In a backup cluster "skipped" nodes should be visible for all cluster member

If a job chain node is skipped by processing a command like

<job_chain_node.modify action="next_state" job_chain="/job_chain1" state="200"/>

a record in table SCHEDULER_JOB_CHAIN_NODES takes place. It contains [host]:[port] from the currently active cluster member.

If another member of the backup cluster takes over the processing it can not "see" those database entries because CLUSTER_MEMBER_ID is a part of database selection to load the job chain configuration.

All "skipped" notes should be visible for all cluster member.