This functionality is currently in preview. If interested, please sign up here: http://aka.ms/migrate/previewsignup
This is an interesting ask. Kindly send out any comments around your PAAS apps, their networking structure, storage requirements etc in the comments,
This is a very new ask. If you have any more requirements around this kindly let us know in the comments.
This is a very new ask. Could you give some idea around any other requirement you have in the case? Types of VMs, networking info etc. We have this on our radar but would like to know more to deliver a great feature.
We have four votes on this so I believe others also share this concern. In case you want to discuss how we can improve it, feel free to directly write to me at firstname.lastname@example.org.
Hi Stanislas -
Planned failover is a planned activity in which you want to ensure 0 data loss. For that we have a clear gesture and a behaviour that will ensure 0 data loss. If during the execution of a planned failover, any thing fails - some script, final delta transfer, creation of any VM, access to storage etc - the failover stops completely. This ensures that the failover does not proceed and the next group VMs which have dependencies also don't fail due to it. Only if every activity succeeds, a planned failover will succeed.
An unplanned failover is when the primary site is not reachable - but that is not the only disaster. Sometimes, the primary site is reachable but you want to initiate a failover to salvage all the latest data, because the site might go down right away. If you initiate a planned failover, it might keep failing at different times due to failures, and you need to retrigger the failover - increasing the RTO.
In a tense situation like disaster, minimizing data loss is a luxurious wish - but quick RTO is also important.
Hence we give an option to opt in - to try and salvage any primary site data. If that fails, the job still continues and recovers the applications to the last known best time. that way you get a mix of both.
When you doing a peace time operation such as maintenance, migration, compliance based failover to recovery site - do a planned failover.
When things are going bad or you have already hit with a disaster - do an unplanned failover.
Let me know your views on this.
We have planned for this ask.
Kindly let us know if you are OK with a data loss when we shutdown automatically during failover.
We are working to understand the requirements of this feature and deliver it. If you have any scenarios or examples that might help us take a decision kindly give a detailed comment below.
Do you have instances of shared VHDX deployments in your environments?
Could you give an idea of how many such VHDx’s might be there in a typical customer environment?
Also how is Shared VHDx used?
1. Over SMB?
2. Over iSCSI with VHD?
3. Over iSCSI with pass through disk?
Today shared VHDx based Virtual machines are not supported to be protected using ASR. Need protection for them.