Bytearray NAME too long in H2


I have encountered a perhaps rare problem.
During the initialization of the flow, I encounter a “…JdbcSQLException: Value too long for column NAME_ VARCHAR(255)…” for the statement INSERT INTO ACT_GE_BYTEARRAY.
In this query, flowable attempts to store full process resource path (.bpmn) and full diagram resource path (.png) in the NAME_ column. Depending on the deployment, these absolute paths can be quite long, I am wondering why this field is defined with the length of only 255. Moreover, the very same values are also inserted into the ACT_RE_PROCDEF table, to fields defined as VARCHAR(4000). So why in PROCDEF and not in BYTEARRAY?

Is there a way to solve this issue without “fear” of having too long name?



Are you using the auto deployment feature?
You can easily change the NAME_ column to store more than 255 characters of course.
Another solution would be to deploy the process definitions with only the filename and not the path (not using the autodeployment feature).

Best regards,


we are using autodeployment, as we are using spring boot module structure. The autodeployment feature is very convenient for us. We would like to keep using it.

We could of course extend the column manually. However would it survive DB schema update, when upgrading Flowable? We would like to avoid any risk of unecessary manual patches.

Would it be maybe possible to update this directly within Flowable repo? Or maybe it’s enough to store filename (without path) during autodeployment?

Best regards,