VMware Horizon Community
Suiname
Enthusiast
Enthusiast

View 5 Linked clones fail to get DHCP IP

Hello all,

  I'm running a small vmware view 5 environment here on 3 esxi hosts.  Each of these hosts has a few nics which are zoned for a specific VLAN which has a range of DHCP addresses we want to use for our VDI environment.  The problem I am running into is that when I create a linked clone pool of about 20 VMs, half of them get IP addresses in the DHCP range and half of them get a 169.254.xxx.xxx IP (which is bogus) and does not have any internet connectivity, causing the provisioning process to be stuck in the "customizing" stage forever and eventually the connection server gives up because it can't reach the VM.  I am able to solve this by manually refreshing each of the clones which does not work, however sometimes I need to do 2 or 3 refreshes before it works.  Does anyone have any insight into why this might be happening?  I already know the following:

-best practices say to do a ipconfig /release on the clone master before you power off/ snapshot for the baseline.  I have done this.

-DHCP range has more than enough addresses (over 100, only 20 clones)

-All the hosts can reach the subnet in question, I have gotten VMs with connectivity to this subnet and migrated it to every VM to ensure it still has internet connectivity.

Thanks in advance.

Reply
0 Kudos
11 Replies
kgsivan
VMware Employee
VMware Employee

I suspect that one of the ESX may not have correct network labeling (for a Virtual machine network). Is it so ?

Else can you ensure the all esx has sufficient vSwitch ports available ?

Reply
0 Kudos
Linjo
Leadership
Leadership

Do you have enough ports available in the vSwitches?

// Linjo

Best regards, Linjo Please follow me on twitter: @viewgeek If you find this information useful, please award points for "correct" or "helpful".
Reply
0 Kudos
Sergei13
Enthusiast
Enthusiast

Also what OS are your VMs working on ?

I have been having all sorts of problems with Windows 7 x64, which are still under investigation (what they say :smileyangry:) with VMware support.

So far I have wasted more than a month on it and still nothing near stability !

Either no netowrk card, or network card but no connectivity, Nic check boxes not ticked in the configuration, only E1000 working, impossible to get VMXNET3 to work (Microsoft patch not required even if mentionned in VMware W7 best practice) ... and on occasion I eventually have some machines available !!!

Suiname
Enthusiast
Enthusiast

Thanks all for your responses.  I do have the subnet labeled  identically on each host, VLAN and network name are the same.  I have  120 ports on each vswitch on each host, so I also have more than enough  switch ports to accomodate 20 clones.  I am using windows 7 x64, I did  not realize that running 64 bit windows linked clones might be such a  painpoint.  I didn't see anywhere in the documentation or best practices guide about not being able to use 64 bit windows 7.  Sergei, does VMware know this is an issue and have they documented it anywhere, or did you just find this out through a support call?  Thanks in advance

Reply
0 Kudos
royberk
Enthusiast
Enthusiast

This is interesting because I successfully deployed Win 7 x64 in View 5.

Is this your first LC pool ? Do you have any non-Win7 x64 machines youve created already?

I would like to see if you create a different pool if this issue persists. That would rule out DHCP options or Networking options within VMWare.

Reply
0 Kudos
Suiname
Enthusiast
Enthusiast

I do have other pools, but they are also win 7 x64 and experience the same issue. I am going to try a win 7 x32 pool and see if there is any difference.

Reply
0 Kudos
royberk
Enthusiast
Enthusiast

That would be a good starting point because I dont necc. think its a Win 7 x64 issue.

Or maybe Im very lucky and they happened to work Smiley Happy

**Edit: Also, how is your base image built? Do you have a VM Network assigned to it . Is the NIC option Connected and Connect on Power On checked?

If for some reason a network setting is misconfigured and upon power up on a compose cant reach back to grab DHCP, this would def. do it.

Reply
0 Kudos
Suiname
Enthusiast
Enthusiast

So I tried creating an x86 pool, I have the same issue.  It seems I will need to be contacting VMware support on this one, really scratching my head as to why this is happening.  I do have the NIC on the parent set to both connected and connect at power on, the network is labeled properly and set the same on all the hosts (both the VLAN and the network label) and I have plenty of ports available on all hosts.  I am puzzled, but hopefully VMware support can give me some answers.

Reply
0 Kudos
royberk
Enthusiast
Enthusiast

How big is your DHCP scope?

Is it running out of addresses to provide?

Reply
0 Kudos
Suiname
Enthusiast
Enthusiast

we have 200+ addresses in this subnet available via DHCP.  There are only 20 clones so they aren't maxing out the DHCP range.

Reply
0 Kudos
Suiname
Enthusiast
Enthusiast

So it turns out that this was in fact a DHCP issue.  One of the network engineers had added a redudant router and their was IP spoof filtering enabled, which caused the active router to think that IPs being issued by the passive router were spoofing IPs and automatically dropping the traffic.  Whoops.  Anyways, turning that off fixed this issue.

Reply
0 Kudos