Possible reasons:

a. Deletes aren’t being forwarded:

    * this is an option in the Vault / Replication settings - check the admin tool to see if Deletes are being forwarded.

b. Deletes are being forwarded, but haven’t yet been forwarded. This could be because:

    * a large number of objects have recently been written/deleted and all the changes haven’t been forwarded yet

    * the communication pipe is too small for the amount of data being written

    * the replication task is paused (check admin)

    * replication task is set to run in a time window (check admin tool)

b. Small differences can come from the MatrixStore system files - small files that are used for admin purposes - the destination vault could have more of these than the source vault.

c. Someone has directly added a number of objects at the destination cluster, or the vault is used as a destination from more than just one source vault. 

d. The destination cluster could have a number of deleted files in its trash can

e. The destination cluster could have orphaned objects - objects that were written from the source cluster but were never indexed (e.g., replication stopped for some reason before the new object could be indexed).