[bug 1819957]If a memcache server disappears and then reconnects when multiple memcacheservers are used (specific to the python-memcached based backends) it ispossible that the server will contain stale data. The default is now tosupply theflush_on_reconnect optional argument to the backend. Thismeans that when the service connects to a memcache server, it will flushall cached data in the server. This change only impacts the pooled backendas it is the most likely (with heavy use of greenlet) to be impactedby the problem and is the recommended production configuration.

See the help from python-memcached:

@param flush_on_reconnect: optional flag which prevents a

scenario that can cause stale data to be read: If there’s morethan one memcached server and the connection to one isinterrupted, keys that mapped to that server will getreassigned to another. If the first server comes back, thosekeys will map to it again. If it still has its data, get()scan read stale data that was overwritten on anotherserver. This flag is off by default for backwardscompatibility.

[bug 1743036] Thebackend_argument value(s) forurl when configuring memcache did not properly handle multiple servers. This is because the URL was passed as a string (comma delimited) instead of a list to the memcache library/client. Theurl argument is now special cased and will split the string on a comma so that it mirrors the behavior of the ListOpt used bymemcache_servers option.