I am using mongoDB and my simple cluster consists of one primary and two slaves for replication.
One of my slaves runs slow at times and is not able to catch with primary. Since primary maintains a capped collection for replication, hence one of slave machines simply goes out of sync with primary.
My question is how to deal with this situation ( apart from performance tuning of slave machine and increasing the size of capped collection) ?
Do I need to restore the slave with a backup of the primary ?
This question is also linked to ::
Can I add a new slave to the replica set on a running cluster ?
If yes, then how will the newly added node sync with primary ?