Table timeout/lifetime Not Reflected in -count

 

1. Less than a handful. It never crossed my mind that we would experience different behavior under load.

 

 

2/3. The counts are from the same table, using two (or three) different methods to arrive at the value:

 

 

a. set session_count_native [table keys -subtable $table_subname -count]

 

b. set session_count_length [llength [table keys -subtable $table_subname]]

 

c. set session_keys [table keys -subtable $table_subname] >> and loop over the keys incrementing a counter

 

 

In the example above "119 v 16355" equates to [b,c] v [a].

 

 

 

Sounds like our best option is to upgrade to the latest/greatest version. Are there any known issues/incompatibilities between the two versions?

 

 

 

Thanks again Spark for the amazing support!