Package Deployment Error - Failed to Lock/Unlock Server

badrudubradu

Member
Hi everyone.

Recently I got this troublesome error when I try to deploy a package : "Error Code : 34, Failed to lock/unlock server".
The package contain only 1 Business Function. This Business Function works fine at my FAT Client, and the package is built successfully.

After that, I've tried to deploy several time with different package and different object, and found that this error only occurred when I try to deploy a package that contain Business Function or Data Structure (or Processing Options).
All Package that only contain UBE and/or application deployed successfully.

Before I deploy, I make sure that the web login is disabled, no user is logged in, no UBE is processing, and no zombie process.
I've tried to search similar case on the web, but still no luck.
If we can, we try not to restart the server. But if restarting the server is the only way to solve this error, we can schedule a server restart at the weekend, hopefully this weekend.

I attached the log files and package deployment report.

Thank you for your help.
 

Attachments

  • R98825D_XJDE0001_D170323_T063821014.pdf
    67.6 KB · Views: 36
  • SvrPkgBuild.txt
    14.9 KB · Views: 29
Hi everyone.

Recently I got this troublesome error when I try to deploy a package : "Error Code : 34, Failed to lock/unlock server".
The package contain only 1 Business Function. This Business Function works fine at my FAT Client, and the package is built successfully.

After that, I've tried to deploy several time with different package and different object, and found that this error only occurred when I try to deploy a package that contain Business Function or Data Structure (or Processing Options).
All Package that only contain UBE and/or application deployed successfully.

Before I deploy, I make sure that the web login is disabled, no user is logged in, no UBE is processing, and no zombie process.
I've tried to search similar case on the web, but still no luck.
If we can, we try not to restart the server. But if restarting the server is the only way to solve this error, we can schedule a server restart at the weekend, hopefully this weekend.

I attached the log files and package deployment report.

Thank you for your help.

Do you have jde.log from Build machine and package Build Kernel log for this action ?
 
More than likely you have some sort of process on your server "locking" a resource. Usually a Zombie kernel is to blame - but you have already identified no zombies.

Since the solution will likely be to kill an "active" process that is actively locking whatever resource is being used, I would suggest waiting until the server restart at the weekend. That should almost certainly resolve the issue.

I noticed that you're on TR8.98 - that might have something to do with this. Theres been a lot of improvement in how package kernels operate since that tools release.
 
Thank you everyone for your reply.
We decided to re-deploy the package this morning before business hour start.
What we did this morning:
- Disable web login and ensure no user is logged in
- Ensure no UBE process is running.
- Stop web server.
- Stop application server.
- On the task manager of application server, we check and manually kill all remaining JDE process (Maybe, this JDE process is what causing business function deployment to fail. We're not sure, and have no solid proof. If you have similar case please share with us, thank you).
- Restart OS of web server.
- Start application server.
- Before starting web server, we tried to deploy that package again to PY. Thank God, the deploy is success.
- Start the web server.
- User login and transaction run normally.

So for this package, more or less this case is solved. But we still need to bring this business function to PD.
We'll try to assemble a new package for PD, build and deploy normally (without restarting or stopping any servers / services).
I'll post the result on this thread.
 
Hi badru,

When you try to deploy not getting lock.
Please check serverlog folder , inside that particular serverlog where you trying to deploy. You will get pid which causing lock issue. Check that pid on server ; most probably pid belongs to call object kernel. You can kill that pid ensure no user assigned to that call object kernel.

Regards,
Bab
 
Hi all.
OK so I guess this case is solved. I encountered (again) this problem last week, and restarting the server always resolve this problem. Last weekend we restarted the server, and this morning I tried to deploy the package that previously "failed to lock/unlock server". Thank you.
 
Quite an unusual issue. though a service restart will resolve it rather than going for complete server restart. But it maximum again depends on call object kernels and way the kernels behave.
Here are few more observations, even if its an app/UBE only, where we can straight go ahead reagrdless of the jobs processing. This case can arise.
You need to manually copy the UBE file (max 1kb) created in the deployment server to the enterprise server (where the batch jobs run) basically at that server the UBE file has not yet created. once u copy it. then the deployment will go fine.
No need for any service restart.
 
Last edited:
Back
Top