Showing results for 
Search instead for 
Did you mean: 

Hybris 6.7 Backoffice Orchestra (F4) mode, Reset configs to defaults in cluster nodes.

0 Kudos

Dear experts,

Currently I got an production issue below deployed in on-premise environment with the information below:

Our system has lot of clustering nodes. We create file server which store the data in /hybris/data/medias/*. In cluster nodes, i create mapping in fstab to mount the file server to each node. It's working well. But currently we got an issue with backoffice when loading the BackOfficeCockpitConfiguration loading failed.

There are a lot of backoffice nodes, we suspected someone login to one backoffice node, then using orchestra mode (F4) and reset the config media file (with code = "cockpitng-config") to default. it will modify the content of file in /hybris/data/media/.... which linked to media model. It takes some minutes to finish. In the meanwhile, other nodes read the content of this file and this content is not finished yet, therefore when marshalling the content to Object, it will be failed and other nodes also try to reload the cockpit configuration and also try to update the content as well. As the result all the backoffice nodes are failed to load the content of cockpit configuration.

The classes are:,

I debug and see there is line of code to lock: getMutex().writeLock().lock(); but as i understand this lock only in local of a node. It may lock the media model but it doesn't lock the content of physical file and other nodes can read the incomplete content of file.

Did anyone get the same problem or do you have any idea to troubleshoot and resolve this problem?


Accepted Solutions (0)

Answers (0)