Node properties and aspects are tightly linked to nodes i.e. any update to nodes or properties is always accompanied by a version increment on the node.
Therefore we can, when deleting a node, cascade-delete the aspects and properties.
The FK constraint on the 'node_id' columns needs to get ON DELETE CASCADE and then a thorough check needs to be done on the different databases. Following this, the SQL upgrade scripts will be fairly simple: drop and recreate the FK constraint.
This will alleviate some of the changes to deleteNodesByCommitTime required for ALF-16029. It needs to be investigated thoroughly
[This comment has been reassigned to firstname.lastname@example.org as part of the Alfresco cloud migration project. The author of this comment was resplin] This issue was reported against a version of Alfresco Enterprise Edition that is in Limited Support. Alfresco Software will not implement further enhancements or improvements for products or versions in this support state. If this issue is impacting your deployment of Alfresco One Enterprise Edition, please contact support so that we can assist you. Customers with a support contract can request a bug fix in certain cases.
Current releases of Alfresco contain many improvements that address old problems. We are closing old issues so that we can better prioritize our efforts.
If you verify that the issue still exists in the latest release of Alfresco Community Edition, please reopen the issue. If you have any trouble reopening the issue, then leave a comment or email us at email@example.com so that we can assist.
Thank you for collaborating with us on improving Alfresco.
[This comment has been reassigned to firstname.lastname@example.org as part of the Alfresco cloud migration project. The author of this comment was douglascrp] Are you going to investigate this?
This is happening in a 5.2.f server.
I do not know a way to replicate the issue.
It simply reappears from time to time.
[This comment has been reassigned to email@example.com as part of the Alfresco cloud migration project. The author of this comment was douglascrp] Because is still happens with 5.2.f.
I raised this issue as a potential improvement to introduce cascade deletion of properties associated with nodes. The effect will probably have benefited the cleanup of deleted nodes, in particular.
It is not a bug and therefore cannot occur or be reproduced.
If there is a bug that needs reporting, it might be around the node cleanup job.
Perhaps run that more frequently, change the schedule of that job, or give the DB more resources.
Either way, a new bug should be reported along with all the details of the environment, the data distributions and the errors (frequency and nature).