Need HELP! THE 2nd node can not join / recognize the cluster

[netappmagic]

Hi,

I finally set up the cluster with 1 node and with adding 2 etra network adapters as instructed by the tutorial video.

When I set up the 2nd node (added 2 extra network adapters as well), and in the end, it prompts me to enter the cluster name, which I did, then it shows me trying to join, and then network setup.. then it loops back to me and ask me if to create or join the cluster again. pls see the attachment. So, i am in the loop.

Are there any solution for that?   This is 2nd time. The first time, I thought I did something wrong, so, started over from scratch.

What is the cluster IP and I can find from the existing 1 node cluster?

If I stopped the prompted questinair, at one point, it asked me to enter cluster IP to recognize. What CLUSTER IP i should use? The cluster with the other node is running fine.

Also, based on the video, when I need to change default subnet in the end of adding 2 extra adapters, however, since vmware play doesn't have VM Editor, I skipped this step.

Appreicate your help!

[shatfield]

Did you set the sysid/serial number on the 2nd node before running option 4?

[netappmagic]

yes, I did.

The question is, I did the change after I modified .vmx file for changing the serial port and adding two extra network adapters. is that alright?

Also, also since VM PLAYER doens't have VM Editor, I did not change the subnet for 2 new added NW adapters, as suggested by that video. Is that alright?

[shatfield]

Odds are no, so since you can't change the subnet of the NAT network in player, you have to figure out what subnet its actually configured for and adjust your IP settings accordingly during the cluster setup script.

The "NAT" network is really VMNET8, so check your host machines ipconfig /all for the ip range/netmask/gateway for VMnet8.  Use IPs in that range instead of whats called out in the simulator documentation, and you won't need the vm network editor.  Just assign nic0 and nic1 to "host-only" and the rest to "nat" and it should all fall into place.

[netappmagic]

I have already run ipconfig /all, and the subnet on  my wondow is 192.168.14.x, so, i used 192.168.14.101 for the cluster mangement IP, 192.168.14.91 for the node1 management ip, as suggested by simulator doc.

however, the IP for e0a and e0b on node1 in the already created cluster is configured by the process as default. the same on node2.

pls find attached the output of "net int show" on node1 in the cluster.

I guess, I am not sure what you wanted me to do.

thanks!

[shatfield]

That looks fine, the cluster lifs on e0a/e0b are autoassigned in the 169 network.  If the join fails, on sibsequent attempts those ips have already been generated so you can just accept the previous values.

Changing the serial needs to be done during the first power on.  Adding nics should be done before the first power on, so it sounds like you've done everything correctly.

Whats the message you get when the join fails?

[netappmagic]

Hi there,

All messages have already attached in two attachments.

I have just re-done node2, this time without adding an extra adapters. Everything went fine, and I am not able to join the cluster.

Now, node1 has 2 more NW cards than node2.

Can I now go ajead tp add those 2 more cards on node2? any extra steps in the sake of adding 2 cards?

[shatfield]

You'll likely have problems booting since all the PCI devices now have slot assignments.   So after adding the bridge code, when you boot you'll get an error about not enough pci slots.  Find the device(s) referenced in the error and change their slot ID to -1 by editing the vmx file.  Then you will probably be able to boot.

[netappmagic]

Hi SHATFIELD ,

it seems alright now. I added 2 extra NIC's on node2, and able to bring it and the cluster up.

The problem now is that I could not putty in and also ping the cluster mgmt IP, 192.168.14.101. Please see my screenshot of "net int show" on both node1 and node2. They all show 192.168.14.101 is THE ip for the cluster mgmt. I can ping node mgmt IP's: 192.168.14.91 and .92, as well as .1

The only possible cause I can think of is that the same IP was used for an already removed cluster. However, the old cluster  is already deleted, plus, I already did "putty -cleanup", and also removed all hosts from Registry.

Could you please offer any advice to fix?

Appreciate your patience.

[netappmagic]

>>Changing the serial needs to be done during the first power on.

I don't quite understand this comment. Based on Installation document, changing the serial ports need to be done before first power on. this is when I did the change. I also subsequently added 2 nics.

Would that be an issue?

[netappmagic]

Hi,

Could somebody pls help.

the cluster mgmt IP is not pingable, and then could not putty in. It looks up, pls see the the attachment in previous message for "net int show" output

[shatfield]

What address did you use for the default gateway? And can you ping it from inside the simulator?

[netappmagic]

Hi there,

As you can see in the attached screenshot.

I can ping anybody else inside the simulator, including the gateway 192.168.14.1, but not the mgmt IP: 192.168.14.101.

Any idea pls, or do I have to rebuild again? Thanks!

[shatfield]

Might be on to something there.

The gateway on the nat network is almost always on .2

But even so, you should be able to putty in to the node interfaces.  If thats working, but the cluster management lif is not, we just have to fix the lif/routing groups.

[netappmagic]

Yes, I can putty in both nodes in the cluster.

Could you please share with me details commands to fix the issue? Otherwise, I am thinking to rebuild.

Based on "Step by Step installation" doc, the gateway should be 192.168.x.1, x  --> 14 here when prompting  me to enter cluster mgmt IP. So, I am not sure next time when I rbuild should I use .2 or .1, or if it matters.

[netappmagic]

I am wondering, if it is 2nd cluster, should I use the same subnet as 1st one, the same "14" when I type ipconfig/all, or different one?