Storage pool capacity management
Embodiments relate to a pool of persistent storage volumes. Capacity of the volumes is managed to ensure continued operation and function of the volumes with respect to their corresponding storage pool capacity threshold(s). A background process groups the volumes within each pool into one or more clusters based on a similarity or commonality characteristic. The background process selectively performs one or more space savings techniques of a copy of a selected volume, measures impact data associated with the techniques, and maintains the data. At such time as the threshold level is approached for a storage pool, one or more space reduction actions take place on constituent volumes in view of the background process and the associated groups. The reduction action includes implementation of a space saving technique, such as compression and/or thinning, of one or more volumes in one or more clusters in an associated storage pool.
1. A computer implemented method for managing capacity of a storage pool comprising:
decoupling estimation of capacity saving from one or more space reduction actions and actual execution of the one or more space reduction actions, comprising:
processing non-compressed data and maintaining live usage statistics associated with volumes in the storage pool operating at a first state;
maintaining a list of candidate volumes in each storage pool for space reduction, the list associated with corresponding capacity savings and performance measurements;
prioritizing volumes for each storage pool in the list as a function of storage growth projection;
selecting a volume from the storage pool, and performing a first space reduction action on a copy of the selected volume, including:
measuring a capacity change associated with the selected volume, and recording capacity change data in a first location;
measuring performance change from a workload on the selected volume, and recording performance change data in a second location; and
producing a subset of optimal space reduction actions for execution, the actions based on the recorded capacity and performance change data; and
on-demand, executing at least one action from the produced subset of optimal space reduction actions on a corresponding volume in the storage pool in the first state, the execution selectively converting one or more non-compressed volumes in the storage pool to a second state, wherein the second state is selected from the group consisting of: compressed and thinned.
2. The method of claim 1 , further comprising inferring capacity saving and performance degradation for a non-selected volume in the pool, wherein the inference is based on the measurements from the selected volume.
3. The method of claim 1 , further comprising predicting a change in space saving since a prior measurement, the prediction employing an I/O access pattern observed for each volume.
4. The method of claim 1 , wherein measuring performance change from the workload on the selected subset further comprises:
performing an application execution workload on an un-reduced copy of each selected subset;
performing the application execution workload on the copy of each selected subset after performing the first space reduction action; and
comparing the performance of the application workload on the un-reduced copy of each selected subset and the copy of each selected subset after performing the first space reduction action.
5. The method of claim 1 , wherein measuring performance change from the workload on the selected subset includes measuring a performance impact associated with a switch from a non-reducing state to a space reducing state of the selected subset.
6. The method of claim 1 , wherein estimating capacity savings further comprises:
clustering two or more volumes in the storage pool into a cluster based on a similarity protocol, and limiting the capacity and performance measurements to a subset of volumes in the cluster.
7. The method of claim 1 , further comprising predicting a change in storage capacity within the storage pool based on an expected new storage allocation.
8. A computer program product for managing capacity of a storage pool, the computer program product comprising a computer readable storage device having program code embodied therewith, the program code executable by a processing unit to:
decouple estimation of capacity saving from one or more space reduction actions and actual execution of the one or more space reduction actions, comprising program code to:
process non-compressed data and maintain live usage statistics associated with volumes in the storage pool operating at a first state;
maintain a list of candidate volumes in each storage pool for space reduction, the list associated with corresponding capacity savings and performance measurements;
prioritize volumes for each storage pool in the list as a function of storage growth projection;
select a volume from the storage pool and perform a first space reduction action on a copy of the selected volume, including:
measure a capacity change associated with the selected volume, and record capacity change data in a first location;
measure performance change from a switched workload on the selected volume, and record performance change data in a second location; and
produce a subset of optimal space reduction actions for execution, the actions based on the recorded capacity and performance change data; and
on-demand, execute at least one action from the produced subset of optimal space reduction actions on a corresponding volume in the storage pool in the first state, the execution to selectively convert one or more non-compressed volumes in the storage pool to a second state, wherein the second state is selected from the group consisting of: compressed and thinned.
9. The computer program code of claim 8 , further comprising program code to infer capacity saving and performance degradation for a non-selected volume in the pool, wherein the inference is based on the measurements from the selected volume.
10. The computer program code of claim 8 , further comprising program code to:
predict a change in space saving since a prior measurement, the prediction employing an I/O access pattern observed for each volume; and
periodically update the measurements, including invalidating any prior measurement data.
11. A computer system comprising:
a processing unit operatively coupled to memory;
a storage pool, having two or more storage volumes, operatively coupled to the processing unit, wherein non-compressed data is processed and live usage statistics associated with the volumes in the storage pool operating at a first state is maintained;
a tool in communication with the processing unit to manage capacity of the storage pool, including:
a decouple module to estimate capacity saving from one or more space reduction actions and actual execution of the one or more space reduction actions, the decouple module to:
select a volume from the storage pool, and perform a first space reduction action on a copy of the selected volume;
maintain a list of candidate volumes in each storage pool for space reduction, the list associated with corresponding capacity savings and performance measurements;
prioritize volumes for each storage pool in the list as a function of storage growth projection;
measure a capacity change associated with the selected volume, and record capacity change data in a first location;
measure performance change from a workload on the selected volume, and record performance change data in a second location;
a selection module to produce a subset of optimal space reduction actions for execution, the actions based on the recorded capacity and performance change data; and
the selection module to on-demand, execute at least one action on the produced subset of optimal space reduction actions on a corresponding volume in the storage pool in the first state, the execution selectively converting one or more non-compressed volumes in the storage pool to a second state, wherein the second state is selected from the group consisting of: compressed and thinned.
12. The system of claim 11 , further comprising the decouple module to infer capacity saving and performance degradation for a non-selected volume in the pool, wherein the inference is based on the measurements from the selected volume.
13. The system of claim 11 , further comprising the decouple module to predict a change in space saving since a prior measurement, the prediction employing an I/O access pattern observed for each volume.
14. The system of claim 13 , further comprising an update of the measurements, including an invalidation of any prior measurement data.