topic Cluster content inconsistencies in Alfresco Archive https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307212#M260342 <HTML><HEAD></HEAD><BODY><SPAN>Hello,</SPAN><BR /><BR /><SPAN>We have a 2 node cluster of Alfresco.&nbsp; We have content being pushed to the repository via WebDAV.&nbsp; We find that every once in a while, the opposite node does not have the updated file.&nbsp; The database has the updated information but, the object (from Hibernate, I presuem) is not updated.&nbsp; So, when any code accesses the file on node1 it's correct.&nbsp; But, when you access it on node2, it is inaccurate. The servers have to restart in order for node2 to reflect the latest changes. </SPAN><BR /><BR /><SPAN>We increase the rmi socket timeout which seems to make a small improvement however, it's still happening and the timeout is LONG(hours).</SPAN><BR /><SPAN>alfresco.ehcache.rmi.sockettimeoutmillis</SPAN><BR /><BR /><SPAN>We are using TCP for jgroups and I've increased logging but, it's like finding a needle in a haystack.</SPAN><BR /><BR /><SPAN>However, I also realize that the ehcache objects are configured to never expire.&nbsp; Our ehcache_custom.xml has entries like this:</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;cache</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; name="org.alfresco.repo.domain.hibernate.NodeImpl"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; maxElementsInMemory="10000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; eternal="false"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; timeToIdleSeconds="900"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; timeToLiveSeconds="900"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; overflowToDisk="false"&gt;</SPAN><BR /><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &lt;cacheEventListenerFactory</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; class="net.sf.ehcache.distribution.RMICacheReplicatorFactory"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; properties="replicatePuts = false,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; replicateUpdates = true,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; replicateRemovals = true,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; replicateUpdatesViaCopy = false,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; replicateAsynchronously = false"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;/cache&gt;</SPAN><BR /><BR /><SPAN>I'm concerned about expiring and potentially moving a load to the database - not desireable.</SPAN><BR /><BR /><SPAN>Does anybody have any advice to help improve the situation?</SPAN><BR /><BR /><SPAN>UPDATE: Digging a little more, I realize I may be going down the wrong path.&nbsp; I believe it's using JGroups instead of EHCache so, my change probably didn't affect anything!&nbsp; Here is the ehcache_custom.xml</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; cacheManagerPeerProviderFactory</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; class="org.alfresco.repo.cache.AlfrescoCacheManagerPeerProviderFactory"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; properties="heartbeatInterval=5000,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; peerDiscovery=automatic,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; multicastGroupAddress=230.0.0.1,</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; multicastGroupPort=4446"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; </SPAN><BR /><BR /><SPAN>I guess this means it's using JGroups?&nbsp; So maybe a timeout in JGroups?&nbsp; The alfresco-jgroups-TCP.xml is this:</SPAN><BR /><SPAN>&lt;config&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;TCP bind_port="${alfresco.tcp.start_port:7800}"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; loopback="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; recv_buf_size="20000000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; send_buf_size="640000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; discard_incompatible_packets="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; max_bundle_size="64000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; max_bundle_timeout="30"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; enable_bundling="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; use_send_queues="false"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; sock_conn_timeout="300"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; skip_suspected_members="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.enabled="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.min_threads="1"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.max_threads="25"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.keep_alive_time="5000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.queue_enabled="false"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.queue_max_size="100"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; thread_pool.rejection_policy="run"</SPAN><BR /><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.enabled="true"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.min_threads="1"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.max_threads="8"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.keep_alive_time="5000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.queue_enabled="false"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.queue_max_size="100"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; oob_thread_pool.rejection_policy="run"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;TCPPING timeout="3000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; initial_hosts="${alfresco.tcp.initial_hosts:localhost[7800]}"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; port_range="${alfresco.tcp.port_range:3}"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; num_initial_members="2"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;MERGE2 max_interval="30000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; min_interval="10000"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;FD_SIMPLE timeout="10000" max_missed_hbs="10" /&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;VERIFY_SUSPECT timeout="1500"&nbsp; /&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;BARRIER /&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;pbcast.NAKACK</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; use_mcast_xmit="false" gc_lag="0"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; retransmit_timeout="300,600,1200,2400,4800"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; discard_delivered_msgs="true"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;UNICAST timeout="300,600,1200" /&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;pbcast.STABLE stability_delay="1000" desired_avg_gossip="50000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; max_bytes="400000"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;VIEW_SYNC avg_send_interval="60000"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;pbcast.GMS print_local_addr="true" join_timeout="3000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; view_bundling="true"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;FC max_credits="2000000"</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; min_threshold="0.10"/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;FRAG2 frag_size="60000"&nbsp; /&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;pbcast.STREAMING_STATE_TRANSFER/&gt;</SPAN><BR /><SPAN>&nbsp;&nbsp;&nbsp; &lt;!– &lt;pbcast.STATE_TRANSFER/&gt; –&gt;&nbsp; </SPAN><BR /><SPAN>&lt;/config&gt;</SPAN><BR /><BR /><BR /><SPAN>Any help would be appreciated.</SPAN><BR /><BR /><BR /><SPAN>Thanks,</SPAN><BR /></BODY></HTML> Wed, 04 Jun 2014 22:24:58 GMT crichsource360 2014-06-04T22:24:58Z Cluster content inconsistencies https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307212#M260342 Hello,We have a 2 node cluster of Alfresco.&nbsp; We have content being pushed to the repository via WebDAV.&nbsp; We find that every once in a while, the opposite node does not have the updated file.&nbsp; The database has the updated information but, the object (from Hibernate, I presuem) is not updated.&nbsp; So, wh Wed, 04 Jun 2014 22:24:58 GMT https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307212#M260342 crichsource360 2014-06-04T22:24:58Z Re: Cluster content inconsistencies https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307213#M260343 <HTML><HEAD></HEAD><BODY><SPAN>Hi,</SPAN><BR /><BR /><SPAN>Which version of alfresco are you using ?</SPAN><BR /><SPAN>As a quick advice, can you please check that both node's system clock are in sync.</SPAN></BODY></HTML> Mon, 09 Jun 2014 12:59:44 GMT https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307213#M260343 yogeshpj 2014-06-09T12:59:44Z Re: Cluster content inconsistencies https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307214#M260344 <HTML><HEAD></HEAD><BODY><SPAN>This is Version 3.1 Enterprise.</SPAN><BR /><SPAN>Definitely, the clocks are synchronized and verified to be working correctly.</SPAN></BODY></HTML> Tue, 10 Jun 2014 14:31:02 GMT https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307214#M260344 crichsource360 2014-06-10T14:31:02Z Re: Cluster content inconsistencies https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307215#M260345 <HTML><HEAD></HEAD><BODY><SPAN>If its 3.1 Enterprise then please call alfresco support.</SPAN></BODY></HTML> Tue, 10 Jun 2014 15:06:19 GMT https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307215#M260345 mrogers 2014-06-10T15:06:19Z Re: Cluster content inconsistencies https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307216#M260346 <HTML><HEAD></HEAD><BODY><SPAN>It's Alfresco OEM'd into Adobe LiveCycle and support is ending for this.&nbsp; I would like to solve this outside of support.</SPAN></BODY></HTML> Tue, 10 Jun 2014 20:08:59 GMT https://connect.hyland.com/t5/alfresco-archive/cluster-content-inconsistencies/m-p/307216#M260346 crichsource360 2014-06-10T20:08:59Z