Overview Scrubbing can be a resource intensive operation but is an important part of detecting bitrot in a cluster. Currently, we support (in Stockwell), the ability to schedule when scrubbing may happen (eg between 10pm and 2am). However the period specified may not provide enough time to perform the scrub. The goal is to ensure that scrubbing is performed on ongoing basis but with intelligent scheduling. Feature: For newly created pools/PGs then scrubbing all happens at the same time. Randomize scrub scheduling so IO impact is reduced. (Merged in Infernalis)
Is this merged in Jewel? I checked the release notes and the git log but couldn't figure it out.
I see the configuration variable "osd_scrub_interval_randomize_ratio" in jewel. So I'd say it is there. The description above mentions that it was merged in Infernalis.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2016:1755