[syslog-ng] [RFC]: afsql improvement plans

[Gergely Nagy]

Martin Holste <mcholste at gmail.com> writes:

>> I'll read up on it, and probably go with supporting both LOAD DATA and
>> HandlerSocket (after thinking a bit more about the INSERT variations,
>> the few things I didn't like about LOAD DATA aren't that severe that
>> would make it worth exploring the INSERT options).
>>
> Good points.  LOAD DATA really is the holy grail of INSERT for both
> speed and universality.  As I wrote earlier, just writing to /dev/shm
> means you don't need tmpfs on Linux.

I'll look into that, thanks, along with mysqlimport (already looking at
it :). I can think of a number of ways to bypass the buffer file, but
the fact that I need to find a workaround is what I didn't like at
first.

However, exploring the other options further... this is still the best.

Also had a look at HandlerSocket in the meantime... that's an entirely
different protocol, so it would probably be best to have a mysql-native
(that uses libmysqlclient, and thus, LOAD DATA & whatever else the
library supports), and a mysql-handlersocket type (and then we still
have the original libdbi-using mysql type - fun times! I don't envy the
documentation team now).

> When looking at an implementation, don't forget the functionality that
> comes with "CONCURRENT" so that it doesn't need to lock the table or
> block.  The manual says there is a performance hit for doing this, but
> I can assure you it is unnoticeable.

Oh, good to know, thanks for the hint!

-- 
|8]