[JBoss JIRA] (ISPN-5614) Write performance regression after ISPN-5484
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-5614?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-5614:
------------------------------------
Fix Version/s: 8.0.0.Final
(was: 8.0.0.CR1)
> Write performance regression after ISPN-5484
> --------------------------------------------
>
> Key: ISPN-5614
> URL: https://issues.jboss.org/browse/ISPN-5614
> Project: Infinispan
> Issue Type: Bug
> Components: Core
> Affects Versions: 8.0.0.Beta1
> Reporter: Dan Berindei
> Assignee: Dan Berindei
> Fix For: 8.0.0.Final
>
>
> Regression test shows a significant drop in throughput in the replicated and distributed write tests.
> This was after adjusting the internal thread pool settings in the JGroups configuration: with the default (min=5, max=20, queue=0), the distributed read test would fail to finish.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
9 years, 1 month
[JBoss JIRA] (ISPN-5623) Retried prepare commands do not wait for backup locks
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-5623?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-5623:
------------------------------------
Fix Version/s: 8.0.0.Final
(was: 8.0.0.CR1)
> Retried prepare commands do not wait for backup locks
> -----------------------------------------------------
>
> Key: ISPN-5623
> URL: https://issues.jboss.org/browse/ISPN-5623
> Project: Infinispan
> Issue Type: Bug
> Components: Core
> Affects Versions: 7.2.3.Final, 8.0.0.Beta1
> Reporter: Dan Berindei
> Assignee: Pedro Ruivo
> Fix For: 8.0.0.Final
>
> Attachments: OptimisticPrimaryOwnerCrashDuringPrepareTest.java
>
>
> When the primary owner crashes during prepare, the prepare command is retried on the new primary owner. But because the transaction topology id stays the same, the new primary owner does not check for backup locks owned by other transactions from the previous topology.
> That makes it possible for the retried prepare command to lock a key that was already locked by another transaction on the primary owner.
> This wasn't a problem before the ISPN-4546 fix, because a primary owner crash would have forced the transaction to roll back.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
9 years, 1 month
[JBoss JIRA] (ISPN-5643) Simplify handling of previous values
by Gustavo Fernandes (JIRA)
[ https://issues.jboss.org/browse/ISPN-5643?page=com.atlassian.jira.plugin.... ]
Gustavo Fernandes updated ISPN-5643:
------------------------------------
Fix Version/s: 8.0.0.Final
(was: 8.0.0.CR1)
> Simplify handling of previous values
> ------------------------------------
>
> Key: ISPN-5643
> URL: https://issues.jboss.org/browse/ISPN-5643
> Project: Infinispan
> Issue Type: Task
> Components: Core
> Affects Versions: 8.0.0.Beta2
> Reporter: Dan Berindei
> Assignee: Dan Berindei
> Fix For: 8.0.0.Final
>
>
> There are 5 ways the user can influence whether the previous value is read from a store/remote node:
> * unsafe.unreliableReturnValues
> * IGNORE_RETURN_VALUES
> * SKIP_CACHE_LOAD
> * SKIP_REMOTE_LOOKUP
> * CACHE_MODE_LOCAL
> Currently the effect of each is slightly different, depending on whether the operation is conditional, whether it's a delta-aware write etc. This is how I think it should work:
> * unsafe.unreliableReturnValues and IGNORE_RETURN_VALUES should skip reading the previous value only if it doesn't change the outcome of the operation (i.e. no effect for reads, conditional operations, or delta-aware).
> * SKIP_CACHE_LOAD should *always* skip the cache loader.
> * SKIP_REMOTE_LOOKUP and CACHE_MODE_LOCAL should *always* skip the remote lookup
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
9 years, 1 month