Hi I fired up mongostat
to see current usage and thinking to optimise hardware usage. Basically I have something like
insert query update delete getmore command % dirty % used flushes vsize res qr|qw ar|aw netIn netOut conn set repl time
*0 *0 *0 *0 1 5|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 942b 19k 27 mongodb0 PRI 2020-07-12T06:36:23Z
*0 3 *0 *0 1 5|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 2k 43k 28 mongodb0 PRI 2020-07-12T06:36:24Z
*0 *0 *0 *0 1 1|0 0.0 27.7 1 4.8G 3.9G 0|0 0|0 180b 18k 28 mongodb0 PRI 2020-07-12T06:36:25Z
*0 *0 *0 *0 1 5|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 400b 20k 28 mongodb0 PRI 2020-07-12T06:36:26Z
*0 *0 *0 *0 1 3|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 290b 19k 28 mongodb0 PRI 2020-07-12T06:36:27Z
*0 3 *0 *0 1 1|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 2k 29k 28 mongodb0 PRI 2020-07-12T06:36:28Z
*0 *0 *0 *0 1 2|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 235b 19k 28 mongodb0 PRI 2020-07-12T06:36:29Z
*0 6 *0 *0 1 1|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 1k 38k 28 mongodb0 PRI 2020-07-12T06:36:30Z
*0 *0 *0 *0 1 5|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 400b 20k 28 mongodb0 PRI 2020-07-12T06:36:31Z
*0 *0 *0 *0 1 8|0 0.0 27.7 0 4.8G 3.9G 0|0 0|0 854b 20k 28 mongodb0 PRI 2020-07-12T06:36:32Z
Looking at vsize and res are they separate memory usage or a subset? eg total memory used 8.7G vs res used 3.9G out of vsize’s 4.8G
With that number would you say having 2 or 3x8GB machines with an arbiter in replication make sense? Or just go with 2x16GB? Thinking to run on Google Cloud where they also have 13GB with 2 vCPUs.
Keen to hear any feedback. Cheers.