CONTAP-588612: Automatic workload rebalancing does not start after upgrading to 9.17.1
Issue
- NetApp ASAr2 platform runs disaggregated ONTAP 9.16.1.
- Due to overprovisioned physical space, one or more of the data flexaggregates is running out of space, despite other data flexaggregates have free-space left.
- However, copy-based placement rebalancing is not happening here, because there is not enough free-space left within the destination flexaggregated, for a copy based volume move:
MyCluster::> aggr show
Aggregate Size Available Used% State #Vols Nodes RAID Status
--------- -------- --------- ----- ------- ------ ---------------- ------------
dataFA_2_p0_i1 59.34TB 742.5GB 99% online 2 MyCluster-01 -
dataFA_4_p0_i1 59.34TB 26.05TB 56% online 1 MyCluster-02 -::*> lun show -fields physical-used
vserver path physical-used
------- ---------------------- -------------
svm1 LUN1_IN_dataFA_2_p0_i1 25.26TB
svm1 LUN2-IN_dataFA_2_p0_i1 30.75TB
svm1 LUN1_IN_dataFA_4_p0_i1 31.44TB
Note: A copy-based move of the 25.26TB physical-used storage unit "LUN1_IN_dataFA_2_p0_i1" from 99% capacity used "dataFA_2_p0_i1" towards "dataFA_4_p0_i1" with 26.05 free-space left, will not work. The remaining freespace would be not enough and ONTAP would block even a manual move attempt.
- Upgrading to 9.17.1 will allow to utilize the newly introduced automatic workload rebalancing, which includes copy-free space rebalancing.
- This automatic workload rebalancing can adjust the capacity between flexaggregates without moving objects like storage units.
- However, automatic copy-free workload rebalancing introduced in 9.17.1, will only get enabled 14days after upgrading to 9.17.1 by design (warmup period for learning space usage over time).
- The cluster is upgraded to 9.17.1.
- After the upgrade, the capacity is not automatically rebalanced based on the new copy-less method introduced in 9.17.1.
