Add cluster connection support to redis (using jedis client)

[yuriyz]

Add cluster connection support to redis (using jedis client)

[yuriyz]

This works in 4.0 and in 3.1.6. If jedis get redirection exception (e.g. master goes down and we got promotion of next server from slave to master) then jedis automatically re-new cluster slots. All this work out of the box (see screenshot above attached).

cc @mogluu

[moabu]

If all servers are introduced redis1.gluu.org:6379,redis2.gluu.org:6379,redis3.gluu.org:6379 we get this error. However, we can connect to all servers via the consul. So the issue is not in redis.

This is the oxCacheConfiguration during testing three redis servers:
oxCacheConfiguration

{"cacheProviderType":"REDIS","memcachedConfiguration":{"servers":"localhost:11211","maxOperationQueueLength":100000,"bufferSize":32768,"defaultPutExpiration":60,"connectionFactoryType":"DEFAULT"},"inMemoryConfiguration":{"defaultPutExpiration":60},"redisConfiguration":{"redisProviderType":"CLUSTER","servers":"redis1.gluu.org:6379,redis2.gluu.org:6379,redis3.gluu.org:6379","defaultPutExpiration":60,"password":null,"decryptedPassword":null,"useSSL":false,"sslTrustStoreFilePath":""},"nativePersistenceConfiguration":{"defaultPutExpiration":60,"defaultCleanupBatchSize":25}}

root@localhost:~# tail -f /opt/gluu/jetty/identity/logs/oxtrust.log
2019-04-13 11:16:50,054 INFO  [main] [org.gluu.oxtrust.ldap.service.TemplateService] (TemplateService.java:102) - file.resource.loader.path = /opt/gluu/jetty/identity/conf/shibboleth3/idp, /opt/gluu/jetty/identity/conf/shibboleth3/sp, /opt/gluu/jetty/identity/conf/ldif, /opt/gluu/jetty/identity/conf/shibboleth3/idp/MetadataFilter, /opt/gluu/jetty/identity/conf/shibboleth3/idp/ProfileConfiguration, /opt/gluu/jetty/identity/conf/template/conf, /opt/gluu/jetty/identity/conf/template/shibboleth3
2019-04-13 11:16:50,156 WARN  [main] [org.gluu.oxtrust.ldap.service.SubversionService] (SubversionService.java:191) - The service which commit configuration files into SVN was disabled
2019-04-13 11:16:51,486 INFO  [main] [org.gluu.oxtrust.ldap.service.ShibbolethInitializer] (ShibbolethInitializer.java:41) - IDP config generation is set to true
2019-04-13 11:16:51,524 INFO  [main] [org.gluu.oxtrust.ldap.service.ShibbolethInitializer] (ShibbolethInitializer.java:87) - ########## shibbolethVersion = v3
2019-04-13 11:16:51,524 INFO  [main] [org.gluu.oxtrust.ldap.service.Shibboleth3ConfService] (Shibboleth3ConfService.java:1325) - >>>>>>>>>> IN Shibboleth3ConfService.generateMetadataFiles()...
2019-04-13 11:16:51,568 INFO  [main] [org.gluu.oxtrust.ldap.service.Shibboleth3ConfService] (Shibboleth3ConfService.java:1391) - >>>>>>>>>> LEAVING Shibboleth3ConfService.generateMetadataFiles()...
2019-04-13 11:16:51,569 INFO  [main] [org.gluu.oxtrust.ldap.service.Shibboleth3ConfService] (Shibboleth3ConfService.java:180) - >>>>>>>>>> IN Shibboleth3ConfService.generateConfigurationFiles()...
2019-04-13 11:16:51,585 INFO  [main] [org.gluu.oxtrust.ldap.service.ApplicationFactory] (ApplicationFactory.java:58) - Cache configuration: CacheConfiguration{cacheProviderType=REDIS, memcachedConfiguration=MemcachedConfiguration{servers='localhost:11211', maxOperationQueueLength=100000, bufferSize=32768, defaultPutExpiration=60, connectionFactoryType=DEFAULT}, redisConfiguration=RedisConfiguration{servers='redis1.gluu.org:6379,redis2.gluu.org:6379,redis3.gluu.org:6379', defaultPutExpiration=60, redisProviderType=CLUSTER, useSSL=false, sslTrustStoreFilePath=}, inMemoryConfiguration=InMemoryConfiguration{defaultPutExpiration=60}, nativePersistenceConfiguration=NativePersistenceConfiguration [defaultPutExpiration=60, defaultCleanupBatchSize=25, baseDn=o=gluu]}
2019-04-13 11:16:51,617 ERROR [main] [org.xdi.service.cache.RedisClusterProvider] (RedisClusterProvider.java:44) - Failed to start RedisClusterProvider.
2019-04-13 11:16:51,617 ERROR [main] [org.xdi.service.cache.RedisProvider] (RedisProvider.java:50) - Failed to start RedisProvider.
^C
root@localhost:~# redis-cli -h redis1.gluu.org
redis1.gluu.org:6379> ping
PONG
redis1.gluu.org:6379>

Even more. If I set server to STANDALONE with one Redis URL 'redis1.gluu.org:6379` we get a connection and cache is working through Redis.

oxCacheConfiguration

{"cacheProviderType":"REDIS","memcachedConfiguration":{"servers":"localhost:11211","maxOperationQueueLength":100000,"bufferSize":32768,"defaultPutExpiration":60,"connectionFactoryType":"DEFAULT"},"inMemoryConfiguration":{"defaultPutExpiration":60},"redisConfiguration":{"redisProviderType":"STANDALONE","servers":"redis1.gluu.org:6379","defaultPutExpiration":60,"password":null,"decryptedPassword":null,"useSSL":false,"sslTrustStoreFilePath":""},"nativePersistenceConfiguration":{"defaultPutExpiration":60,"defaultCleanupBatchSize":25}}

[moabu]

This issue can be closed. Tested in 3.1.6 and 4.0 with Cluster Redis. Sentinel was causing conflict when Jedis was trying to connect to redis cluster. I will update the proposed UMA for HA Redis cache configuration to accommodate the change.