[ZODB-Dev] State of persistent cache after commit

[Tim Peters]

[Rad Widmer]
> Yes, thanks for the info. I was off-base with the invalidation theory,
> thanks for clearing that up.

That's OK -- there are a lot of pieces to any persistence system, and it's
not at all obvious how they fit together.

> The default cache_size is the most likely cause of my performance
> problems. I did a quick check with a much larger cache_size (50,000) and
> the results look promising.

It's a good use for RAM, if you've got it to spare <wink>.  Note that the
memory cache size is in units of number of objects, not in units of bytes;
the memory cache machinery doesn't know how many bytes an object consumes
(and that's a harder thing to know than you might think!).