Hi All,
We are getting no process models enabled error. I need to republish the process models and recreate the JMS Queues every day. The host is set to restart every day so the queue are getting vanished, i need republish the process models to re-create the queue. Even setting Persistent property not helping us.
Please suggest how to make permanent fix or retain queues
Shafi
Hi Mohamad,
can you provide some more details, please ?
Which version of wM?
Which type of messaging (UM vs. Broker)?
List of Fixes for the above if any are applied?
Sounds strange to me when the client queues on the messaging are not persisted over restarts.
This should be checked with Support team. As long as the support system is not available due to maintenance you will have to send an email with the description of the issue. This email address can be found on www.softwareag.com.
Regards,
Holger
Hi Holger,
Please find details
wM version - 10.3.0.0
messaging type - JMS over UM
fixes -
IS_10.3_Core_Fix7
Trading Networks Server 10.3 Fix 4
Trading Networks Server 10.3 Fix 5
IS_10.3_SPM_Fix3
IS_10.3_SPM_Fix4
PRT_10.3_Fix4
TE_10.3_TE_Fix2
TE_10.3_TE_Fix3
Even the queues made persistent still they go vanish every day after VM restart. Please let me know if am missing any configuration.
Below are the messaging JMS setting
| Enable CSQ | Yes |
|---|---|
| Maximum CSQ Size (messages) | [UNLIMITED] |
| Drain CSQ in Order | Yes |
| Publish Wait Time while Reconnecting (milliseconds) | 30000 |
| Enable Follow the Master for Producers | Yes |
| Enable Request-Reply Channel and Listener | Yes |
| — | — |
| Enable Follow the Master for Consumers | Yes |
| Logging Type | SERVER LOG |
| — | — |
| Enable Producer Message ID Tracking | No |
| Producer Message ID Tracking: Include Channels (semicolon delimited) | [ALL] |
| Enable Consumer Message ID Tracking | No |
| Consumer Message ID Tracking: Include Triggers (semicolon delimited) | [ALL] |
Connection Alias settings
| Create Connection Using | JNDI LOOKUP |
|---|---|
| JNDI Provider Alias | DEFAULT_IS_JNDI_PROVIDER |
| Connection Factory Lookup Name | test_um |
| Create Administered Objects On Demand (Universal Messaging) | No |
| Enable Follow the Master (Universal Messaging) | Yes |
| Monitor webMethods Connection Factory | No |
| Class Loader | INTEGRATION_SERVER |
| — | — |
| Maximum CSQ Size | [UNLIMITED] |
| Drain CSQ in Order | Yes |
| Create Temporary Queue | No |
| Enable Request-Reply Listener for Temporary Queue | No |
| Enable Destination Management with Designer (Broker and Universal Messaging) | No |
| Create New Connection per Trigger | No |
Hi,
is the VM configured to persist the last state before shutting down/restarting or is it reverting to the state available on startup?
Which VM technology are you using?
Regards,
Holger
Hi Shafi,
When the channels and queues are created you should see them on the file system of your Universal Messaging installation - if you navigate to the data directory of your Universal Messaging server (usually something like <UM_Installation_Directory>/UniversalMessaging/server//data) you should folder structure and files corresponding to the Topics and Queues that you have created.
As Holger hinted, can you confirm whether after the restart the data directory is the same as before the restart.
The first thing that comes to mind is that the VM may be running in non-persistent mode - it’s worth checking it out.
Thanks,
Stefan
Hi Holger and Stefan,
We are into Azure VMs. i see the queue and topics after every VM restart in data directory of UM. But i should check with team whether VM persistence enabled at VM level or not.
Do you think any issue in fixes or connection settings.
Shafi
Hi Shafi,
This is odd. When you say the queues are gone after restart, where do you check this? If you are able to connect to the UM with an Enterprise Manager, do you a see the Topics and Queues in the navigation tree on the left?
If they are visible in Enterprise Manager prior to the restart, and missing after that (and yet the files on disk are present), we should check the internal state of the server. The Topics and Queues definitions are actually stores in a file called channels.nst under the data/RealmSpecific folder of the UM server. It’s a binary file so you won’t be able to make much of it, but if you could upload it here (a copy taken before and another one taken after the machines is restarted) I’d be happy to check it out and confirm if it looks alright. Another thing that may be worth checking is a realm export that you can generate using the Enterprise Manager (again before and after the restart). If the files look ok, and we are indeed not seeing the stores in Enterprise Manager, we should check the nirvana.log file of UM for anything odd during bootstrap of the server.
I doubt it’s a connection settings issue, or anything to do with fixes, but who knows - maybe the Topics and Queues are actually there, but not visible to the process runtime…
Thanks,
Stefan
Hi Stefan,
Thanks for sharing. I was able to replicate same in our lower environments and for quick solution i did uninstall the UM, reinstalled it by commands using Enterprise installer. And added the UM fix mentioned by SAG again on it. Followed by process model deployment. Now since couple days, the application behaving good and no issues found with queue behavior. Let me follow the same with other evns as well.
Stay safe and Thanking you
Shafi