Run Failover automatically if VM down

[gian202b]

Hello,
I've been doing a lot of research but have not been able to find a solid solution as of yet.

This is what I'd like to do. Based on Veeam ONE alarms - if VM looses connection for x amount of time I would like to run a script that would start the VAO failover. (I would add multiple checks in the Pre-Failover to make sure that the VM is really unreacheable)

I found on previous posts instructions on using Veeam ONE and powershell to run directly on VBR but I want to use VAO.
I also tried with using the VAO API but it's not running as smoothly as I'd like.

If anyone has any insight into this I'd appreciate the help.
Thanks

[Oleg Patrakov]

Hi
Welcome to our forums and to VAO

Actually, the only way to do it now would be to use APIs. And we are not planning to add any "auto-start" in the future, as it's an intricate matter and decision should be done on the user side, according to the processes used in the company.

Could you share what exactly are your struggles with VAO API? What is your VAO version?
Maybe I can help you pulling through them now on your version, or at least we can make some improvements for the next one.

Cheers,
Oleg

[gian202b]

Hi Oleg,
Thanks for your reply.

VAO version 1.0.0.895, which I believe is the latest.

Issues with API - I started looking into the swagger UI to understand exactly how they work, then moved to powershell and was able to login, get plan ids, enable and disable plans without an issue.

Once I tried running my test failover plan I would get an error while processing the replica VM stating "Object reference not set to an instance of an object." (Fails within seconds)
I also noticed that the plan report is a empty when this happens.

This is the body of the POST
{
"mode": "FailoverNow",
"planLaunchOptions": {
"restorePointRule": {
"date": "2019-01-29T14:41:11.749Z",
"mode": "NearestBefore"
},
"executeChainedPlans": true
}
}

Any insight you'd be able to give me would be appreciated.
Thanks

[Alec King]

Hi,
Thanks for the additional info. The errors you are getting from the VAO API are definitely not expected, and we'd like to investigate further.
Could you open a Veeam Support Case please?
Support team will gather all the logs and other info we need.
If you could post the support case ID here when you get it, that would help us connect with them faster.

Thanks in advance!

[gian202b]

Here is the Case info: Failover plan fails when run from VAO API - Case # 03387688
Based on first reply it seems as though it's an issue with an internal command not getting a variable it's expecting. They will be doing further investigation.
I will post a reply once solution has been determined.

Thanks

[Alec King]

Thanks! My R&D team are also investigating. Hope to have an answer for you soon!

[gian202b]

Hello,

Just wanted to give everyone an update. It's been a while.
After much back and forth with support and R&D team, they confirmed it was a bug that would be worked out in the VAO v2.


I have just upgrade to v2 and will be doing some testing to make sure it was fixed.

[Alec King]

Thanks for the update - please post here your progress with VAO v2, thanks!

[Mildur]

Was there any change to this feature request?
Do we still need the RestAPI to do an autostart of a Orchestration Plans?

[Alec King]

Hi, it's correct we do not auto-failover. We considered this a somewhat risky approach, as many different factors could make a VM seem failed when it is not (such as network disconnect from vCenter).
You could set this up in Veeam ONE which can run a script that triggers Orchestrator API when a certain Alarm fires.

[Mildur]

Thanks Alec. i understand the risk with starting a failover automatically. No one wants to have a split brain situation

We will have a project next year, which could led us using the VAO product.
Veeam One is already in use. A script would be an option.

[Marcus1414]

I have VRO deployed and I do have existing VBR replication to DR VMware vSphere. Would like to know is it possible for me to setup a plan that execute automatic failover to DR if it detects my production VM is down? Or I can only setup a schedule?

[Mildur]

Hi Markus

I moved your question to this topic. Funny, I asked the same question before I was working for Veeam.

We don't provide automatic failover. There are a lot of parameters on decide to if a source infrastructure is offline. This decision must be done by the backup administrator. Please see the answer of my Alec King.

But you may use external scripts to monitor your source environment and start a Failover plan.
The script should make sure that the source VMs are definitely powered down or offline. If not, it will lead to split brain. Meaning some users could work on the source server, some on the restored server.

Best,
Fabian