Bad network performance on new ESX host (part 2)
Some weeks ago I posted an article about some bad network performance when moving VMs to a new ESX host.
Last week I got an e-mail which announced some updates and among it was the one promised to us by the engineer that helped us with our problem.
Description of the patch is exactly what we where experiencing with our VMs:
“When you access remotely using Remote Desktop Protocol (RDP) from a client machine to a Microsoft Windows virtual machine with TCP segmentation offload (TSO) enabled and EnablePMTUDiscovery parameter set to 0, the network connection is slow and sometimes, the connection might not work. This issue is reported on Broadcom bnx2x NICs, such as BCM57710, BCM57711, BCM57711E, etc.” http://kb.vmware.com/kb/1026129
After installing the patch on one of the ESX hosts we tested with a test VM on a patched and an unpatched ESX host and there was a clear difference in performance. The logon on the VM was much faster on the ESX host with the patch.
For us the patch fixed the problem and we are now deploying it on all the ESX hosts.
We got a happy customer again.
- Untangle the mess! Network issues with VMware VMs by Sander Martijn
- Microsoft’s Convenience Update breaks VMware… by Alex Muetstege
- Resolved: VMware vSphere 5.5 Update 1 NFS disconnects by Erik Scholten
- PowerShell Friday: Enabling SSH with PowerCLI by Anne Jan Elsinga
- Veeam Backup & Replication 7.0 patch 4 released by Erik Scholten