Comments on: 2-node hyperconverged cluster with Windows Server 2016

By: Backrightup

Backrightup — Tue, 29 Jun 2021 09:51:59 +0000

DevOps is a new trend in software development over the past few years. Although the term is relatively new, it is actually a combination of many practices that have been in use for years. DevOps allows software to be released quickly and efficiently, while still maintaining high levels of security. for more visit Azure devops backup

By: Stu

Stu — Fri, 05 Feb 2021 10:38:26 +0000

I worked it out thanks to this article. Instead of using a single physical NIC for management traffic, I used a SET with a virtual management adapter across both servers and cluster worked. No more issues with UDP port 3343. Thanks Romain 🙂

By: Stu

Stu — Wed, 03 Feb 2021 20:16:50 +0000

Forgot to say I have also posted issue with UDP 3343 here: https://docs.microsoft.com/en-us/answers/questions/249241/cluster-network-validation-fail-udp-port-3343.html?childToView=257695#answer-257695

By: Stu

Stu — Wed, 03 Feb 2021 20:07:26 +0000

Hi Romain,
I also have the same problem as Damon Johnson and firewall disabled completely. If I put hosts in differnet subnets they pass the cluster validation but fail to create the cluster due to following error in the cluster log:
‘INFO [NODE] Node 1: New join with n2: stage: ‘Attempt Initial Connection’ status (10060) reason: ‘Failed to connect to remote endpoint 10.63.37.5:~3343~’
I emailed you as suggested and would appreciate any help or advice you may have with this.
Thanks

By: Federico Mapelli

Federico Mapelli — Tue, 29 Dec 2020 01:54:06 +0000

Hi Romain, Great Article! Clusters is a new world for me, so I’m probably going to ask obvious questions.

I’ve two servers with Hyper-V Server 2019 and many VM’s on each. DC’s are 2 VM’s, 1 in each server so i have a sort of HA using 2 dns, dhcps, dfs-namespace and replica. and so on…they work fine. Obviously some applications and service don’t have HA.

I’d like to adopt a new configuration using Failover Clustering using storage space direct because a can’t buy a san or something similar. So you article is a great guide for me.

There are some aspect i can’t understand…
I was thinkng to make a cluster with 2 server with Hyper-V Server 2019 on (not windows server ), is it possible?

You suggest to put dc outside cluster as a phisical server, but it could be a point of failure.. so i have more physical server for dc, COrrect?

Why don’t you put DCs into the cluster? i’ve read that the cluster boot also there isn’t an active domain controller from WS2016. I think the answer is about the time DC take to boot on the alive node when the other node fail…and if i have more than 1 dc into the cluster nothing change… dcs native ha is the best way… is it right?

Last question is about the HD: now, i’ve a lot of SAS 2.5″ 10/15K in raid 5 or 50 configuration plus other 3.5 sas storage. You say storage direct work only in jbod mode (not raid) and if i have raid controller they are useless and degrade performance.. How can i save my money using my hd and controller?

Really, your solutions should be great for me but your is a test lab and use Hyper-v to virtualize the 2 DCs and the cluster itself but could this solution have sense in a production enviroment?

Thank you so much… i’m sorry but i’m not an it manager

By: Reha-Klinik NATURANA

Reha-Klinik NATURANA — Thu, 10 Dec 2020 12:19:16 +0000

Hi Romain, thanks so much for this article. I am at building my first 2 node cluster on Windows Server 2019 and looking for a step by step tutorial about the network configuration I found yours. I am not sure about “Set-NetAdapterRSS” because each of my servers runs 2 Xeon CPUs with 8 cores each. How many cores should I assign to each network adapter?

By: marekjs

marekjs — Fri, 23 Oct 2020 12:24:01 +0000

Hi Jean. I have feeling, you have to recover from an image, or build from scratch…

By: Jean

Jean — Tue, 04 Aug 2020 12:14:59 +0000

Dear Romain, great article. If i lost one node(replace), how i rebuild the node?

By: Stefan

Stefan — Wed, 29 Jul 2020 07:12:19 +0000

Dear Romain,
I have made a complete new setup after my old one went into strange and not really understandable problems. For this setup I have put my DC out of the cluster as I feel more comfortable with this..

However, during the new setup with Server 2019 I am running into two issues :

1. Test-Cluster is not working as it says that test “Network” was not found in the list of loaded tests.

2. However, I can run the test with the Failovercluster-Manager. It runs into a Network warning saying that DHCP status of vEthernet(Management-0) is differing from network “Cluster Network 3”. Of course I have given vEthernet(Management-0) a fix address.

The strange thing is that “Cluster Network 3” at that moment does not exist yet…

Any idea ?

Thanks a lot,
Stefan

By: Ivaylo Valkov

Ivaylo Valkov — Tue, 14 Apr 2020 11:25:57 +0000

Great video, congrats! Quick question. What will happen if one node fails? I do not mean shutdown/restart but total fail like power, for example, or NIC issue. The VMs running on the affected node will be transferred to the new right, but in what state? What about VMs with large amount of RAM?