Showing posts with label SOA Enterprise Manager Console. Show all posts
Showing posts with label SOA Enterprise Manager Console. Show all posts

Monday, 16 September 2013

Recovering failed instances / messages in SOA 11g using Enterprise Manager

Recovering failed instances / messages in SOA 11g using Enterprise Manager

Want to recover failed instances in SOA? Below are the high level steps to recover.

- Login to SOA EM console http://virtualhost.example.com/em

- On the left side tree navigation, go to Farm_soa_domain > SOA > soa-infra (WLS_SOA1)

- Right click on soa-infra (WLS_SOA1) and then select Service Engines > BPEL



- In the BPEL Service Home page, you will see tabs - Dashboard, Statistics, Instances, Faults, Deployed components, Recovery.

- Click on Recovery tab

- Select the following from the drop downs.
Type:  Invoke , Duration: Select appropriate duration you need to check, Message State: All

- Select the relevant conversation ID and click on 'Recover' tab. Repeat this for all the conversation IDs that need to be recovered.

Please note you can recover or cancel selected messages marked as Recoverable. If messages are marked as Exhausted you can select them and click Reset to return them to the auto-recovery queue, click Mark Cancelled so that recovery is never attempted on them, or click Recover to attempt to recover them immediately.


Tuesday, 9 April 2013

How to check Oracle HTTP Server (OHS) component version?


How to check Oracle HTTP Server (OHS) component version?

Below are a couple of ways to check with Oracle HTTP Server (OHS) version used. There may be more ways to check it.

1) Go to $OHS_HOME/bin

Run ./httpd -version

It should return result as below based on the version of OHS


$ ./httpd -version
Server version: Oracle-HTTP-Server/2.2.15 (Unix)
Server built:   Apr  2 2011 23:51:27
Server label:   APACHE_11.1.1.5.0_LINUX.X64_110325.2001

Please note:

Ensure LD_LIBRARY_PATH variable is set and libaprutil-1.so.0 is present in that location.

Else you will get error as below when running httpd -version command.


./httpd: error while loading shared libraries: libaprutil-1.so.0: cannot open shared object file: No such file or directory

2) Check in Enterprise Manager

- Login to Enterprise Manager console /em.
- Select ohs component (eg: ohs1, ohs2 ....)
- Right click on it and select General Information
- You will see the version of OHS
- However this version info may not be reliable as it may show you can incorrect version in EM specially after upgrades if all upgrade steps are not correctly followed.


Even if you have run below commands after OHS upgrade, sometimes incorrect version may show up in EM.

 ./opmnctl unregisterinstance

 ./opmnctl redeploy

 ./opmnctl registerinstance

Friday, 29 March 2013

AIAReadJMSNotificationProcess [1.0] in unknown state in SOA 11g EM Fusion Middleware Console


AIAReadJMSNotificationProcess [1.0] in unknown state in SOA 11g EM Fusion Middleware Console


Are you seeing AIAReadJMSNotificationProcess [1.0] composites in unknown states in your SOA enterprise manager console?


Errors as below:

<Error> <oracle.integration.platform> <SOA-20020> <Deployment of composite "AIAReadJMSNotificationProcess" failed: Error in getting XML input stream: http://soa.localhost.com:port/soa-infra/services/default/AIAErrorTaskAdministrationProcess/AIAErrorTaskAdministrationProcess.wsdl: Connection reset.>
====> CubeEngine load AIAErrorTaskAdministrationProcess/AIAErrorTaskAdministrationProcess took 476 ms


AIAReadJMSNotificationProcess appears in unknown state on SOA cluster nodes


Composite is in Failed state because it may not have got deployed when soa_server has been started. This may be because you have dependencies between your SOA Composites and AIAReadJMSNotificationProcess that has been deployed before AIAErrorTaskAdministrationProcess it depends on. 

For quick fix:

- restart soa_servers but it may happen again because deployment order of SOA Composites cannot be configured

or 

- redeploy AIAReadJMSNotificationProcess composite


For long term fix, follow:

Doc ID 1151973.1 Boot Order Of The Composites Upon Soa Suite Restart


Doc ID 1369101.1 SOA Server With AIA Unable to Look Up WSDL's At Startup

associated with bug: 12979076 - SOA11G MANAGED SERVER WITH AIA FAILS LOOKING UP WSDL AT STARTUP


Wednesday, 6 March 2013

Soa-Infra Log Configuration Screen Appears Blank in SOA 11g Enterprise Manager Console


Soa-Infra Log Configuration Screen Appears Blank in SOA 11g Enterprise Manager Console

In a SOA cluster environment when navigating in Oracle Enterprise Manager Fusion Middleware Control using this path: Farm_soadomain -> SOA -> soa-infra and right click 
Logs -> Log Configuration, the Log Level tab appears blank.

It appears blank in SOA standalone environments too.

In our case, we upgraded our SOA servers from 11.1.1.3.0 to 11.1.1.5.0 recently.


Log configuration Page appears blank

This may be caused because the WLST script upgradeJRF('/DOMAIN_HOME') command on the node or system where the Administration Server is located for each domain you want to update was not run.

Reference: Bug 12658557 - APPLY JRF TEMPLATE FOR SOA_CLUSTER AND ADMINSERVER SHOULD BE ADDED TO SOA EDG

To resolve this, apply the JRF Template to the SOA_Cluster as documented in EDG.

After the domain is created with the Configuration Wizard, you must target a number of resources not included in the WebLogic server installation to the SOA_Cluster.


To target these resources:

1. Log in to Oracle Enterprise Manager Fusion Middleware Control
2. On the navigation tree on the left, expand Farm_<domain_name>, WebLogic Domain, and then <domain_name>, and select SOA_Cluster.
3. Click Apply JRF Template button right below the SOA_Cluster Information banner in the EM Console's right pane, see screenshot below.
4. Wait for the confirmation message to appear on the screen.
This message should confirm that the JRF Template has been successfully applied to the SOA_Cluster cluster.
5. Restart the SOA server





We were facing this issue on our standalone SOA servers which were recently upgraded from 11.1.1.3.0 to 11.1.1.5.0.

Applied this workaround. However, we hit another error:


Unable to invoke applyJRF because the domain $MW_HOME/111/user_projects/domains/soa_domain has not been extended with JRF template.oracle.sysman.emSDK.app.exception.EMSystemException




Friday, 22 February 2013

Unable to update Email Driver Properties via SOA 11g Enterprise Manager console or changing driver.xml?


Unable to update Email Driver Properties via SOA 11g Enterprise Manager console or changing driver.xml? 

You want to configure outbound email notifications and you are trying to update email driver properties in Enterprise manager?

Were you able to update email driver properties? Did the changes get applied successfully after server restart?

If no is the answer, did you trying changing the email driver properties via driver.xml located under soa_server1/applications/usermessagingdriver-email/configuration and restart the server? Did the changes show up?

If no is still the answer and you are still wondering what's wrong, why are the changes not getting saved, then probably you need to continue reading further !

Hmm ... now try to recall if at some point of time did you ever start managed server from Admin server domain? Did you? May be you did ! Not a problem, perfectly normal, happens with many.

Now because a managed server is started from the Admin Domain, copy of the configuration files are created in the Admin Domain as well under:

admindomain/config/fmwconfig/server/soa_server1/applications/*

 Now that the config files are created in the admindomain, starting the configuration files in the managed domain is overridden by the configuratin in the admin domain ! 


manageddomain/config/fmwconfig/server/soa_server1/applications/*

Ohh no, didn't realize it was this big a mistake !


So what next?

1) Stop the server. 

2) Delete/rename the managed server details from the admin domain admindomain/config/fmwconfig/server/soa_server1 to admindomain/config/fmwconfig/server/soa_server1.bkp

3) Start the server

This will make sure that the configuration files are always loaded from the managed domain.

Now give it a try and update the email driver properties and check if it works ! I'm sure it does unless you goofed up with the server restart again !