WebLogic Console - No application files exist when redeploying application

Question

When stopping/starting a deployment via the WebLogic Admin console, sometimes we get the following error:

Unable to access application source information in '/opt/product/oracle/local/managedservers/mydomain/servers/serverA/stage/apputil/apputil.war' 
for application 'apputil’. 
The specific error is: [Deployer:149158] No application files 
exist at '/opt/product/oracle/local/managedservers/mydomain/servers/serverA/stage/apputil/apputil.war'

Yet, if I stop and start the managed server, the deployment appears to come back.

My question is, WHY do these war files disappear from the managed server seemingly randomly, while the server is running? This typically happens after we haven't touched a deployment for some time (6 months). Other war files for other deployments are there. It does not affect the running of the app, until we try to stop/start it.

This is what the filesystem looks like before and after.

[oracle@serverA stage]$ pwd;ls -alstr
/opt/product/oracle/local/managedservers/mydomain/servers/serverA/stage
total 20
4 drwxr-x--- 8 oracle dba 4096 Mar 19  2014 ..
4 drwxr----- 3 oracle dba 4096 Mar 19  2014 app-crypto-util
4 drwxr----- 2 oracle dba 4096 Mar 19  2014 appadmin
4 drwxr----- 2 oracle dba 4096 Mar 19  2014 appsm
4 drwxr----- 5 oracle dba 4096 May  1 15:29 .

[oracle@serverA stage]$ ls
appadmin  app-crypto-util  appsm

Restart managed server here...

[oracle@serverA stage]$ ls
appdmin  app-crypto-util  appsm  apputil

[oracle@serverA stage]$ ls -alstr
total 24
4 drwxr-x--- 8 oracle dba 4096 Mar 19  2014 ..
4 drwxr----- 3 oracle dba 4096 Mar 19  2014 app-crypto-util
4 drwxr----- 2 oracle dba 4096 Mar 19  2014 appadmin
4 drwxr----- 2 oracle dba 4096 Mar 19  2014 appsm
4 drwxr----- 2 oracle dba 4096 Jun 25 14:35 apputil
4 drwxr----- 6 oracle dba 4096 Jun 25 14:35 .

[oracle@serverA stage]$ ls -alstr apputil/apputil.war 
28660 -rw-r----- 1 oracle dba 29347298 Jun 25 14:35 apputil/apputil.war

Answer 1

This may happen when AdminServer & Managed Server are in different machines, or the war is being sent from a different machine to AS.

use arguments: -remote -upload

i.e.:

java weblogic.Deployer -adminurl t3://200.10.10.125:7001 -verbose -username weblogic -password welcome1 -deploy -targets WLCluster -name sample -remote -source sample.war -upload

Answer 2

Referring to the stage directory modification date, it appears this dir is created/edited when an event occur, maybe by a script, and if it is the case, the problem should come from that script when copying apps wars.

So in my POV when stopping a deployment, maybe weblogic stop the apputil managed server, wich delete the war from th stage dir, and when starting a deployment after stopping it, it does not start the specified managed server before, and try to redeploy all apps wich cause the exception.