avierengel at gatrixx.com
Wed Nov 2 08:53:49 PST 2005
Am Mittwoch, den 02.11.2005, 11:46 +0100 schrieb Elizabeth Mattijsen:
> At 11:34 AM +0100 11/2/05, Andreas Vierengel wrote:
> >Also I must admit that I wrote my own oo-perl-client (which is of course
> >slower as the standard one, because I don't use closures). I didn't like
> >the idea of invalidating key-distribution if one server crashes.
> Do you mean that the "no_rehash" flag didn't do that for you with
If I remember correctly it does nothing for key-requests on the crashed
server. My implementation uses the rest of the available servers to
distribute the keys which were originally destined for the crashed one
for any subsequent request.
Also I added some "features" like chaining several requests together in
one poll(), send(), read() to bring down the number of
systemcalls if setting many key/values at once. It reduces network
latency, thats all.
It's roughly similar to the multi_get, but I solved it on network layer
for set-like commands :)
More information about the memcached