Article number
000004342
Affected Versions
All
Source Hypervisor
All
Target Hypervisor
All

Problem with Datastore Space Available, Resulting in Different Alerts

Viewed 557 times

Summary

An administrator may see some alerts regarding datastore space available for Zerto use, yet the datastore may not be full.

Root Cause

The size of the journal datastore must have at least 30GB or 15% free space, from the total datastore space, whichever is smaller.
 
For example, in a 2TB datastore, 15% free space equals 307GB. Zerto wouldn’t consider the datastore full if it has 307GB of free space remaining. It would only alert if it was 30GB of free space, as it’s the smaller figure.
Alternatively, in a 100GB datastore, 15% free space equals 15GB. Therefore, Zerto wouldn’t alert on 30GB free space. It would alert on 15GB.

NOTE:
  1. Separating the replicated volumes and journals would make a difference in a VM replication disk usage, as the datastore usage is aggregated over all volumes replicating to it. For example, if replicating 2 volumes of a single VM, each one 100 GB to a given datastore, and additionally using the same datastore for 50 GB of journal space (doesn't matter for which VMs), we have 250 GB usage on this datastore
  2. When the total remaining capacity is either 30GB or 15% of the total datastore capacity (lowest of the two), replication to this datastore will be stopped and all VPGs using it may see RPO rising. The reason for this mechanism is if Zerto won't halt replication, the VRA itself may be halted when the datastore is full thereby affecting our ability to failover.

Symptoms

If the datastore fails the validation, VPG Creation on that datastore will not be possible.

In the case of existing VPGs and the validation fails, one of the following alerts will be seen in the Zerto GUI, depending on what the datastore is used for:

Cannot write to journal disk {journal} because datastore is almost full.

Cannot write to recovery disk {disk_name} of virtual machine {vm_name} because datastore is almost full.

Cannot write to clone disk {disk_name} of virtual machine {vm_name} because datastore is almost full.

The storage {storage_name} is close to full capacity and further protection of VMs in VPGs using this storage cannot continue until you increase the storage size. The minimum free space required is {n}GB.

Solution

To resolve this issue, one of the following options must be followed:
  1. Update VPGs to move the necessary disks (recovery or journal) from the suspect datastore to a lesser used datastore.
  2. Manually perform storage vMotion on recovery and/or journal disks as required from the suspect datastore to a lesser used datastore.