VMware Cloud Community
BaghdadIT
Contributor
Contributor

SQL Heartbeat between two nodes on two different hosts

Good Day VM Guru's

I am trying to build a two node Microsoft SQL Cluster each node on a separate ESXi 6.0 host. I really don't want to build a VLAN just for heartbeat connection unless absolutely necessary, I added a NIC to each VM and IP'ed it (10.50.70.14, 10.50.70.15) but no connection between the two. How do I go about doing this?

0 Kudos
12 Replies
rcporto
Leadership
Leadership

Both NICs are assigned to the same port group and share the same VLAN ? How is the guest firewall configured ? It is allowing ICMP traffic ?

---

Richardson Porto
Senior Infrastructure Specialist
LinkedIn: http://linkedin.com/in/richardsonporto
0 Kudos
BaghdadIT
Contributor
Contributor

Hello

The firewalls are off on Both Nodes,They are on the same Port Group (as far as I can tell) and the same VLAN (470).

0 Kudos
rcporto
Leadership
Leadership

And the VLAN 470 is created on physical switches and allowed on ports where your ESXi are connected ?

---

Richardson Porto
Senior Infrastructure Specialist
LinkedIn: http://linkedin.com/in/richardsonporto
0 Kudos
BaghdadIT
Contributor
Contributor

Yes, It is

0 Kudos
ThompsG
Virtuoso
Virtuoso

Hi BaghdadIT,

Just noticed you have a previous post open where you mention that you are using PVLAN's - is this the same case with this post? If you are using PVLAN's for this solution then make sure that your external physical Switches understand this technology.

If you migrate both VM's to the same host can you ping them then?

Kind regards.

0 Kudos
BaghdadIT
Contributor
Contributor

It is the same thing, I was trying to determine which would be the best way to do it, I think the PVLANS will cause me more grief so I want to try with just a PortGroup. Layer 2 is the way I want to go, But because of my network setup I am having issues with the portgroup.

0 Kudos
ThompsG
Virtuoso
Virtuoso

Hi,

So lets break this down and let me know if we have anything wrong Smiley Wink

  • Two ESXi 6.0 hosts connected via a pSwitch [ESXi01 and ESXi02]
  • The hosts have a portgroup configured with VLANID xxx (Not using PVLANS just straight VLANs)
  • pSwitch ports that the vSwitch is connected to are tagged with the correct VLANs
  • Two Windows Server VMs are configured with network connections from same portgroup [Server01 and Server02]
  • Windows firewall is disabled for testing and there is no firewalls routers between the two hosts

With this configuration when you try to ping from Server01 located on ESXi01 to Server02 located on ESXi02 there is no response and the ping fails.

Now for question time: Have you tried moving Server01 to ESXi02 so that both VMs are on the same host? Does the ping still fail at this point?

0 Kudos
BaghdadIT
Contributor
Contributor

Yes Sir, that is all correct and I apologize for my late response, I had another project take priority

0 Kudos
ThompsG
Virtuoso
Virtuoso

No problems at all - completely understand what that is like Smiley Happy

So even on the same host on the same portgroup, the two Windows servers cannot ping each other? If this is the case then it is likely there is a problem with the networking within the OS guests. I would start by checking that the Windows firewall is definitely disabled on both VMs and that they are both in the same subnet.

With the VMs on the same host, what do you get if you run a tracert (Trace Route) between them?

0 Kudos
BaghdadIT
Contributor
Contributor

With them on the same host, I can ping them on those addresses , only when I put them on separate hosts

0 Kudos
BaghdadIT
Contributor
Contributor

So I just want to make sure we are on the same page, these are going to be 'Heartbeat" connections, so the seconf NIC is only configured with an IP and a subnet mask,

0 Kudos
ThompsG
Virtuoso
Virtuoso

Hi there,

Excellent - so you have confirmed that there is no OS related issue between the two guests as the servers can ping each other when they are on the same host. This means it is likely we are dealing with an ESXi or physical networking configuration issue.

Not sure the best way to proceed here but perhaps you could post some screenshots of the network configuration for the portgroup both VMs are connection to from both hosts? You could also do this via the ESXCLI if you wanted.

Also if you are able to also provide the physical port configuration from your pSwitches that the physical network adapters are connected to?

Sorry another dumb question - completely understand that this is to allow cluster heartbeats across but do you have the same issue with the network that clients will be connecting to or is that to be determined yet?

Kind regards.

0 Kudos