For stores where there is a very high rate of sets compared to gets, it
is beneficial to retrieve from the local server, skipping the network
overhead, at the expense of having to set many severs when a key is set.
This patch adds a memcache option to enable clustering. When on, only
one "server" is allowed, which will be where fetches are from, while
sets/updates/deletes/purges will occur to the all the servers in the
"set server" list.
To run unit tests, define TEST_CACHESTORE_MEMCACHE_TESTSERVERS with
multiple (return delimited) servers.
For stores where there is a very high rate of sets compared to gets, it
is beneficial to retrieve from the local server, skipping the network
overhead, at the expense of having to set many severs when a key is set.
This patch adds a memcached option to enable clustering. When on, only
one "server" is allowed, which will be where fetches are from, while
sets/updates/deletes/purges will occur to the all the servers in the
"set server" list.
To run unit tests, define TEST_CACHESTORE_MEMCACHED_TESTSERVERS with
multiple (return delimited) servers.
This issue makes several fixes for the session loader and the session store.
* maxsize argument now works for session caches.
* fixed performance hole when interation occurs frequently.
* fixed cache purge bug occuring when multiple caches are defined before being used.
* improved lastaccess handling.
Big thanks to Marina who contributed the following commits:
* Always make sure the elements in cache are sorted so we need to remove only elements in the beginning of array
* Remove expired elements from session store to free memory
* Minor bug fixes
MDL-38162 fixes the issue about testing the connection to the memcached
daemon, which means testing once the connection to the pool of servers
and not to each of the servers while being added to the pool.
It misses to move the comment to that code too.
Various reasons can lead to the cache stores not being available
(missing extension, configuration...). Under those situations
the test must be skipped because for sure it will break if
we continue.
isset is much faster than array_key_exists. For caches we need to
ensure as much performance as possible. To all null to continue to
be stored and provide correct results, non ttl also now saves in
$this->store[$key][0]. As isset($this->store[$key]) will return
true in all cases.
time() as found in cache::now() can be slow on some platforms and
we shouldn't bother checking it unless needed. It also stops non-ttl
caches paying the ttl price.
Unit tests were adjusted to ensure null is still handled correctly
when we are only using isset.
By holding onto a reference of the global $CFG object we
can be sure that it isn't destroyed before we are done
with it.
This allows for caches to be used within destructors for
purposes such as delayed writes.
