Consequences of increase Oplog to large size
-
03-03-2021 - |
문제
I'm trying to add a new member (m*-server: 4.2.7) in my replica set (m*-servers: 4.2.1),
and it can't sync in time, after the initial sync it changes state to RECOVERING and have: (~ 4.3 - 6.31 hrs) behind the primary
; Oplog size: "usedMB" : 50769.34, "timeDiffHours" : 1.19 ( ~0.79 -1.19)
.
I can't optimize queries or reduce size of dbs (~1.9T), or make backup and copy data (there is not enough time),
so think only solution for me is replSetResizeOplog
on all members.
Question: What the consequences can be I if increase size of Oplog larger than Upper Bound: 50GB (WiredTiger Storage Engine) to for example: 400 Gb?
올바른 솔루션이 없습니다
다른 팁
Nothing negative, if you don't count the usage of disk space. Actually you need just make your Primary nodes OpLog bigger and then wait at least that 6 hours to fill it enough. But of course it's good to have same size OpLog on every RS node. I try to keep OpLog size at least 48h. That gives you "time" to solve problems, if those rise up.