Cachestorage/Recovery vault for site recovery Manager best practices

[Chandrasekhar_Arya]

What are the suggested best practices for Cache storage . What is the default recommendation on no of VMs that should be part of same cache storage . 

Where should the recovery vault for site replication should be shall I configure on DR site with Geo replication enabled assuming 

[#2]

It depends on what type of cache storage (app cache like redis or something else) and what is it used for?

[#3]

@Chandrasekhar_Arya 

Hi 

• Use GPv1 storage when you are using a custom storage account cache .Usage of GPv2 is not recommended because transaction costs for V2 are substantially higher than V1 storage accounts . Also premium storage are not supported. 
• The storage account can be in the same or a different subscription but should be in the same region as the virtual machines being protected 
• Ensure that Allow trusted Microsoft services'.  is checked in Firewalls and Virtual Networks when the access is allowed from Selected Networks and not from All networks . Also, ensure that you allow access to at least one subnet of source Vnet.
• Use multiple storage account when using large amount of virtual machines because there are some limits with the number of disks that can be replicated . Those  limits are specific to Azure to Azure and zone to zone DR scenarios.

The table below can help to calculate how many storage account you would need for a given number of virtual machines and disks . 

Storage account type Churn = 4 MBps per disk Churn = 8 MBps per disk

V1 storage account	300 disks	150 disks
V2 storage account	750 disks	375 disks

You can find the support matrix here and almost all the informations listed above 

: Support matrix for Azure VM disaster recovery with Azure Site Recovery - Azure Site Recovery | Micro...

[#4]

@ibrahimambodj 

Thanks for that info.

@Chandrasekhar_Arya 

It is the use case /purpose for which you are using a "cache" that I am interested in firstly.

[#5]

Cache storage is needed to copy the disks info in the source datacenter for DR purpose refer this article https://docs.microsoft.com/en-us/azure/site-recovery/azure-to-azure-architecture that show the Cache storage account You need a cache storage account in the source network. During replication, VM changes are stored in the cache before being sent to target storage. Cache storage accounts must be Standard.