VMware vSAN on HPE Synergy

[cariebens]

We're refreshing our hardware (currently using 3PAR and HPE blades) and HPE are really pushing vSAN on Synergy. We're also considering using a Nimble array with DL360s (known as dHCI). I'm reading mixed reviews about vSAN, keen to hear your thoughts. Initially I was going to avoid HCI as balancing disks across nodes seems like it would be a pain if we require more storage (even if we spec the nodes with free slots), however with Synergy we can have blades and disk draws in the same frame, so we can have a whole bunch of disks and then logically attach them to individual blades meaning that we don't need to worry about free slots in individual nodes - if we need more disks, we just add them to the draw.

We'd have a 12 node stretched cluster (6 in each DC) and would run it as FTT=2/N+2.

[support_s]

System recommended content:

1. HPE Synergy Installation and Recommended Practices Guide VMware vSAN ReadyNode on HPE Synergy for Gen10 and Gen9

2. Deploying VMware vSphere on HPE Synergy Best Practice Guide

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

[DanCernese]

As an HPE employee, I can share that I work *a*lot* with very large customers who have gone all-in with the D3940 storage for HPE Synergy compute nodes running vSAN.  The performance does require sufficient and stable upstream networking when crossing logical enclosures (for example when designing clusters across fault domains).  Care must be taken, obviously, when specing out the proper SSD drives (i.e. read/write/lifetime/etc) which is not specific to Synergy.

Using the D3940 provides all the speed of local storage with a redundant SAS fabric (and when configured with dual I/O cards and proper dual ported drives), full redundancy of storage.. ..and a drawer you can slide open during production to pop in more drives.

[Richard75]

Because of the size of your environment VSAN can perfome well and have a good fault tolerants.

But i agree with you on having separete storage solution, in mine opinion it is a more robust solution. in the all integrated solution the dependency of fw/driver between all component (escpecially network and hdd) is more fragile. if you have maintenance you have one hour for a server, otherwise it starts to move data. bet practice is to keep all diskgroups in vsan cluster for all nodes the same. consider if you have to expand your storage and no compute