VMware Cloud Community
lywwing
Enthusiast
Enthusiast

Does HA can detect HBA failure and failover

hi,

ESXi 4.1 with VMware HA, vMotion, but no VMware FT.

I have two fibers in one VM host, I disconnect all fibers, it don't trigger fail-over, and no error report in vCenter, but VM guest is dead.

Does it normal? or I need extra option e.g. VMware FT?

Many thanks.

Reply
0 Kudos
5 Replies
AndreTheGiant
Immortal
Immortal

It's correct.

VMware HA does not handle storage failure.

You must have a storage designed for HA and multipath to handle this cases.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
Reply
0 Kudos
lywwing
Enthusiast
Enthusiast

hi,

What is 'storage designed'? in my system, there have no fiber switch, the IBM DS3215 just have two controller connecting to two VM host servers.

VMware HA don't have multipath? why my two VM host server can access the storage at the same time? and both VM host server can start VM guest (active-active) in the same storage, I think the VMware cluster already have multipath to tackle about this.

Many thanks.

Reply
0 Kudos
AndreTheGiant
Immortal
Immortal

I've written that for storage part you must design for HA and check that works:

dual HBA, dual controllers, ...

Multipath is NOT related to VMware HA... it's need to handle SAN path failure.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
Reply
0 Kudos
nnmansoori
Enthusiast
Enthusiast

The overall availability of the virtual machines and the entire vSphere cluster
are both dependent on the shared storage infrastructure. Designing in high availability into
this infrastructure element is paramount. If the storage is not available, VMware HA will
not be able to recover, and the aggregate community of VMs on the entire cluster can be
affected.

for Storage redundancy, you should use multipathing.

Vmware ESX/ESXi 4.x contains their own native multipathing software that is built into the kernel. This failover software, called Native Multipathing Plugin (NMP), has three policies:

  • Fixed
  • Round Robin policy
  • Most Recently Used (MRU)      policy

When using ESX/ESXi 4.1 or later, there are 2 types of Fixe policies settings. Of the 2 options, for CLARiiON storage systems, the VMW_SATP_FIXED_AP is used of the VMW_SATP_FIXED setting.

for more information, you can visit the following link:

http://blog.mythincloud.com/2011/07/vmware-native-multipathing-and-failover.html

Nastaran M

Reply
0 Kudos
lywwing
Enthusiast
Enthusiast

hi,

thanks for your reply.

Many thanks.

Reply
0 Kudos