"An existing connection was forcibly closed by the remote host" after restarting Redis about nhibernate.caches.redis HOT 9 CLOSED

Hey - thanks for bringing this up! I was planning on fixing this at some point, but you convinced me to fix it right now! 😄

from nhibernate.caches.redis.

Cool, thank you for such a fast fix! :-)

from nhibernate.caches.redis.

I did some tests and this exception is indeed swallowed now so Cache can operate without issues even when connection is broken but even when Redis is back it does not seem to recover and keeps throwing this exception. Is it some issue with ServiceStack.Redis maybe ?

from nhibernate.caches.redis.

Hmm - I have an integration test that should cover this (I just updated it to be more robust) and it's still passing for me. Could you try:

1. Run the sample.
2. Add a couple of posts.
3. Kill Redis.
4. Refresh

What happens here when running the sample? It resumes fine for me (retrieving the posts from the cache after it's been restored). Is there a different exception thrown? If so, could you paste it here?

Thanks!

from nhibernate.caches.redis.

Yeah that's a weird one.

My scenario is like this:

• I'm using PooledRedisClientManager
• When Redis server goes down and back again (simply restarted or is switched to slave backup server) cache will stop working. It is caused because RedisNativeClient is throwing exceptions that are swallowed by your Cache Provider (good thing). Problem is that these connections will never recover and this exception will happen constantly.

Exception is System.Net.Sockets.SocketException: An existing connection was forcibly closed by the remote host

at System.Net.Sockets.Socket.Send(IList1 buffers) at ServiceStack.Redis.RedisNativeClient.FlushSendBuffer() at ServiceStack.Redis.RedisTransaction.Exec() at ServiceStack.Redis.RedisTransaction.Commit() at NHibernate.Caches.Redis.RedisCache.ExecuteEnsureGeneration(Action1 action) in c:\DEV\NHibernate.Caches.Redis\src\NHibernate.Caches.Redis\RedisCache.cs:line 315
at NHibernate.Caches.Redis.RedisCache.Get(Object key) in c:\DEV\NHibernate.Caches.Redis\src\NHibernate.Caches.Redis\RedisCache.cs:line 156

I'm using yours RedisSessionProvider that seems to behave in a way that only first connection after Redis restart is throwing an exception and subsequent are ok.

from nhibernate.caches.redis.

I've found something out just now.

As I'm using both Cache and Session providers I was creating two PooledRedisClientManagers (as Cache and Sessions can be stored on different servers - they are not currently though). Just now I've tried to reuse the same PooledRedisClientManager and there seem to be no problem ?!?

I guess it's all down to ServiceStack.Redis not your provider really...

I will ask ServiceStack ppl than, thanks for your help!

from nhibernate.caches.redis.

Yeah - I'm wondering if the two client managers are sharing state...?

from nhibernate.caches.redis.

Check out #8 where I've switched to StackExchange.Redis... hopefully all Redis-related problems can go away.

from nhibernate.caches.redis.

@kowalgta I'm going to close this due to lack of activity. Also, I just pushed the 2.0 release which now uses StackExchange.Redis as the Redis provider. Please give it a shot and let me know if you have any problems.

from nhibernate.caches.redis.