PowerScale OneFS: CloudPools space not available on cluster after cloud archive

After adding a CloudPools license, configuring a cloud tier and creating a filepool policy to move data to the cloud, the expectation is that the cluster will gain free writable space equivalent to the amount of data moved to the cloud. While this is generally the case, some factors may affect this outcome.

Cloud Job Still Processing

First, verify that the FilePool policy is in the correct order and did identify files for CloudPools to archive. This can be done by looking at the cloud jobs:

# isi cloud jobs list -v

Find the job with the correct "Job Engine Job" ID which corresponds to the job engine ID for the SmartPools job. Then run:

# isi cloud jobs view <ID>

The output here should show a number in "Total Files." If not, then the FilePool policy should be reviewed.

If there is a number there, verify that the job has been run to completion. You are looking for "Total Pending" "Total Failed" "Cancelled" If the job is still processing, the files are still taking up space. 

Snapshots

If the job has completed successfully, then look at snapshots. If there is a snapshot of the directory, and the snapshot is taken prior to cloud archive, the snapshot maintains its pointers to the blocks, so space savings are not realized until the snapshots either expire naturally or are deleted. 
 

Deduplication

If the cluster has had a deduplication policy run against the directory that has been stubbed to cloud, or any parent of the directory, then of the space savings available from these files has already been realized. The full size of the file is written to cloud, and the file is replaced with stub files. However of the file was previously pointers to the shared deduplication tree.

# isi_classic job history -j dedupe

# isi job reports view <job ID>