StarWind Software

Hello community
Thank you for having me here.
I have two questions. I set up starwind vSan at the moment. Unfortunately I haven't access to the GUI because of the free license. New I want to use a dual port mellanox connect x3 with iSER.
Is there a way to activate it with powershell, or maybe over the starwind.conf file?

My second question is: I want to use L2 cache on NVME disks. I already read an how to use L1 and L2 cache. If I understand it right, that the L2 cache will not be synchronised between the nodes? What I don't understand then is, what will happen in a failure of a node,? The cache is then not available. How will that work in a HA environment? Or is my understanding wrong?

Thank you very much for your time and help.
Regarda
Kevin

Hi Kevin,

I have two questions. I set up starwind vSan at the moment. Unfortunately I haven't access to the GUI because of the free license. New I want to use a dual port mellanox connect x3 with iSER.
Is there a way to activate it with powershell, or maybe over the starwind.conf file?

You can check the interfaces with iSER enabled in StarWind.cfg file, <iSerListen value=""/> with the list of IPs on which iSER works.

My second question is: I want to use L2 cache on NVME disks. I already read an how to use L1 and L2 cache. If I understand it right, that the L2 cache will not be synchronised between the nodes? What I don't understand then is, what will happen in a failure of a node,? The cache is then not available. How will that work in a HA environment? Or is my understanding wrong?

Cache is replicated between the nodes, no matters L1 or L2 at all. L2 cache is working only in Write-through mode and this means that you will not receive the full power of NVME, the reads can be only boosted. By the way, your environment read intensive or write intensive? Because you can use NVME disks as the fully functional StarWind device with maximum read/write performance.

Hi

Thank you for your fast reply.
OK, I set here my two IP's. RDMA traffic is viewable in the permormance monitor. It's seems as would it work. Thank you.

OK, it's replicated without specifiy it especially in the _HA.swdsk file?
Mhm, I think the workload depends, I can't say it. I have two Exchange servers, application, firewalls, owncloud, proxy's running on this storage.
How do you mean use the NVME as a fully functional device? Define it directly as a HA device? But then I can't boost all storage operations, only for this servers which are stored on this device... or how would you config this?


Thank you.
Regards Kevin

Kevin,
Also, please check if your NIC supports iSER.

OK, it's replicated without specifiy it especially in the _HA.swdsk file?

You need to specify the size of L2 cache during creating a device and during the process of replica creation.

If you are going to use NVME as the L2 cache, you will boost only read operations and the writes will be the same as the main storage. L2 caching is working only in write-through mode.
You can put some of your performance hungry VMs on StarWind HA you can create on NVME disks.