A Dtrace of the StorageFileWatch process will show an entry similar to the following:
(StorageFileWatch) <45456> EV~E Event ID: 6511 Storage File Watch failed to delete an item that could not be verified. |[Partition: Ptn1] |[Vault ID: 23E1FEE93V8C63548B207D4C128A12D9B1110000EV01] |[Saveset Filename: evroot/29AE3056B4D201458F33EB3F303DC5V4/2017_Oct_21/B01Ab450-aa8a-415g-8c70-169468bcdN04/0_11/00016] |[Transaction ID: 0027FS8ADEB72E43F04FB0A2D43CA121]
V-437-6511
The Enterprise Vault Event Viewer log will also contain this Event.
Issue occurs if the data has not been replicated between HCP nodes when using HCP streamer 1.4.3.3. The HTTP response is 404 Not Found, however the way that the streamer then passes this to Enterprise Vault (EV) is different than in previous versions of the streamer . With 1.4.3.3 this does not occur and it passes information that states file not found. File not found being passed to EV then triggers the logic to roll back the item as it thinks it is invalid. The event log message stating the item cannot be deleted due to the WORM functionality is logged.
If the data has been replicated before the replica is checked then the issue will not occur.
Issue has been resolved by Hitachi using the streamer version 1.4.4.3.
Possible workarounds/minimizing issue occurring includes, increasing thre replication scan of the partition which is set on the properties of the HCP Partition via the Enterprise Vault Admin Console, as per the image below.
Please note the updated streamer change is not retrospective and therefore items already affected will not be resolved and the error can continue to occur. Please contact Enterprise Vault Support for assistance.
Using Hitachi Content Platform Streamer Version 1.4.3.3 the Storage Queue Location is growing in size and Event ID: 6511 Storage File Watch failed to delete an item that could not be verified is logged repeatedly.