0

I have a MongoDB replica set. I periodically put a lot of data into the master (like 1T); every time, all the secondary nodes enter "recovering" status and never come back.

I already increased the binlog size to a very large value; is there any panacea configuration that can solve this problem?

Vince Bowdren
  • 8,326
  • 3
  • 31
  • 56
billtian
  • 6,589
  • 4
  • 18
  • 28
  • Can I ask two points of clarification; when you say _binlog_, do you mean the [oplog](https://docs.mongodb.com/manual/core/replica-set-oplog/)? Secondly, have you checked the secondary servers' logs to see exactly why they say they're entering 'recovering' status? – Vince Bowdren May 16 '17 at 21:27
  • Yes, I mean oplog, sorry for the ambiguity. The secondary nodes log says "we are too stale to use xxxx as a sync source". It just looks like that the rsync is not fast enough when we pushing the large amount of data into the primary. Not sure if there is a configuration like delay sync to solve this problem. – billtian May 17 '17 at 22:23

0 Answers0