REPO-1915 Make sure to have latest jcr node states instead of possibly stale nodes
Since the lock via the LockManager does not involve any jcr locking, it does *never* trigger a cluster sync. This means that we can get the lock for a node that has been processed and finished just by another cluster node that also already freed the lock again. Making sure that after we obtained the lock to invoke a session refresh should result in that #isPendingTrigger returns false if another cluster node already did process the job.
Showing with 1 addition and 0 deletions