VMware
 

Knowledge Base

Search the Knowledge Base:

Products:
Search In:
 

Limitation of Disaster Recovery Failover LUNS for Some iSCSI Arrays and Some Hosts

Details

What is the Limitation?

When using iSCSI arrays with ESX Server hosts, the maximum number of targets per host is 64 as described in the SAN Compatibility Guide for ESX Server 3.5 and ESX Server 3i. However, when using VMware Site Recovery Manager, failover hosts with some configurations have a limit of approximately 23 recovery LUNs. This limitation applies to recovery LUNs only.

What is the Actual Limit?

The actual limit is determined by how many production LUNs (non-disaster recovery targets) are also accessed by the host. The formulas that dictate the specific limit in any given configuration are as follows:
<number of production LUNs on host> + 11*<recovery LUNs on host> <256
<number of production LUNs on host> + <recovery LUNs on host> <64

Note that this limitation applies to any disaster recovery setup you might regularly test; it is not specific only to Site Recovery Manager.

What Server and Array Configurations Are Affected?

Configurations with this limitation are those that meet all of the following criteria:

  • iSCSI arrays that issue a different iSCSI target per LUN
  • iSCSI arrays that create a new target for each iSCSI LUN snapshot
  • Hosts with software iSCSI initiators
If you are not sure whether your array behaves as described above, please contact your iSCSI vendor .

What is the Effect of the Limitation?

Because VMFS supports 2 TB LUNS and most users use LUNs in the 200-500 GB range, it is unlikely you are affected by this limitation. It is more likely if you:
  • Use clusters with many hosts
  • Are providing disaster recovery protection for virtual machines with very modest resource needs that are highly consolidatable
  • Are providing disaster recovery protection for virtual machines with very large disks
  • Use small LUNs
 

Solution

To work around the limitation if you are affected:
  • Split up clusters. Recovery scenarios involving more than 23 LUNs can be implemented as long as the LUNs are split up between VI clusters so that no host has more than 23 recovery LUNs.
  • Increase LUN size. Use bigger LUNs to serve the virtual machines to a cluster. Use Storage VMotion to non-disruptively migrate from smaller to larger LUNs if necessary.
  • Move from software initiators to hardware initiators.

Feedback

Rating: 1 - Lowest 2 3 4 5 - Highest (0 Ratings)   

Did this article help you?
This article resolved my issue.
This article did not resolve my issue.
This article helped but additional information was required to resolve my issue.
What can we do to improve this information? (2000 or fewer characters)
Submit
Rating: 1 - Lowest 2 3 4 5 - Highest (0 Ratings)   
Actions