This file is used during active-sync-prepare to receive a snapshot of the whole persistent store which is sent to the standby. Subsequently, the file is used as a swap for the memory spool. To reduce the size, check the size of your page.db at the active and, when it is small, restart the standby to force a new sync which will create a new spool file.
Is this new behaviour? I can't recall seeing this file a few versions ago. Also, what if my page.db isn't small, or what is small, what is not?
Subsequently, how does the HA router benefit from this spool file? The state of the routers is supposed to be in sync, I fail to see the use of this file on the active instance.
We do have one particular queue that sometime holds a lot of messages which are between 2 and 3 MB in size. If this is a sufficient reason for the spool file to grow, should I increase the maximum cache size to a larger value than 5120 KB to prevent file growth or doesn't it work like that?