Thank you for going for it!
Just in case here are the guidances you might find helpful:
VMware 2 node HC
Hyper-V 2 node HC
StarWind HA device configuration

@ardeyn Thank you very much for mentioning StarWind.

StarWind HCA should fit your needs perfectly. It is based on Dell hardware and will provide you VMs with pure 100% synchronous HA. This solution is also fully supported by a single vendor and comes to your site preconfigured.
We are also Dell OEM partner, thus less expensive than our competitors.

@DustinB3403

@DustinB3403 said in Replacing the Dead IPOD, SAN Bit the Dust:

@dafyre said in Replacing the Dead IPOD, SAN Bit the Dust:

Why not RLS a la StarWind ?

Because the hosts aren't uniform.

In case they need a SAN replacement, they can consider going for StarWind Storage Appliance. That solution will give them a synchronous HA storage pull that can be shared via iSCSI to their existing hosts. StarWind crew will get those onsite fully preconfigured and will help with designing the migration plan as well as executing it.

StarWind Virtual SAN is something that would allow achieving your goal. We can synchronously replicate the storage across 2 sites thus provide you with the ability to build geo-cluster. Then instead of upgrading your SQL to the enterprise edition, you could simply go with the FCI approach that is available in the standard edition.

An additional benefit of connecting a local drive would be the option of creating a shared storage replica. There is also an option of connecting a device called AcloudA, it would allow you to create a volume that would be connected directly to the cloud. Take a look here for more details - http://aclouda.com/

Theoretically, the guest system is totally isolated by the VM and cannot even "see" the host, let alone attack it; so the guest cannot break out of the VM. Of course, in practice, it has occasionally happened. An attack requires exploiting a security issue (i.e. a programming bug which turns out to have nasty consequences) in the VM implementation or, possibly, the hardware features on which the VM builds on. There are few exit routes for data out of the VM; e.g., for Internet access, the VM is emulating a virtual network card, which deals only with the lowest level packets, not full TCP/IP -- thus, most IP-stack issues remain confined within the VM itself. So bugs leading to breakout from VM tend to remain rare occurrences.

There are some kinds of attacks against which VM are very effective, e.g. fork bombs. From the point of view of the host system, the VM is a single process. A fork bomb in the guest will bring to its knees the scheduler in the guest OS, but for the host this will be totally harmless. Similarly for memory: the VM emulates a physical machine with a given amount of RAM, and will need about that amount of "real" RAM to back it up efficiently. Regardless of what the guest does, the VM will never monopolize more RAM than that. (You still want to limit VM RAM size to, say, at most 1/2 of your physical RAM size, because the extra "real" RAM is handy for disk caching; and the host OS will want to use some, too.)

Taken into account that malware/ransomware is implemented on the file level, the best method of protection would be based on a block level recovery tool.

In addition, the best way to overlook the possibility of loosing your data would be the implementation of the the 3-2-1 rule. Its a quite common safety measure in the data infrastructure. We actually implement it quite often, it is based on the replication of your data between 3 nodes as well as creating 2 real-time replication copies of the data between the nodes and storing a single copy of your data in a VTL on the cloud.

For any other additional information, I would like to suggest you to take a look at the following article - https://knowledgebase.starwindsoftware.com/explanation/the-3-2-1-backup-rule/

In case you eventually decide you need HA. I would recommend trying out StarWind Virtual SAN Free it will eliminate the need in physical SAN devices and instead use internal server storage to share it via NFS or SMB 3.0.

@dafyre said in EMC ScaleIO Available for Free:

This looks pretty interesting. It would be a competitor to say Starwind & co?

It cannot be a competitor to us as it is not allowed in production, where we do allow using our Free version in prod.

@djdanilo I would highly recommend looking at StarWind Virtual SAN.
With it you will be able to configure redundant storage pool. It will allow you to create a 2 node cluster at your branch office side and present synchronous storage pool to the cluster as a CSV. This way in case of a failure, your SQL server will simply migrate to the partner node and continue running, as for the storage, there will be 0 downtime since StarWind has synchronous active-active replication.

You could build a 3 node HyperConverged Cluster using old servers with StarWind Virtual SAN as a storage provider and Free Hyper-V as the Hypervisor. In case you decide going that path, ping me offline and i will help you out with the configuration.

@StrongBad
All you would have to do is PM your contact details for me to pass them on to the representatives which would contact you in the nearest time.

Theoretically, the guest system is totally isolated by the VM and cannot even "see" the host, let alone attack it; so the guest cannot break out of the VM. Of course, in practice, it has occasionally happened. An attack requires exploiting a security issue (i.e. a programming bug which turns out to have nasty consequences) in the VM implementation or, possibly, the hardware features on which the VM builds on. There are few exit routes for data out of the VM; e.g., for Internet access, the VM is emulating a virtual network card, which deals only with the lowest level packets, not full TCP/IP -- thus, most IP-stack issues remain confined within the VM itself. So bugs leading to breakout from VM tend to remain rare occurrences.

There are some kinds of attacks against which VM are very effective, e.g. fork bombs. From the point of view of the host system, the VM is a single process. A fork bomb in the guest will bring to its knees the scheduler in the guest OS, but for the host this will be totally harmless. Similarly for memory: the VM emulates a physical machine with a given amount of RAM, and will need about that amount of "real" RAM to back it up efficiently. Regardless of what the guest does, the VM will never monopolize more RAM than that. (You still want to limit VM RAM size to, say, at most 1/2 of your physical RAM size, because the extra "real" RAM is handy for disk caching; and the host OS will want to use some, too.)

Taken into account that malware/ransomware is implemented on the file level, the best method of protection would be based on a block level recovery tool.

In addition, the best way to overlook the possibility of loosing your data would be the implementation of the the 3-2-1 rule. Its a quite common safety measure in the data infrastructure. We actually implement it quite often, it is based on the replication of your data between 3 nodes as well as creating 2 real-time replication copies of the data between the nodes and storing a single copy of your data in a VTL on the cloud.

For any other additional information, I would like to suggest you to take a look at the following article - https://knowledgebase.starwindsoftware.com/explanation/the-3-2-1-backup-rule/

An additional benefit of connecting a local drive would be the option of creating a shared storage replica. There is also an option of connecting a device called AcloudA, it would allow you to create a volume that would be connected directly to the cloud. Take a look here for more details - http://aclouda.com/

StarWind Virtual SAN is something that would allow achieving your goal. We can synchronously replicate the storage across 2 sites thus provide you with the ability to build geo-cluster. Then instead of upgrading your SQL to the enterprise edition, you could simply go with the FCI approach that is available in the standard edition.

@openit

You can also address them directly to me in PM here. Also, I have PMed you my email in case you prefer that way of comunication

@Tim_G

I would go RAID 5 on SSD when i am short on budget, but i need SSD performance.
In case I am capable of going SSD RAID 10, i will

I would recommend looking at what StarWind Virtual SAN can do for you. This is a Windows native software that can be installed directly to the host and does not need to be an appliance, but still can be inside a VM if needed. Also with our software you can build a 2 node HyperConverged cluster (Windows/VMware/Xen) with highly available storage pull that supports SMB-direct and iSER. You can also take a look at HP VSA.

You could build a 3 node HyperConverged Cluster using old servers with StarWind Virtual SAN as a storage provider and Free Hyper-V as the Hypervisor. In case you decide going that path, ping me offline and i will help you out with the configuration.

@DustinB3403

@DustinB3403 said in Replacing the Dead IPOD, SAN Bit the Dust:

@dafyre said in Replacing the Dead IPOD, SAN Bit the Dust:

Why not RLS a la StarWind ?

Because the hosts aren't uniform.

In case they need a SAN replacement, they can consider going for StarWind Storage Appliance. That solution will give them a synchronous HA storage pull that can be shared via iSCSI to their existing hosts. StarWind crew will get those onsite fully preconfigured and will help with designing the migration plan as well as executing it.

@ardeyn Thank you very much for mentioning StarWind.

StarWind HCA should fit your needs perfectly. It is based on Dell hardware and will provide you VMs with pure 100% synchronous HA. This solution is also fully supported by a single vendor and comes to your site preconfigured.
We are also Dell OEM partner, thus less expensive than our competitors.