Cold storage migration of virtual machine data is slow with the log entry: VMB: 1104: ignoring range
book
Article ID: 326236
calendar_today
Updated On:
Products
VMware vSphere ESXi
Issue/Introduction
Symptoms:
Migrating a powered off virtual machine from one datastore to another is significantly slower than other migration methods.
Cold storage migration of a virtual machine disk is slow.
The vmkernel log file may contain entries similar to:
VMB: 1104: ignoring range [0x10 to 0x89] VMB: 1104: ignoring range [0x103 to 0x200] VMB: 1104: ignoring range [0x8a to 0x9f]
Environment
VMware vSphere ESXi 5.1 VMware vSphere ESXi 5.0
Cause
The reduced throughput on some systems is caused by increased latency in Network File Copy (NFC) operations on some systems, which is caused by a memory caching issue.
Note: NFC file copy operations (such as cloning or moving) are generally slower than storage hardware assisted copy operations (such as VAAI).
Resolution
This is a known issue with ESXi 5.1 and 5.0.
To resolve this issue upgrade ESXi 5.0 to 5.0 update 3 and ESXi 5.1 to 5.1 P03.
To work around the memory caching issue, set boot options to ignore the first 2MB of memory.
To set boot options to ignore the first 2MB of memory:
Open an SSH session to the ESXi host.
Open /bootbank/boot.cfg with a text editor.
Add the variables vmbIgnoreStartMPN=0 and vmbIgnoreNumMPNs=512 to the end of the existing kernelopt line.