Issue with Camunda BPMN

Camunda 7 Topics
1


HARISH

2d

Hello there, Team. We are experiencing the same problem that Sureshv did. Please see the following for more information:

Problem: 1. Camunda BPMN flow changes are reverting to the previous version.

Details:

  1. We have our BPMN files in the resources folder, and as we make changes to BPMN, we will replace the files.
  2. We have a clustered environment in which various Camunda BPMN-based services share a shared database.
  3. Camunda Modeler Version 4.11.1; Camunda Version 7.18.7.

Workaround:
We’ve asked our Dev-Ops Admin to restart the environment. The prior BPMN changes are reversed after 2 or 3 hours.

Please let us know if you need any additional information. Thanks in advance.

Regards,
Harish Anne

2

Hello my friend!

Do you have any logs that could help us find the problem?

You reported that when you restart the environment after a while, new versions appear…, this may be related to the “cache”.

When you upload your updated project, did you check the cluster status to see if all nodes were online and healthy?

William Robert Alves

3

Hello there, William.
We don’t have any logs in the Camunda Cockpit. Whenever we encounter an issue, we request that our Dev-ops team deploy the latest Jenkins build. The workflow is not being deployed via the cockpit.

When we deploy a new Jenkins build, the BPMN flow gets updated to the current version (which we require), but after a while, the BPMN file reverts to an earlier version. So, in order to remedy this, we are requesting that our Dev-ops team produce a new Jenkins build and restart the server.

Regards,
Harish Anne

4

Hello William

Also, we are having the clustered environment. Can you let us know how to delete the Camunda Cache?

Regards,
Harish Anne

5

Hello William,

Hope this message finds you well. We are still facing the issue. Can you help us out in resolving? Please let us know if you need any further details.

Regards,
Harish Anne

6

Hello buddy, how are you?
Sorry for the delay, I ended up not seeing this previous message and due to my busy week, it’s taking me a while to respond.

Ask your DevOps team to check the logs to see if the cluster is not experiencing synchronization issues between cluster nodes, as this would result in changes not being able to propagate to all nodes, causing the change to be rolled back.

To be able to help you in a more precise way, I believe the best way is to track the logs in search of possible errors… this doesn’t seem like Camunda behavior itself… it seems like something related to Infra.

Is your project Spring? If so… do you want to send me your dependencies file (maven or gradle) here?

William Robert Alves

7

If you could provide us with more information about your project that would also be great…

Such as:

  • Spring version;
  • Database used for Camunda;
  • JDK version;

and others…

William Robert Alves

8

Hi @HARISH,

to get closer to the root cause, take a look at the version number of the process definitions.

Does the latest version look like the older version, but has a higher version number in the cockpit?

Then an old version of the BPMN file gets deployed again.

Does the version number of the BPMN diagram decreases?

Then someone undeploys the latest version of the BPMN with the DELETE deployment command: Camunda Platform REST API

Hope this helps, Ingo