The following alarm could appear if adequate space is not available in the replication log volume.
Details of a Replication Log Volume
Each protected domain uses a single replication log for all the VMs belonging to it.
The replication log volume consists of multiple sections (illustrated below):
Replication Log Section
Keeps a record of data being replicated to the object store.
I/O coming into the DRVA is written to the Object store.
If there is a network bottleneck the I/O will get written to the replication log.
(It acts as a temporary placeholder.)
Metadata Section
Contains metadata of the VMs including garbage collection (DB).
Free Space Section
When Metadata becomes full, if there is available Free Space in the replication log volume, the management server will automatically re-assign some of the Free Space to Metadata.
The “log space low” message can appear when the Replication Log Section is full or > 90% for a certain period of time. This message will stop appearing once the space is freed up.
Best Practice
Replication log volumes should be sized to 64GB per protected domain (regardless of the number of VMs being protected).
The number of vDisks should not exceed 500 per DRVA.