VMware Cloud Community
babaeih55
Enthusiast
Enthusiast

vSphere HA detected a possible host failure of this host

Hello,

I have been building a lab environment using the hardware below with VMware Workstation 8.

Physical Hardware:

Dell Precision 690

Intel Xeon CPU 3.73GHz (2 processors)

Memory: 12GB

Windows 7 Pro 64bit

Virtual Setup:

Workstation 8

VM for a DC

VM for vCenter

VM for FreeNAS 8

VM for ESXi5 1

VM for ESXi5 2

I have everything built out but after configured the HA.  I have been getting the following errors:

1. Under summery tap of the ESXi5 host.

            vSphere HA State:  Host Failed

The vSphere HA Master Agent connected to vCenter Server cannot reach the host over the management network or via the configured heartbeat datastores.

It will attempt to restart the vSphere HA protected VMs that were running on this host at the time that the failure was observed.

2. Under Tasks & Events of the ESXi5 host.

            Configuring vSphere HA:

A general system error occurred: Failed to increase memory reservation and limit for host/vim/vmvisor from -1 to 100.  HA may not configure on the host.

I have tried searching the forum and haven't had any luck.  I have also search Google and no luck.

Any help would be greatly appreciated!

Thanks in advance!

BB

Reply
0 Kudos
16 Replies
Virtualinfra
Commander
Commander

Try this out

Select the ESX host in vCenter Server.

  1. Click the Configuration tab, then click System Resource Allocation > Advanced.

  2. Under System Resource Pools, select host > vim> vmvisor, then click Edit Settings.

  3. In the Edit vim Resource Allocation window, set the Memory Resources limit to Unlimited


Also you need to check if the both ESXi host are able to communicate each other.Turn OFF firewall on vCenter server

Also take a look at the below Errors for HA on vCenter 5.0

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=103363...

Thanks & Regards Dharshan S VCP 4.0,VTSP 5.0, VCP 5.0
babaeih55
Enthusiast
Enthusiast

Thanks for the suggestions Virtualinfra.

1. I have tried setting the Memory Resources limit to Unlimited with no luck.

2. On turning the firewall of.  Can you please elaborate on that?  I have disabled the Windows firewall on the vCenter server.  I have connected to each host through vCenter and opened the ports for Fault Tolerance, vSphere HA Agent, and vmotion with no luck.  I have connected to each of the ESXi5 host through putty and tried to ping the others management console and the vMotion vSwitch IP with no response.  I have tried vmkping and still no response.

3. I have reviewed the link with the vSphere HA and FT Error Messages and wasn't able to find any of my errors. 

I have been tring to resolve this issue for 3 days searching everywhere with no luck.  I have also attached screenshots of errors and the firewall port info.Error1.jpg

Error2.jpgFirewall Ports.jpg

Thanks again for all your help!

BB

Reply
0 Kudos
Virtualinfra
Commander
Commander

1. Is that both the host are pinging each other.

2. is both host are resolving DNS( to nslookup).

3. have tried disable and enable HA - if not please try...

4. what is the RAM allocated to Each VM.( please assing atleast 3 GB..

5. Please put the screen shot of host/vmvisor memory is reservered or not

Please award points for correct/helpful responses by clicking the tabs

Thanks & Regards Dharshan S VCP 4.0,VTSP 5.0, VCP 5.0
Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

1. Is that both the host are pinging each other.

a. When I ping from ESXi5-01 to ESXi5-02 I don't get a response and vice versa.

2. is both host are resolving DNS( to nslookup).

a. I can resolve the name on both hosts.

3. have tried disable and enable HA - if not please try...

a. I have tried disabling and enabling HA.  I have also tried "Reconfigure for vSphere HA" under the Summary tab with no luck ether.

4. what is the RAM allocated to Each VM.( please assing atleast 3 GB..

a. So are you saying increase the memory on my VM that is in my cluster from 1Gb to 3GB?

5. Please put the screen shot of host/vmvisor memory is reservered or not

a. Screenshot below

vmvisor Resource Allocation.jpg

Thank you again!

Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

I have also attached screenshots of the networks on both hosts.

ESXi5-01 Network Setup.jpgESXi5-02 Network Setup.jpg

Thanks,

BB

Reply
0 Kudos
Virtualinfra
Commander
Commander

Both the ESXi host are not pinging. If the do not communicate HA wont work..

1st we need to resolve this Issue and if you enable HA it will work..

What is the gate way set to each of the server?

Is both the servers are able to ping the gate way?

Trouble shoot why both the host are not ping.. if both the host starting ping then your HA will work fine.

HA work by keeping the heart beat communication between both the ESXi host. If there is not connectivity then you wont be able to configure the HA.

"Please award points for correct/helpful responses by clicking the tabs"

Thanks & Regards Dharshan S VCP 4.0,VTSP 5.0, VCP 5.0
Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

So here is what I have...

I can ping ESXi5-01 and ESXi5-02 from my DC and vCenter server.  I can resolve the name.  I can ping the management console and the vMotion vSwitch. 

From both ESXi5 host I can pint the DC, vCenter server and storage.  I can also ping the vMotion vSwitch while connected to one of the ESXi5 host (by this I mean I can be connected to ESXi5-02 and I can ping ESXi5-01 vMotion vSwitch and get a response.)  However, I can not do the same when trying to ping the management console of the VM network.

I have changed things on the VMware Workstation 8 network by changing the way the VMs use bridged mode, NAT and Host only with no luck.

This thing is driving me nutts!  :smileyconfused:

Any other ideas?  If you need more info please let me know.

Thanks again!!!

BB

Reply
0 Kudos
martin_schmidt
Enthusiast
Enthusiast

Hi,

Facing the same here.
Installed three ESXi 5 servers as nested VMs on an ESX 4 host.

As soon as I enable HA, one host gets elected as master. The other two complain being network partitioned.
Tried each and every HA advanced parameter (isolationaddress etc), but no success.

fdm.log:
2011-10-27T20:16:55.050Z [6A9AEB90 error 'Election' opID=SWI-d07609b5] [ClusterElection::SendAll] [50 times] sendto 10.10.10.80 failed: Host is down

Ping to VC, DNS etc. works fine, but not between ESXi hosts.
Disabled firewall on hosts completely (esxcli network firewall set --enabled false).
But no success.

~ # ping 10.10.10.80
sendto() failed (Host is down)

Still no solution. 😞

Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

I resolved the problem with being able to communicate between the ESX host.  The problem I was having was caused by having the same MAC address assigned to both management networks and vMotion vSwitch on both hosts.  When you are nesting ESXi5 on another virtual platform you are sharing the physical NICs. 

After I resolved this issue by adding more physical NICs.  I was able to ping between the ESX host and my entire network.  However, I still can't get my HA to work.  Moreover, I have errors on both of my ESXi5 host now. 

HA Error1.JPGHA Error2.JPG

Any ideas?

Thanks!

BB

Reply
0 Kudos
martin_schmidt
Enthusiast
Enthusiast

Also fixed my issue.
Root cause was probably that I cloned my ESXi VMs.
So I reinstalled two of the three and no more communication issues between all of them.

Regarding your issue:
If your networking is fine, try:

Disable firewall on vCenter server.


Disable HA on the cluster.
Remove ESX hosts from vCenter.
Re-add ESX hosts to vCenter.
Re-enable HA.

Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

Martin,

I'm glad that you have your setup working.  That's for the suggestion and I tested it out with no luck.  I was reading the following article and with vSphere 5 I need to have 2 datastores in order for HA to communicate.  I have added a second datastore and still no luck. 

http://www.vladan.fr/vsphere-5-new-high-availability-agent-introduced/

Thanks!

BB

Reply
0 Kudos
babaeih55
Enthusiast
Enthusiast

I FINALLY got it!!!  I had to delete the two hosts and recreated them.  Once that was done HA started working like a champ!!!

Thanks!

BB

Reply
0 Kudos
depping
Leadership
Leadership

You don't need two datastores in order for vSphere HA to work. You can disable this, just check my article on that topic:

http://www.yellow-bricks.com/2012/04/05/the-number-of-vsphere-ha-heartbeat-datastores-for-this-host-...

Reply
0 Kudos
PramodKhalate
Enthusiast
Enthusiast

Nice article Deeping... it will very helpful..

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
Reply
0 Kudos
TraianBratescu
Contributor
Contributor

Hi,

Had the same issue with a lab created from cloned esxi hosts; Everything seemed ok until i tried to enable HA and than I got errors like:

Reconfigure vSphere HA host Operation timed out.

Cannot install the vCenter agent service. Unknown installer error

Network partitioned

HA disabled

After some hours Smiley Happy I discovered that the vmknic had the same mac addresses which caused all the trouble; I found these links usefull:

     *     VMware KB: vmk0 management network MAC address is not updated when NIC card is replaced or vmkernel ...

     *     http://www.vhersey.com/2012/08/cloned-esxi-duplicate-vmk0-mac-addresses/

     *     Change the MAC Address of VMKNIC on ESXi 5 - Bonus Bits Wiki

Hope this will save some time,

Traian

Reply
0 Kudos
vmstoani
Contributor
Contributor

bb

Reply
0 Kudos