Thu, 18 Mar 2004 10:11:46 +0100
this part of the code has changed between 0.11 & 0.12, but your diff is on=
So is this relevant for 0.12 too?
On Tuesday 16 March 2004 03:00, Jason Titus wrote:
> We found an issue connecting to multiple caches on the same host. The
> dead_hosts hash contains both the hostname:port AND just the hostname.
> This means that if one process goes down on a machine, all the
> connections to the other ports will fail (silently).
> Here is a proposed patch that folks may or may not like. I'm not sure
> what the logic was of including the hostname as well as the
> hostname:port combo. Perhaps there is a benefit that I'm not thinking of.
> Let me know what you think. Without this, we have to manually call
> 'Cache::Memcached::forget_dead_hosts()' between each call.
> Also, is there any plan to add in some kind of connection check to
> Cache::Memcached? It is kind of a bummer to only find out that a
> connection isn't working when you try and 'get' or 'set' something. A
> failure on 'new' would make sense if there is no cache at a particular
Phone: +49 (0) 89 490 267 726
=46ax: +49 (0) 89 490 267 701
Mobile: +49 (0) 179 674 3611
mailto: firstname.lastname@example.org =A0 =A0 =A0 =A0=20