How memory usage is represented in the unix free command

[Tim Haines]

Hi there,

I have one server with 8GB of RAM with 2 dbs, each holding a collection.  Stats shows one collection has a 2G storage size and a total index size of 2G.  The other collection has a storage size of 1G, and a total index size of 0.5G.  So, a total of 5.5G of data + index.

When I run free -m I see:

             total       used       free     shared    buffers     cached

Mem:          7958       6459       1499          0        167       5659

-/+ buffers/cache:        631       7327

Swap:          975          0        975

I'm surprised by 2 things.  1) That the cache is larger than the memory reported as being used.  and 2) that there's 7.3G of cache reported as being free, and only 631M reported as being used, when I have 5.5G of data+ index.

How does the memory being used by Mongo get represented in the free output?  Is it a mixture of the free and used cache?

Cheers,

Tim.

[David Hows]

Hi Tim,

This is to do with the way that Linux keeps track of RAM.

Its likely that the pages in RAM in which those datafiles reside in have not been touched in a while and as such they have been added to the list of free pages.

Cheers,

David