Hello all, I'm fairly new to memcached, and the documentation I've been reading on its use for Drupal, along with searching for answers in this group gives me enough to know that it is not normal for sessions to spin on the browser until they timeout when memcached is stopped.
The architecture is a cluster of two instances. Initially views would no longer work as soon as one of the instances was brought down. I found that we were not using unique key prefixes and corrected that. Now, whenever one memcached instances is stopped, things still work fine. However, once I bring both memcached instances down, it doesn't treat hits as a cache miss. It doesn't appear to even go to the database for data, but just times out after the four minute period we've set for idleTimeout in PHP. I presume this means that I am doing something fundamentally wrong here. I am not using any kind of custom setting in memcached other than just specifying a memory limit and the user to run it as. Do you think this means that I should look at the Drupal application itself for clues on why it behaves in this way, or am I missing something on the memcache front? Thanks for any help.
