[JBoss JIRA] (ISPN-4846) State transfer keeps trying to fetch transaction data after the cache was stopped
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-4846?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-4846:
------------------------------------
Fix Version/s: 8.2.0.Final
(was: 8.2.0.CR1)
> State transfer keeps trying to fetch transaction data after the cache was stopped
> ---------------------------------------------------------------------------------
>
> Key: ISPN-4846
> URL: https://issues.jboss.org/browse/ISPN-4846
> Project: Infinispan
> Issue Type: Bug
> Components: Core, State Transfer
> Affects Versions: 7.0.0.CR1
> Reporter: Dan Berindei
> Fix For: 8.2.0.Final
>
>
> StateConsumerImpl doesn't check if the cache is stopped while fetching transaction data, it only stops when it's no longer able to find providers for transactions.
> However, JGroupsTransport throws a generic CacheException when the channel is stopped. The state transfer thread can enter a busy-wait loop, retrying to get the transaction data and immediately getting the CacheException, filling the log with messages like this:
> {noformat}
> 19:32:28,237 WARN (remote-thread-NodeN-p42592-t1:) [StateConsumerImpl] ISPN000209: Failed to retrieve transactions for segments [10, 11, 12, 13, 14, 15, 17, 16, 19, 18, 21, 20, 23, 22, 25, 24, 27, 26, 29, 28, 42, 43, 40, 41, 46, 47, 44, 45, 51, 50, 49, 48, 55, 54, 53, 52, 59, 58, 57, 56] of cache testCache from node NodeM-53416
> org.infinispan.commons.CacheException: java.lang.IllegalStateException: channel is not connected
> at org.infinispan.commons.util.Util.rewrapAsCacheException(Util.java:655)
> at org.infinispan.remoting.transport.jgroups.CommandAwareRpcDispatcher.invokeRemoteCommand(CommandAwareRpcDispatcher.java:176)
> at org.infinispan.remoting.transport.jgroups.JGroupsTransport.invokeRemotely(JGroupsTransport.java:536)
> at org.infinispan.remoting.rpc.RpcManagerImpl.invokeRemotely(RpcManagerImpl.java:290)
> at org.infinispan.statetransfer.StateConsumerImpl.getTransactions(StateConsumerImpl.java:766)
> at org.infinispan.statetransfer.StateConsumerImpl.requestTransactions(StateConsumerImpl.java:685)
> at org.infinispan.statetransfer.StateConsumerImpl.addTransfers(StateConsumerImpl.java:629)
> at org.infinispan.statetransfer.StateConsumerImpl.onTopologyUpdate(StateConsumerImpl.java:331)
> at org.infinispan.statetransfer.StateTransferManagerImpl.doTopologyUpdate(StateTransferManagerImpl.java:195)
> at org.infinispan.statetransfer.StateTransferManagerImpl.access$000(StateTransferManagerImpl.java:43)
> at org.infinispan.statetransfer.StateTransferManagerImpl$1.rebalance(StateTransferManagerImpl.java:116)
> {noformat}
> We should check is the cache is stopped before retrying in StateConsumerImpl.requestTransactions. I also think we should change the stop order - it would make sense to stop the remote executor threads and the RpcDispatcher before we stop the channel.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 1 month
[JBoss JIRA] (ISPN-4817) Cluster Listener replication of listener to remote dist nodes needs privilege action
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-4817?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-4817:
------------------------------------
Fix Version/s: 8.2.0.Final
(was: 8.2.0.CR1)
> Cluster Listener replication of listener to remote dist nodes needs privilege action
> ------------------------------------------------------------------------------------
>
> Key: ISPN-4817
> URL: https://issues.jboss.org/browse/ISPN-4817
> Project: Infinispan
> Issue Type: Bug
> Components: Loaders and Stores
> Affects Versions: 7.0.0.CR1
> Reporter: William Burns
> Priority: Critical
> Fix For: 8.2.0.Final
>
>
> The ClusterListenerReplicateCallable needs to add the listener in a privileged block so that it can work properly.
> I was also thinking if we need to change all of our messages to send along the current Subject so it can be properly applied to remote invocations as well. In this case we wouldn't need to wrap the listener invocation.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 1 month
[JBoss JIRA] (ISPN-4722) CLI remove is not cluster-wide
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-4722?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-4722:
------------------------------------
Fix Version/s: 8.2.0.Final
(was: 8.2.0.CR1)
> CLI remove is not cluster-wide
> ------------------------------
>
> Key: ISPN-4722
> URL: https://issues.jboss.org/browse/ISPN-4722
> Project: Infinispan
> Issue Type: Bug
> Components: CLI
> Affects Versions: 6.0.2.Final, 7.0.0.Beta1
> Reporter: Galder Zamarreño
> Assignee: Tristan Tarrant
> Fix For: 8.2.0.Final
>
>
> In CLI, the "remove" command does not delete entries in all nodes of a clustered environment, only the local copy. However, the "put" command does write in all nodes. Is it the expected behavior? See example below:
> {code}
> node 1
> put k1 v1
> get k1 -> v1
> node 2
> get k1 -> v1
> node 1
> remove k1
> get k1 -> null
> node 2
> get k1 -> v1
> {code}
> I know that these commands provided by CLI are not used in real world, but they are useful to demonstrate the correct configuration of a JDG cluster.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 1 month
[JBoss JIRA] (ISPN-4624) Allow custom partition handling strategy
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-4624?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-4624:
------------------------------------
Fix Version/s: 8.2.0.Final
(was: 8.2.0.CR1)
> Allow custom partition handling strategy
> ----------------------------------------
>
> Key: ISPN-4624
> URL: https://issues.jboss.org/browse/ISPN-4624
> Project: Infinispan
> Issue Type: Feature Request
> Components: Core
> Affects Versions: 7.0.0.Beta1
> Reporter: Dan Berindei
> Labels: partition_handling
> Fix For: 8.2.0.Final
>
>
> Users should be able to configure a custom PartitionHandlingStrategy. It should be able to specify a behaviour on merge as well.
> We might want to merge this with the RebalancePolicy, which was also supposed to be configurable.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 1 month
[JBoss JIRA] (ISPN-5042) Remote gets caused by writes could be replicated only to the primary owner
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-5042?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-5042:
------------------------------------
Fix Version/s: 8.2.0.Final
(was: 8.2.0.CR1)
> Remote gets caused by writes could be replicated only to the primary owner
> --------------------------------------------------------------------------
>
> Key: ISPN-5042
> URL: https://issues.jboss.org/browse/ISPN-5042
> Project: Infinispan
> Issue Type: Enhancement
> Components: Core, State Transfer
> Affects Versions: 7.1.0.Alpha1
> Reporter: Dan Berindei
> Priority: Minor
> Labels: 7.0
> Fix For: 8.2.0.Final
>
>
> For write operations that need the previous value, a write CH-only owner that doesn't have a key locally will attempt to retrieve the key from the read CH-owners.
> Sending the remote get command to all the previous owners will create extra load on the cluster during state transfer, so it should be more efficient to send the remote get only to the primary owner. Even though the latency of some write operations will be higher, the average latency should be better.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)
10 years, 1 month