"local" directory automatically purged on master node

Folder pointed to by local property in dremio conf was clean/deleted.

2020-07-25 10:41:56,665 [scheduler-38] INFO c.d.e.catalog.SourceMetadataManager - Source ‘"__jobResultsStore"’ refreshed names in 0 seconds. Details:
Shallow probed 0 datasets: 0 added, 0 unchanged, 0 deleted
….
2020-07-25 11:31:56,619 [scheduler-11] INFO c.d.e.catalog.SourceMetadataManager - Source ‘"__jobResultsStore"’ skipping metadata refresh since it is currently in a bad state of Unavailable: File /mnt/foo/dremio/cache/pdfs/results does not exist.

master/coordinator node:

local: /mnt/foo/dremio/cache
dist: /mnt/foo/dremio/cache/pdfs

Along with above, all folders/files under cache folder above deleted. I seems it was by luck that db folder which was under local path above was locked and did not get the big-hammer purge, even though it should have been 14 day incrementally.

results.max.age_in_days = 14 days coincided with above event from what I could tell. Maybe there is an issue here?

Any ideas?