I have a RAID 5 setup with almost 15TB of data capacity.
In the past I have let it go unchecked for too long, and found that multiple drives had SMART errors and I want to prevent bad blocks. I am implementing a weekly scrub but I am wondering what will happen if an rsync timed schedule starts while the drives are scrubbing. Will the rsync just fail until the scrub is completed? Or will there be a multithreaded system where both of these operations will happen at once? Or worst case -- would it interrupt the scrubbing process so it would have to start over.
My schedules are as follows:
Daily job from 4 different local computers rsync to my local NAS at different times. 3:00am, 4:00am, 5:00am, and 6:00am.
Daily job from my local NAS to a cloud server at 12:00pm.
Weekly job Sunday at 3:00pm to scrub the raid.
Surely, a scrub can take enough time that the other jobs would start while the raid is still scrubbing. Basically, I'm not sure what the implications are of writing new data to the drives as they are being scrubbed....
What are the best practices for handling this, with the priority being that the RAID gets scrubbed once a week and backups happen around that?
Thank you
Normally, scrubbing runs at low priority and only when the logical drive is idle. Basically, access while scrubbing only slows down the latter. Note that infrequent access can experience higher latency due to the drives being busy during scrubbing.
You should schedule scrubbing for periods when you expect the array to be idle, ie. put it outside your backup window.