Hi Team,

I am looking for the recommendation on volume size in PBHA and Non-PBHA environment. Knowing the fact that that we can create the volume size up to 256 TB.

I am specifically looking a scenario where customer says his data store size on VMWare is 60TB.

So is it OK if we create a single volume of 60TB or we can recommended multiple volume (10*6) on which they can create data store and use, this is in PBHA environment and the PBHA storage 7300 & 7200 storage are within the same site.

I couldn't find any guideline recommendation on volume sizing. 

Could you please provide recommendation here.

Thanks and Regards,

Subodh Jaiswar

------------------------------
Subodh Jaiswar
------------------------------

In a PBHA environment (e.g. using two systems, or within the same site as in your case), I would recommend using multiple volumes rather than a single monolithic 60 TB volume, for these reasons:

• Replication load balancing: Multiple volumes can be spread across volume groups and I/O groups so replication traffic is balanced. This helps avoid bottlenecks or a single hotspot.

• Redundancy / isolation: If one volume or VG hits an issue (e.g. replication, journal resource exhaustion), others may still function.

• Manageability & flexibility: You can grow, migrate, rebalance or tune individual volumes independently.

• Risk containment: Smaller units reduce blast radius in case of any internal problem or failure.

• IBM's own caution: As mentioned, IBM suggests having multiple replicating volume groups per I/O group, rather than relying on a single VG handling everything under PBHA. public.dhe.ibm.com+1

• Avoid single-group bottlenecks: The release notes explicitly mention potential node warmstarts or performance impacts if only one VG is replicating in an I/O group. public.dhe.ibm.com

So in your scenario of ~60 TB, splitting it into multiple volumes (for example, 10 volumes of 6 TB each) is a safer and more flexible approach, especially when the two PBHA nodes are in the same site.

You should also ensure your interconnect (ISL / replication link) bandwidth and latency are sized adequately to handle the aggregated write load, and that the journal / replication resources in the PBHA nodes are sufficient for that many volumes.

Hi Mustapha Gbogboade,

Thank you for your response and recommendation.