Memcached Clusters?

Moritz Maisel maisel at sipgate.de
Tue Oct 24 07:41:17 UTC 2006


On Oct 24, 2006, at 07:54 AM, Artur Bergman wrote:
> It is a cache, why do you need to retain the data?

I think there is another aspect making me want to have synced servers: 
It would avoid race conditions with servers goining up and down. 
If there were 3 Servers X, Y and Z. A client uses cache Y. After that Y goes 
down. Now the client takes another cache. When B comes up again it is again 
used by the client and we get outdated data ...

The big question is, how synchronisation would affect performance of the 
cache. So another way to get those race conditions under control without that 
lack of performance would be a flush_all() in any case a cache-server joins 
the cluster. The problem is that as far as I know the client has no 
information about those events (at least in the perl client lib).

Any ideas?

Moritz


More information about the memcached mailing list