Re: How to force resync from REPSET primary?

[Chris Cunningham]

Hello,

Setting the rs.syncFrom() only tells the Secondary who to get data from, so running that command is not going to start the initial sync process.

Please be aware that as soon as you created that capped collection on the Primary, it was instantly written to the MongoDB oplog and propagated out to all of the Secondaries in that Replica Set. This means that the oversized capped collection was created immediately on all Secondaries. You can find more information on how the sync process works in the Replication page.

I have a few additional questions for you in regards to what lead to your current status:

• What is your MongoDB Version? In the mongo shell execute the command db.version() to find the version you are running.
• Is the subscriptionproductdb database where you created this capped collection?
• What size are you expecting the subscriptionproductdb database to be now?
• Have you deleted the large capped collection you created by accident and re-created it with the correct size?
• Can you please provide some additional details on how you tried to delete the large capped collection you created? Was it through the mongo shell or did you use some other method?
• When you deleted the files from the filesystem, which node did you do this on?
• When you started that node back up, did you execute the command rs.initiate() before joining it to the existing Replica Set?
• Log into each node of the Replica Set and run rs.status() and verify that each member of the set have the same content.

Additionally, I noticed that the username that was used to connect to the mongod is mmsuser. Is there a reason for having this username? Are you currently running MongoDB Cloud Manager?

Thanks,

Chris

​