Oracle's "opmnctl: starting opmn and all managed processes..." won't start

restart
infra server
application server

<<We had a successfully running Oracle application & database servers running together >>
sorry ignored that both are at same server

-->> restart this server

-->> Did you try to connect as admin (ias_admin) and then start all

-->> is everything (i.e control files, logfiles etc) in /u01/app/oracle

How you upgade the infrastucture Database?
Upgrade comprises 2 steps:
1. Upgrading of the software
2. Upgrading of the Catalog (catupgrd.sql script).
Did you run the second step?

Why I ask you? Because you have also to recover the backuped
state of the Database. Please check if the .dbf files are located
in /u01... subdirectory.

Please provide the error message you get.


making a recovery from cold backup

It appears that the OC4J process hangs with no response and the OID process shows "406 Unacceptable request"
I ran each process with "opmnctl verbose startproc ias-component=" and my results are:
$ ./opmnctl verbose startproc ias-component=dcm-daemon
Response: 0 of 1 processes started.     (See TYPICAL verbose result:)
./opmnctl verbose startproc ias-component=DSA
(See TYPICAL verbose result:)
./opmnctl verbose startproc ias-component=LogLoader
(See TYPICAL verbose result:)
************************************************************************************
TYPICAL verbose result:
________________________________________________________
opmnctl: starting opmn managed processes...
HTTP/1.1 200 OK
Content-Length: 632
Content-Type: text/html
Response: 1 of 1 processes started.

<?xml version='1.0' encoding='ISO-8859-1'?>
<response>
<opmn id="domainname:6200" http-status="200" http-response="1 of 1 processes started.">
  <ias-instance id="cisinf.domainname">
    <ias-component id="DSA">
      <process-type id="DSA">
        <process-set id="DSA">
          <process id="12976674" pid="31108" status="Alive" index="1" log="/u01/app/oracle/product/Ora10gAS_Inf/opmn/logs/DSA~1" operation="request" result="success">
          </process>
        </process-set>
      </process-type>
    </ias-component>
  </ias-instance>
</opmn>
</response>
*******************************************************************************************************
[oracle]$ ./opmnctl verbose startproc ias-component=OC4J
HTTP/1.1 200 OK
Content-Length: 0
Content-Type: text/html
Response: Ping succeeded.

opmnctl: starting opmn managed processes...

(This just hangs with no response.)

$ ./opmnctl verbose startproc ias-component=OID
HTTP/1.1 200 OK
Content-Length: 0
Content-Type: text/html
Response: Ping succeeded.

opmnctl: starting opmn managed processes...
HTTP/1.1 406 Unacceptable request
Content-Length: 186
Content-Type: text/html
Response: no processes matched this request

<?xml version='1.0' encoding='ISO-8859-1'?>
<response>
<opmn id="domainname:6200" http-status="406" http-response="no processes matched this request">
</opmn>
</response>

$ ./opmnctl status
</response> Processes in Instance: cisinf.domainname
-------------------------------------------------------------------------------------------------------
ias-component      process-type           pid             status
--------------------------------------------------------------------------------------------------------
DSA                      DSA                      31108             Alive
LogLoader               logloaderd               31031             Alive
dcm-daemon               dcm-daemon               N/A             Down
OC4J                     OC4J_SECURITY        0             NONE
HTTP_Server         HTTP_Server            30855            Alive
OID                      OID                           0             NONE

What am I missing?

Catupgrd.sql will not work on this version (10.1.0.4.2) that I have gone back to with my backup recovery.
Yes, all the control files and logfiles and .dbf files etc are under the /u01 directory structure.
Is there any information that is kept or used from system files that would reside outside the database directory structure?

I will recommend to reinstall the Oracle Application server instead
to recover it. There is no known mechanism and experience to
restore Application Server installations.

Be aware also that Oracle 10g R1 is very bugy software.
If your Database (not the Infrastructure DB, I mean the main DB with the application Data)
works fine and is on another box I will suggest to reinstall the Application server.
If the DB is on the same box you have to backup very carefully the DB
before doing this,

OracleiAS 10g includes 2 components: The infrastructure server and the middle ware server.
All Oracle application softwares(like Forms,Discoverer) are packed in the middle ware server.
The middle ware server can work on if its dependent  infrastructure server is available.

If both components are installed in the same box, especially on Windows plateforms,
you have to start the  infrastructure server first, then the middle ware server.
Unfortunately, both components are configured to be as Windows services,
but Windows does not garentee to start the services in their proper order.

There are methods to achieve what AlbertYou says in Windows environment. The sympoms are that the process of Infrastructure DB is alphabetically after the Middtier process. If you ask me I will post the methods.

But the asker speaks about upgrades, not reinstallation. I do not think that upgrade will change the service names.

We have had the Infrastructure and Mid tier working properly prior to attempting a patch.  The system will now not start.  It hangs trying to start the OC4J process. ( As best I can tell).  
We are now attempting to recover from our backup prior to applying the patch.  Have only, so far, recovered the Oracle directories.  I am believing that the application server is looking for something else outside the Oracle structure.  Does anyone know what the pieces and parts are that the Infrastructure and the Mid tier looks at while starting up and in what order it does this?