[JBoss JIRA] Created: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
3:19 a.m.
With L1 disabled a new node join may result some entries are not invalidated for non-owner
in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
I have a test case that have 30,000 entries in cache in dist mode with number of owner set
to 2 and L1 is disabled. A new node join the cluster, then the total number of entries in
all cache is larger than 60,000. Even worse, after I execute cache.clear() on all caches
in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
3:21 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Changgeng Li updated ISPN-1012:
-------------------------------
Attachment: test-4.2.1final.tgz
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: test-4.2.1final.tgz
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
9:22 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Erik Salter updated ISPN-1012:
------------------------------
Attachment: test-4.2.1final.zip
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: test-4.2.1final.tgz, test-4.2.1final.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
9:24 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Erik Salter updated ISPN-1012:
------------------------------
Attachment: test-4.2.1final.zip
Here's an updated test case that replaces cache.size() with the reliable
cache.size().entrySet().
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: test-4.2.1final.tgz, test-4.2.1final.zip, test-4.2.1final.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
11:29 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Erik Salter updated ISPN-1012:
------------------------------
Attachment: DataCountTest.java
Per request, this is a version of the test that uses ISPN's existing test suite.
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:21 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Manik Surtani updated ISPN-1012:
--------------------------------
Fix Version/s: 4.2.2.BETA1
Component/s: Distributed Cache
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Fix For: 4.2.2.BETA1
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
5:45 a.m.
New subject: [JBoss JIRA] Resolved: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Manik Surtani resolved ISPN-1012.
---------------------------------
Fix Version/s: (was: 4.2.2.BETA1)
Resolution: Cannot Reproduce Bug
A couple of things here:
* Your test (DataCountTest) did not disable L1. After disabling L1, I don't see the
"extra" entries.
* The clear() commands sometimes get issued before a rehash completes, which is why some
transient state being moved across gets applied after the clear().
I've updated the test a bit, and it is now here: https://gist.github.com/936439
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
8:59 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Erik Salter updated ISPN-1012:
------------------------------
Attachment: test-5.0.0.CR3.zip
We re-ran the original test against 5.0.0.CR3 to verify the findings (and we made sure to
explicitly disable Ll).
We're still seeing data count inconsistencies after re-adding a node.
Also, after waiting 30 seconds after issuing a clear, there is still data in the caches.
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
9:07 a.m.
New subject: [JBoss JIRA] Reopened: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Dan Berindei reopened ISPN-1012:
--------------------------------
Reopening as Erik can still reproduce the issue.
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:38 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Manik Surtani updated ISPN-1012:
--------------------------------
Fix Version/s: 5.0.0.CR5
5.0.0.FINAL
Priority: Critical (was: Major)
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Manik Surtani
Priority: Critical
Fix For: 5.0.0.CR5, 5.0.0.FINAL
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
1:27 a.m.
New subject: [JBoss JIRA] Assigned: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Dan Berindei reassigned ISPN-1012:
----------------------------------
Assignee: Dan Berindei (was: Manik Surtani)
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Dan Berindei
Priority: Critical
Fix For: 5.0.0.CR5, 5.0.0.FINAL
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
12:20 p.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Manik Surtani updated ISPN-1012:
--------------------------------
Fix Version/s: 5.0.0.CR6
(was: 5.0.0.CR5)
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Dan Berindei
Priority: Critical
Fix For: 5.0.0.CR6, 5.0.0.FINAL
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:03 p.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Dan Berindei updated ISPN-1012:
-------------------------------
Status: Pull Request Sent (was: Reopened)
Git Pull Request: https://github.com/infinispan/infinispan/pull/391
* (In the ISPN-1171 commit) I modified ConfigurationValidatingVisitor to disable
L1OnRehash when L1 is disabled.
* I modified DefaultCacheManager to apply the validating visitor in all the
constructors for the global configuration and when creating the cache for the named cache
configurations.
* I fixed CloneableConfigurationComponent so cloned configuration objects don't
share the overridden elements list.
* In the ISPN-1171 commit I modified ConfigurationValidatingVisitor to disable
L1OnRehash when L1 is disabled.
* I modified DefaultCacheManager to apply the validating visitor in all the
constructors for the global configuration and when creating the cache for the named cache
configurations.
* I fixed CloneableConfigurationComponent so cloned configuration objects don't
share the overridden elements list.
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Dan Berindei
Priority: Critical
Fix For: 5.0.0.CR7, 5.0.0.FINAL
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
10:36 a.m.
New subject: [JBoss JIRA] Updated: (ISPN-1012) With L1 disabled a new node join may result some entries are not invalidated for non-owner in DIST mode
[
https://issues.jboss.org/browse/ISPN-1012?page=com.atlassian.jira.plugin....
]
Galder Zamarreño updated ISPN-1012:
-----------------------------------
Status: Resolved (was: Pull Request Sent)
Fix Version/s: 5.0.0.CR6
(was: 5.0.0.CR7)
Resolution: Done
With L1 disabled a new node join may result some entries are not
invalidated for non-owner in DIST mode
-------------------------------------------------------------------------------------------------------
Key: ISPN-1012
URL: https://issues.jboss.org/browse/ISPN-1012
Project: Infinispan
Issue Type: Bug
Components: Distributed Cache
Affects Versions: 4.2.1.FINAL
Reporter: Changgeng Li
Assignee: Dan Berindei
Priority: Critical
Fix For: 5.0.0.CR6, 5.0.0.FINAL
Attachments: DataCountTest.java, test-4.2.1final.tgz, test-4.2.1final.zip,
test-4.2.1final.zip, test-5.0.0.CR3.zip
I have a test case that have 30,000 entries in cache in dist mode with number of owner
set to 2 and L1 is disabled. A new node join the cluster, then the total number of entries
in all cache is larger than 60,000. Even worse, after I execute cache.clear() on all
caches in the cluster, there are still lots of entries left.
The attached unit test can reproduce this issue.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
4708
days inactive
4786
days old
infinispan-issues@lists.jboss.org
13 comments
5 participants
participants (5)
-
Changgeng Li (JIRA) -
Dan Berindei (JIRA)
-
Erik Salter (JIRA)
-
Galder Zamarreño (JIRA)
-
Manik Surtani (JIRA)