Les Hazlewood-2 wrote > > Hi Jason, > > In your local test environment, if you disable the > RMICacheReplicatorFactory - e.g have Ehcache on a single box, does > your test case still fail? > > In other words, its great that you have a repeatable test. Does the > test fail in any scenario other than the RMICacheReplicatorFactory? > Is there another type of CacheReplicatorFactory that could be swapped > out for testing? E.g. LocalCacheReplicatorFactory that uses > in-process memory? > >
Unfortunately, I don't really have a repeatable test. In my test environment with one application server and some stakeholders hitting it during the day and QA at night, it will happen once or twice a day, so you don't have to wait a long time to see it, but I can't create the problem at will. That's why I was asking if anyone was really using this particular cache replicator in production. If people are using it without problems, it's clearly my problem. If no one is, then it's less clear. If many have tried and failed, that doesn't prove it's a problem with the cache replication but it's a prime suspect. If no one is doing it in production, it's less clear, but still seems fishy for the RMI replication. Incidentally, yesterday, working on the assumption that the problem is with "false misses" in the cache, I implemented an alternative to org.apache.shiro.session.mgt.eis.EnterpriseCacheSessionDAO, which also extends CachingSessionDAO, but also writes sessions to the database and reads from the database in the event of a cache miss. My sessions are quite small, but this still might not be a good solution. However, if the problem goes away, then I've definitely implicated the cache and not some other thing that's causing my sessions to be invalidated. -- View this message in context: http://shiro-user.582556.n2.nabble.com/Anyone-using-RMI-cache-replicator-in-production-tp7172831p7177564.html Sent from the Shiro User mailing list archive at Nabble.com.
