We have a sharded mongo cluster, and a set of application servers, each with their own mongos process. Each of the app servers has 4GB ram, and we are seeing over 2GB usage from mongos, when the servers are idle. This looks like some kind of memory leak to me. I found a previous thread talking about similar symptoms, but that was suppose to be fixed I believe.
Could anyone shed any light on this?
To be clear, this is after 30 days uptime, so is accumulating relatively slowly, but still.
댓글 없음:
댓글 쓰기