StarWind iSCSI SAN
StarWind Native SAN for Hyper-V
 

Downtime duration if a host fail in a 2-node configuration

Pure software-based VM-centric and flash-friendly VM storage (iSCSI, SMB3, NFS, iSER and NVMe over Fabrics) including free version. Same software embedded on StarWind ready nodes.

Moderators: art (staff), anton (staff), Anatoly (staff), Max (staff)

Downtime duration if a host fail in a 2-node configuration

Postby contatto_decollo » Wed Mar 04, 2020 7:53 am

Hello everybody,

I am interested in testing a StartWind Virtual SAN into an Hyper-V 2-node failover cluster and understanding the behavior of this configuration in case of failure of one of the two hosts.

Let's consider that just one virtual machine is running on such a basic configuration and that a failure (disaster or power failure in general) occurs on one of the two hosts.

Here you are few questions about which I would really appreciate your answers...

1. How would the configuration act in order to preserve the running state of the virtual machine? Would the machine be moved together with the RAM content or would the virtual machine be rebooted from scratch on the other host?

2. According to your experience, simulations, tests, how long would be the duration of the downtime when the failover mechanism moves the virtual machine on the other host? (Less than 10 seconds, less than 30 seconds, less than one minute or longer?)

3. Is it possible to implement the described configuration without a domain (having the servers just into the same workgroup)?

Thanks in advance for your support

Best regards

Giulio
contatto_decollo
 
Posts: 1
Joined: Wed Mar 04, 2020 7:11 am

Re: Downtime duration if a host fail in a 2-node configuration

Postby yaroslav (staff) » Sat Mar 07, 2020 9:20 pm

Greetings!

Let's consider that just one virtual machine is running on such a basic configuration and that a failure (disaster or power failure in general) occurs on one of the two hosts.

If disaster or power outage strikes out, one location is down. Given that, if you have servers in one location they both are gone. If you have them in 2 different locations (i.e., stretched cluster) one location will survive. Do you want me to tell you more about building a stretched cluster?

1. How would the configuration act in order to preserve the running state of the virtual machine? Would the machine be moved together with the RAM content or would the virtual machine be rebooted from scratch on the other host?

When there is hardware failure (say, a server is down), failover occurs. VM gets restarted on the other host. If shutdown was not performed gracefully, no RAM flush occurs. During graceful shutdown, RAM should get flushed.
To ensure true fault tolerance, it is possible to configure VM replicas. Find more here https://www.nakivo.com/blog/vm-failover-guide/.
Sorry, that's beyond StarWind functionality as we provide highly available storage for your services. StarWind VSAN enables to mirror the storage between hosts so that VMs can run on any host.

2. According to your experience, simulations, tests, how long would be the duration of the downtime when the failover mechanism moves the virtual machine on the other host? (Less than 10 seconds, less than 30 seconds, less than one minute or longer?)

By default, it takes 120 seconds in Windows Server 2016 for a VM to migrate. This interval can be shrunk to 45 seconds though. Here is the readout on that https://techcommunity.microsoft.com/t5/ ... a-p/372027. Why it takes so long? For the failover cluster, it is important to distinguish a momentous network interruption and server a going down. Both are associated with inability of nodes to communicate.

3. Is it possible to implement the described configuration without a domain (having the servers just into the same workgroup)?

Having servers in domain enables you to enjoy all the goods of failover clustering. You can configure a cluster without a domain, but what's the point of doing that? Here is StarWind's recommendation on how to configure domain controllers: https://knowledgebase.starwindsoftware. ... san-usage/. Long story short: you can use local VMs. Here is the readout on domainless clustering https://www.mssqltips.com/sqlservertip/ ... ry-part-1/.

Should you require any assistance and if you consider using StarWind VSAN for your production, I recommend scheduling a tech call with your account manager.
If you still have any questions, I am here to help.
yaroslav (staff)
Staff
 
Posts: 218
Joined: Mon Nov 18, 2019 11:11 am

Re: Downtime duration if a host fail in a 2-node configuration

Postby reusegrinning » Fri Mar 13, 2020 4:37 pm

Perfect! Thank u!
Download: run 3 apk online.
reusegrinning
 
Posts: 1
Joined: Fri Mar 13, 2020 4:35 pm

Re: Downtime duration if a host fail in a 2-node configuration

Postby Michael (staff) » Sun Mar 15, 2020 1:10 pm

Welcome!
Michael (staff)
Staff
 
Posts: 308
Joined: Thu Jul 21, 2016 10:16 am


Return to StarWind Virtual SAN / StarWind Virtual SAN Free / StarWind HyperConverged Appliance / StarWind Storage Appliance

Who is online

Users browsing this forum: Google [Bot] and 7 guests