error failed upload BLOB during project deploy

Michael Illarietti

During a deployment of a project , the deployment was successful (with positive message at the end on Designer and also in table ELOG) but the BLOB field
of EPROCEDURE table, which should contain the procedure has size 0.

In event viewer I found an error to deploy Oracle system ( impossibility to extend a tablespace) and then failed to upload the BLOB field with the procedure.

The result is that:
- You got the OK after deployment and also by AdministrationTools version appears to be deployed correctly. Also in Table ELOG everything looks good.
- The procedure deployed appartentemente didn't malfunction
- During the retrieve by the designer is unable to retrieve that version, as it gives a fatal error, caused to the lack of field in table EPROCEDURE EPROCEDURE.

Kaman_Paul

Yes I also raised this issue (on MSSQL) when there's a timeout and publication succeeds but storing the procedure as the BLOB in eProcedure fails.  Consequently you get an error when retrieving that version.  This can be problematic as the web runtime is using the latest published procedure but eProcedure would then be out of sync.

Prakash__Kanchan

Yup, we've seen this too ...its random and dangerous because the fail is "silent"

 

What we noticed is that after the deployment says "success" if you wait about a minute or two, there's generally an error message in designer.

 

Thats consistent behavior for us - so at least we have something to watch for to know that deployment really failed.

 

 

Richard Galt

I've experienced this issue on our development server and logged a call with OpenText and was told it was a performance Issue with our SQL Server!

 

When you deploy a process we discovered that the green tick appeared to show it was successful but it would continue to deploy and return a deployment error. This deployment version would then be corrupt and unavailable to retrieve.

 

It appears to be an issue with the amount of time it takes to access the eprocedure table when the deployment occurs. To fix this issue we just cleared down current deployed versions and low and behold it was working fine again.

 

Tony Wheeler

If you see the symptoms described in the original post, what I suspect is happening is that the connection between the deployment service to the engine is timing out.

 

Try adding a liberal 'Timeout' value to the connection string located at around line 14 in "C:\Program Files (x86)\Metastorm\BPM\Deployment\DeploymentService.exe.config" (like below) and restart the deployment service. The deployment service uses this connection to talk to the database and if no value is present, the default value may be too low when network or database is slow or the process to blobify is very large.

 

<connectionStrings>
        <add name="Metastorm" connectionString="Data Source=cpu\sqlexpress;Initial Catalog=Metastorm;User ID=MyEngineAccount;Password=MyPassword;Timeout=600;" providerName="System.Data.SqlClient"></add>
    </connectionStrings>

 

 

 

Michael Illarietti

Hi Tony, I try to add  Timeout=600 and the deployament service doesn't work.

 

My DB is Oracle. Does this parameter work well with Oracle?

 

My connection string is:

 

 

Michael Illarietti

Does anyone know if this parameter exist for oracle database?

 

Timeout=600;