not aware whether you can force manual discovery without
going through Discover Nodes process. You can, however, stop the
agent and do discovery. This might (I hope) expedite the
discovery process and lead you to manual discovery quickly.
BTW with a manual discovery you lose major functionality of OEM
sarath

OEM intelligent agent

I'm running 8.0.5 on Red Hat 6.0. The system has been running happily for two months, but I'm now starting to play with the management tools and having a problem.
When I start the OEM intelligent agent via
"lsnctrl dbsnmp_start"
I get the following error message in DBSNMPW.log:
NMS-00315: Message 315 not found; No message file for product=network, facility=
NMS
From OEM Doc I found the following:
NMS-00315: Failed to create tcl package index.
Cause: The tcl package index could not be created.
Action: Make sure all the shared libraries in the $ORACLE_HOME/network/agent/library directory are correct.
I checked the $ORACLE_HOME/network/agent/library and found it was empty.
I then started orainst and checked to see that OEM intelligent agent was installed, it shows as being installed
Has anyone run into this problem? Does anyone have a solution?
Thanks

hello,
I have a problem with the installation of a agent in OEM on linux (Fedora).
this error has occur :
Oracle9i Enterprise Edition Release 9.0.1.5.0 - Production
With the Partitioning option
JServer Release 9.0.1.4.0 - Production
SQL> SQL> SQL> WEBSITE Creation FAILED = EM Website ERR:Host not found
CREATE DEFAULT WEBSITE FAILED : ORA-20998: Host not found
BEGIN
ERROR at line 1:
ORA-20999: CREATE DEFAULT WEBSITE FAILED : ORA-20998: Host not found
ORA-06512: at line 234
Disconnected from Oracle9i Enterprise Edition Release 9.0.1.5.0 - Production
With the Partitioning option
JServer Release 9.0.1.4.0 - ProductionWebsite target creation script failed.
Apr 19, 2004 11:26:04 AM oracle.agent.ConfigureAgent startAgentService
INFO: Starting the agent...
Starting the agent...TZ set to Europe/ParisOracle Enterprise Manager 10g Release 10.1.0.2.0.
Copyright (c) 1996, 2004 Oracle Corporation. All rights reserved.Starting agent .................... failed.
Apr 19, 2004 11:26:38 AM oracle.agent.ConfigureAgent startAgentService
INFO: Started agent Successfully
Started agent Successfully
The Management Agent Configuration Assistant has failed. The following failures were recorded:
Enterprise Manager Website target configuration failed.
Agent Service Failed to start.
Exit status: 1
Configuration assistant "Agent Configuration Assistant" failed
Can you help me please to resolve this problem.
ps:
OEM is succesfully installed in my computer.

OEM 9i Supporting 10g Intelligent Agents

Hello.
We currently use OEM 9.2.0.1 for monitoring and tracking of our 9i database instances. I want to add in our application server when we go live with iAS 9.0.4.1. We're currently running 1.0.2.2, which of course is well behind the times. Will the 10g iAS Intelligent Agent communicate with a 9i EM? If so, are there any limitations, or special considerations I need to take into account? Thanks.

Nope. Totally different architectures.

Linux Intelligent Agent - OEM cannot discover node

Similar Messages

Maybe you are looking for