Homelab Neophyte Obtains a 3PAR 8200, Don't Know What to Do.

[EndlessFailover]

This forum has been very helpful when I first got the 3PAR. As a pinout diagram here helped me DIY a serial console cable and let me successfully get access to the 3PAR 8200 controller nodes. After which has been smooth sailing installing the service processor and resetting the nodes.

My plan is to use RAIDZ to create volumes with redundant drives underneath, and serve them via some standard, whether it be minio S3, NFS, or even SMB to the VMs, so application data can be stored in a resilient location.

As my homelab does not have a fiber channel switch, the plan is to connect the 3PAR nodes via iSCSI to a VM that can create RAIDZ drives, and then serve them to the VMs, however, I do not know if this is feasible on a 3PAR system.

I don't know if it is possible to expose the drives in "HBA" mode to a managing VM via iSCSI. The CPU and RAM inside the controller nodes make me feel like they are the ones in charge aggregating the drives before exposing them to a host.

If RAIDZ is not possible with the 3PAR system, what is the usual way it is setup by a system administrator in a data center via iSCSI? I really hope this won't prove to be too difficult to the point that I will have to let go of this very nice piece of enterprise equipment with potential to do a 4 way failover.

[MammaGutt]

This forum has been very helpful when I first got the 3PAR. As a pinout diagram here helped me DIY a serial console cable and let me successfully get access to the 3PAR 8200 controller nodes. After which has been smooth sailing installing the service processor and resetting the nodes.

My plan is to use RAIDZ to create volumes with redundant drives underneath, and serve them via some standard, whether it be minio S3, NFS, or even SMB to the VMs, so application data can be stored in a resilient location.

As my homelab does not have a fiber channel switch, the plan is to connect the 3PAR nodes via iSCSI to a VM that can create RAIDZ drives, and then serve them to the VMs, however, I do not know if this is feasible on a 3PAR system.

I don't know if it is possible to expose the drives in "HBA" mode to a managing VM via iSCSI. The CPU and RAM inside the controller nodes make me feel like they are the ones in charge aggregating the drives before exposing them to a host.

If RAIDZ is not possible with the 3PAR system, what is the usual way it is setup by a system administrator in a data center via iSCSI? I really hope this won't prove to be too difficult to the point that I will have to let go of this very nice piece of enterprise equipment with potential to do a 4 way failover.

Where to start.

You cannot expose the drives in HBA modes thru the controllers. If you have external cages you may or may not be successful in accessing those drives directly without a controller shelf, but the intention is not doing this. Also keep in mind that the drives used in 3PAR systems have 520byte sectors (for T10 dif) so you would need to do a low-level format if you were to take these drives somewhere else.

So the next questions if probably, what type of HBAs are installed in your system? Most systems were sold with FC only. If you don't have the dedicated iSCSI HBAs, you can't use the system for iSCSI. It would probably be cheaper for you (assuming low number of hosts in your home lab) to get FC HBAs for your servers and connecting them directly to the 3PAR nodes without switches vs getting hold of those iSCSI HBAs for 3PAR 8000.

The system supports RAID1, RAID5 and RAID6 and it will present a block device to your hosts.

[EndlessFailover]

So the next questions if probably, what type of HBAs are installed in your system? Most systems were sold with FC only. If you don't have the dedicated iSCSI HBAs, you can't use the system for iSCSI. It would probably be cheaper for you (assuming low number of hosts in your home lab) to get FC HBAs for your servers and connecting them directly to the 3PAR nodes without switches vs getting hold of those iSCSI HBAs for 3PAR 8000.

I was very lucky, the controller nodes have both the iSCSI SFP+ slots populated and FC slots populated, so both protocols are available to me. In this case, is RAIDZ supported / recommended?

Looking at this from another angle, if 3PAR was designed with RAID6 in mind, that probably means a hardware RAID6 config has been battle tested in production environment to provide enough protection for an array, and RAIDZ might not be necessary if I have access to hardware RAID6.

One more question I have in mind is expansion, is populating more slots in the enclosure, and adding an enclosure directly after the support contract expired still possible? Or would I be locked in the current drives configuration if the support contract were to terminate?

[MammaGutt]

So the next questions if probably, what type of HBAs are installed in your system? Most systems were sold with FC only. If you don't have the dedicated iSCSI HBAs, you can't use the system for iSCSI. It would probably be cheaper for you (assuming low number of hosts in your home lab) to get FC HBAs for your servers and connecting them directly to the 3PAR nodes without switches vs getting hold of those iSCSI HBAs for 3PAR 8000.

I was very lucky, the controller nodes have both the iSCSI SFP+ slots populated and FC slots populated, so both protocols are available to me. In this case, is RAIDZ supported / recommended?

Looking at this from another angle, if 3PAR was designed with RAID6 in mind, that probably means a hardware RAID6 config has been battle tested in production environment to provide enough protection for an array, and RAIDZ might not be necessary if I have access to hardware RAID6.

One more question I have in mind is expansion, is populating more slots in the enclosure, and adding an enclosure directly after the support contract expired still possible? Or would I be locked in the current drives configuration if the support contract were to terminate?

As stated, RAID1, RAID5 or RAID6.

It is hardware accelerated with dedicated ASICs.

Look at your license if it is drive limited or not.