http://www.slony.info/bugzilla/show_bug.cgi?id=235

--- Comment #2 from Christopher Browne <cbbrowne at ca.afilias.info> 2011-08-23 14:50:30 PDT ---
It seems to me that in HEAD, we should look at revising the logic in a more
substantial way.

The logic that was added around 1.2 has the conceit that we imagine that SYNCs
are likely to 'typically' take some period of time to process, from whence
comes the configuration option "desired_sync_time".

This is fairly nonsensical, as how long a SYNC will take to process really
depends on how much data got committed between two SYNCs, which isn't so
predictable.

The logic might be simplified to something more like:

- We define a "max" number of syncs to process, and don't ever consider doing
more than that.

- If there are fewer than that "max", then group whatever is available.

There's some logic surrounding increasing towards the "max"; it's pretty likely
that trying to do this is just nonsense, and we should just head straight to
the configured maximum.

If the administrator feels that only a few SYNCs should be processed together,
then they can set sync_group_maxsize fairly small, and get the desired result.

If they want lots of grouping, then they can set it high, and get the desired
result.

-- 
Configure bugmail: http://www.slony.info/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
You are the assignee for the bug.