I am running repair with -pr option. The issue which we are facing that when we trigger repair disk space usage reaches 100% on some nodes. I know running subrange repair is one option but we have not integrated reaper as of now. Is there any other way we can avoid this scenario?
Asked
Active
Viewed 155 times
-1
-
You can run it on a single table at a time, in case not doing it already. – Payal Nov 02 '20 at 09:00
-
Yes we are running one table at a time. Looks like subrange repair is the only way out – Manish Khandelwal Nov 02 '20 at 11:45
1 Answers
1
Good practice is to always have up to 50% disk utilization in case scenario like this happens.
Your best option would be to make sure you have less than 50% disk utilization before you trigger repair. You can check for any snapshots and remove them after taking backup.
k0ner
- 1,086
- 7
- 20