January 2017 - infinispan-issues - Jboss List Archives

[JBoss JIRA] (ISPN-7322) Improve triangle algorithm: ordering by segment

by Dan Berindei (JIRA)

[ https://issues.jboss.org/browse/ISPN-7322?page=com.atlassian.jira.plugin.... ] Dan Berindei commented on ISPN-7322: ------------------------------------ [~belaban] I think the point here is that we *want* more than {{cluster size - 1}} threads processing incoming commands at the same time, because each sender could have 100 threads writing to the cache, and our {{DataContainer}} is optimized for many threads accessing it concurrently. > Improve triangle algorithm: ordering by segment > ----------------------------------------------- > > Key: ISPN-7322 > URL: https://issues.jboss.org/browse/ISPN-7322 > Project: Infinispan > Issue Type: Enhancement > Components: Core > Reporter: Pedro Ruivo > Assignee: Pedro Ruivo > Fix For: 9.0.0.Beta2 > > > Current triangle algorithm uses regular message (FIFO ordered) between the primary owner and backup owners of a key. While it ensures that the backup owners receives the stream of updates in the same order, it makes everything slower since it doesn't allow different keys to be handled in parallel. > "Triangle unordered" solves this problem by sending OOB messages (not ordered) between the primary and backup. To keep the consistency, Infinispan introduces the TriangleOrderManager that orders the updates based on the segment of the key. > While it is not as perfect as ordering per key, the segments are static; this removes the complexity and avoids handling the cluster topology changes and key adding/removal while improves the performance. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (HRJS-31) Old connections not disconnected when switching sites

by Tristan Tarrant (JIRA)

[ https://issues.jboss.org/browse/HRJS-31?page=com.atlassian.jira.plugin.sy... ] Tristan Tarrant updated HRJS-31: -------------------------------- Status: Resolved (was: Pull Request Sent) Resolution: Done > Old connections not disconnected when switching sites > ----------------------------------------------------- > > Key: HRJS-31 > URL: https://issues.jboss.org/browse/HRJS-31 > Project: Infinispan Javascript client > Issue Type: Bug > Affects Versions: 0.3.0 > Reporter: Galder Zamarreño > Assignee: Galder Zamarreño > Fix For: 0.4.0 > > > When a site change happens in the client, old connections are left lingering and they're not disconnected. Before installing any new routers for any sites, irrespective of whether the switch was manual or a as a result of a failover, old connections need to be disconnected. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (HRJS-30) Move testsuite server orchestration to a domain

by Tristan Tarrant (JIRA)

[ https://issues.jboss.org/browse/HRJS-30?page=com.atlassian.jira.plugin.sy... ] Tristan Tarrant updated HRJS-30: -------------------------------- Status: Resolved (was: Pull Request Sent) Resolution: Done > Move testsuite server orchestration to a domain > ----------------------------------------------- > > Key: HRJS-30 > URL: https://issues.jboss.org/browse/HRJS-30 > Project: Infinispan Javascript client > Issue Type: Enhancement > Affects Versions: 0.3.0 > Reporter: Galder Zamarreño > Assignee: Galder Zamarreño > Fix For: 0.4.0 > > > As more tests have been added to the Infinispan JS testsuite, the existing simple server orchestration script/environment are turning out to be a bit painful to deal with. > On one hand, Haskell's {{stack}} can be sometimes difficult to install in some envs, e.g. Ubuntu. I've also seen issues with Turtle library used in Haskell and copying contents inside symbolic links in Red Hat Linux. > On top of that, as security and cross-site tests were added, more configurations were added and more servers need to be added, making it track the configs a bit more confusing. > Finally, the methods used to start/stop servers in self-contained tests, e.g. xsite and cluster failover, a little bit clunky. > So, with this in mind, we thought of using more suitable orchestration techniques for Infinispan servers. We considered docker compose and kubernetes, but since the servers just simply need to run in a single machine and there's not a huge need to have clean environments, we decided to use a single Infinispan server domain to orchestrate all servers. > The advantages of domain is that is easy to start all servers with the right configuration very quickly, and start/stop servers from tests is very easy and quick compared with current approach. On top of that, having domain based orchestration will make it easy to port over the servers, their configs and script over to testsuites of other langs. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-7322) Improve triangle algorithm: ordering by segment

by Dan Berindei (JIRA)

[ https://issues.jboss.org/browse/ISPN-7322?page=com.atlassian.jira.plugin.... ] Dan Berindei updated ISPN-7322: ------------------------------- Status: Resolved (was: Pull Request Sent) Fix Version/s: 9.0.0.Beta2 Resolution: Done > Improve triangle algorithm: ordering by segment > ----------------------------------------------- > > Key: ISPN-7322 > URL: https://issues.jboss.org/browse/ISPN-7322 > Project: Infinispan > Issue Type: Enhancement > Components: Core > Reporter: Pedro Ruivo > Assignee: Pedro Ruivo > Fix For: 9.0.0.Beta2 > > > Current triangle algorithm uses regular message (FIFO ordered) between the primary owner and backup owners of a key. While it ensures that the backup owners receives the stream of updates in the same order, it makes everything slower since it doesn't allow different keys to be handled in parallel. > "Triangle unordered" solves this problem by sending OOB messages (not ordered) between the primary and backup. To keep the consistency, Infinispan introduces the TriangleOrderManager that orders the updates based on the segment of the key. > While it is not as perfect as ordering per key, the segments are static; this removes the complexity and avoids handling the cluster topology changes and key adding/removal while improves the performance. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-7349) Statistics attribute on the cache-container tag is not true by default

by Jakub Markos (JIRA)

Jakub Markos created ISPN-7349: ---------------------------------- Summary: Statistics attribute on the cache-container tag is not true by default Key: ISPN-7349 URL: https://issues.jboss.org/browse/ISPN-7349 Project: Infinispan Issue Type: Bug Components: Configuration Affects Versions: 9.0.0.Alpha2 Reporter: Jakub Markos Priority: Minor The xsd for the 'statistics' attribute looks like this: {code} <xs:attribute name="statistics" type="xs:boolean" default="true"> <xs:annotation> <xs:documentation>Determines whether or not the cache container should collect statistics. Keep disabled for optimal performance.</xs:documentation> </xs:annotation> </xs:attribute> {code} but unless you specifically configure {code} <cache-container name="default" statistics="true"> {code} the 'channel' and 'protocol' jgroups MBeans won't be available through JMX. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-7230) XSite view should be logged at INFO

by Galder Zamarreño (JIRA)

[ https://issues.jboss.org/browse/ISPN-7230?page=com.atlassian.jira.plugin.... ] Galder Zamarreño reassigned ISPN-7230: -------------------------------------- Assignee: Galder Zamarreño > XSite view should be logged at INFO > ----------------------------------- > > Key: ISPN-7230 > URL: https://issues.jboss.org/browse/ISPN-7230 > Project: Infinispan > Issue Type: Feature Request > Components: Cross-Site Replication > Affects Versions: 9.0.0.Alpha4, 8.2.5.Final > Reporter: Galder Zamarreño > Assignee: Galder Zamarreño > > XSite view should be printed as INFO message. Currently, the only log messages printed are: > Site A: > {code} > 11:19:00,398 INFO [org.jgroups.protocols.relay.RELAY2] > (Timer-2,gs-macbook-pro-3) _gs-macbook-pro-3:LON: > joined bridge cluster 'xsite' > {code} > Site B: > {code} > 11:19:31,329 INFO [org.jgroups.protocols.relay.RELAY2] > (Timer-3,gs-macbook-pro-3) _gs-macbook-pro-3:NYC: > joined bridge cluster 'xsite' > {code} -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-6988) Create UI for off-heap configuration

by Vladimir Blagojevic (JIRA)

[ https://issues.jboss.org/browse/ISPN-6988?page=com.atlassian.jira.plugin.... ] Vladimir Blagojevic updated ISPN-6988: -------------------------------------- Status: Pull Request Sent (was: Coding In Progress) Git Pull Request: https://github.com/infinispan/infinispan-management-console/pull/174 > Create UI for off-heap configuration > ------------------------------------- > > Key: ISPN-6988 > URL: https://issues.jboss.org/browse/ISPN-6988 > Project: Infinispan > Issue Type: Feature Request > Components: Console > Reporter: Pedro Zapata > Assignee: Vladimir Blagojevic > Fix For: 9.0.0.Final > > > Configuration of off-heap memory usage should be made available through the console. Also, monitoring of off-heap usage must be shown at the cluster node level. -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-6599) PutAll operation in the Hot Rod client only partially completed during topology changes

by Gustavo Fernandes (JIRA)

[ https://issues.jboss.org/browse/ISPN-6599?page=com.atlassian.jira.plugin.... ] Gustavo Fernandes edited comment on ISPN-6599 at 1/10/17 1:45 PM: ------------------------------------------------------------------ The issue appeared again on 9.0.0.Beta1. Does not happen on Alpha4 or 8.2.5.Final. I used the attached reproducer to check it, running in a loop. Steps: * Change in the pom infinispan.version to 9.0.0.Beta1 * Still in the pom, use version 0.4 for <artifactId>infinispan-spark_2.11</artifactId> * Build the uber jar with {code}mvn clean assembly:assembly -DdescriptorId=jar-with-dependencies{code} * Run in a loop {code}java -cp target/node-1.0-SNAPSHOT-jar-with-dependencies.jar com.gustavonalle.infinispan.perf.test.Failover /path/to/server/9.0.0.Beta1 {code} * It will eventually error with {{Exception in thread "main" java.lang.RuntimeException: Invalid cache size! Expected 1000 but it was 999}} meaning entries were lost during topology change was (Author: gustavonalle): The issue appeared again on 9.0.0.Beta1. Does not happen on Alpha4 or 8.2.5.Final. I used the attached reproducer to check it, running in a loop. Steps: * Change in the pom infinispan.version to 9.0.0.Beta1 * Still in the pom, use version 0.4 for <artifactId>infinispan-spark_2.11</artifactId> * Build the uber jar with {code}mvn clean assembly:assembly -DdescriptorId=jar-with-dependencies{code} * Run in a loop {code}java -cp target/node-1.0-SNAPSHOT-jar-with-dependencies.jar com.gustavonalle.infinispan.perf.test.Failover /path/to/server/9.0.0.Beta1 {code} * it will eventually error with {{Exception in thread "main" java.lang.RuntimeException: Invalid cache size! Expected 1000 but it was 999}} meaning entries were lost during topology change > PutAll operation in the Hot Rod client only partially completed during topology changes > --------------------------------------------------------------------------------------- > > Key: ISPN-6599 > URL: https://issues.jboss.org/browse/ISPN-6599 > Project: Infinispan > Issue Type: Bug > Components: Server > Affects Versions: 8.2.1.Final, 9.0.0.Alpha1 > Reporter: Gustavo Fernandes > Assignee: Dan Berindei > Fix For: 9.0.0.Alpha2, 8.2.2.Final, 9.0.0.Final > > Attachments: reproducer.zip, start.zip, trace-2.zip, trace-3.zip, trace.zip > > -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-6599) PutAll operation in the Hot Rod client only partially completed during topology changes

by Gustavo Fernandes (JIRA)

[ https://issues.jboss.org/browse/ISPN-6599?page=com.atlassian.jira.plugin.... ] Gustavo Fernandes edited comment on ISPN-6599 at 1/10/17 1:45 PM: ------------------------------------------------------------------ The issue appeared again on 9.0.0.Beta1. Does not happen on Alpha4 or 8.2.5.Final. I used the attached reproducer to check it, running in a loop. Steps: * Change in the pom infinispan.version to 9.0.0.Beta1 * Still in the pom, use version 0.4 for <artifactId>infinispan-spark_2.11</artifactId> * Build the uber jar with {code}mvn clean assembly:assembly -DdescriptorId=jar-with-dependencies{code} * Run in a loop {code}java -cp target/node-1.0-SNAPSHOT-jar-with-dependencies.jar com.gustavonalle.infinispan.perf.test.Failover /path/to/server/9.0.0.Beta1 {code} * It will eventually error with {{Exception in thread "main" java.lang.RuntimeException: Invalid cache size! Expected 1000 but it was 999}} meaning entries were lost during topology change was (Author: gustavonalle): The issue appeared again on 9.0.0.Beta1. Does not happen on Alpha4 or 8.2.5.Final. I used the attached reproducer to check it, running in a loop. Steps: * Change in the pom infinispan.version to 9.0.0.Beta1 * Still in the pom, use version 0.4 for <artifactId>infinispan-spark_2.11</artifactId> * Build the uber jar with {code}mvn clean assembly:assembly -DdescriptorId=jar-with-dependencies{code} * Run in a loop {code}java -cp target/node-1.0-SNAPSHOT-jar-with-dependencies.jar com.gustavonalle.infinispan.perf.test.Failover /path/to/server/9.0.0.Beta1 {code} * It will eventually error with {{Exception in thread "main" java.lang.RuntimeException: Invalid cache size! Expected 1000 but it was 999}} meaning entries were lost during topology change > PutAll operation in the Hot Rod client only partially completed during topology changes > --------------------------------------------------------------------------------------- > > Key: ISPN-6599 > URL: https://issues.jboss.org/browse/ISPN-6599 > Project: Infinispan > Issue Type: Bug > Components: Server > Affects Versions: 8.2.1.Final, 9.0.0.Alpha1 > Reporter: Gustavo Fernandes > Assignee: Dan Berindei > Fix For: 9.0.0.Alpha2, 8.2.2.Final, 9.0.0.Final > > Attachments: reproducer.zip, start.zip, trace-2.zip, trace-3.zip, trace.zip > > -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

[JBoss JIRA] (ISPN-6599) PutAll operation in the Hot Rod client only partially completed during topology changes

by Gustavo Fernandes (JIRA)

[ https://issues.jboss.org/browse/ISPN-6599?page=com.atlassian.jira.plugin.... ] Gustavo Fernandes reopened ISPN-6599: ------------------------------------- The issue appeared again on 9.0.0.Beta1. Does not happen on Alpha4 or 8.2.5.Final. I used the attached reproducer to check it, running in a loop. Steps: * Change in the pom infinispan.version to 9.0.0.Beta1 * Still in the pom, use version 0.4 for <artifactId>infinispan-spark_2.11</artifactId> * Build the uber jar with {code}mvn clean assembly:assembly -DdescriptorId=jar-with-dependencies{code} * Run in a loop {code}java -cp target/node-1.0-SNAPSHOT-jar-with-dependencies.jar com.gustavonalle.infinispan.perf.test.Failover /path/to/server/9.0.0.Beta1 {code} * it will eventually error with {{Exception in thread "main" java.lang.RuntimeException: Invalid cache size! Expected 1000 but it was 999}} meaning entries were lost during topology change > PutAll operation in the Hot Rod client only partially completed during topology changes > --------------------------------------------------------------------------------------- > > Key: ISPN-6599 > URL: https://issues.jboss.org/browse/ISPN-6599 > Project: Infinispan > Issue Type: Bug > Components: Server > Affects Versions: 8.2.1.Final, 9.0.0.Alpha1 > Reporter: Gustavo Fernandes > Assignee: Dan Berindei > Fix For: 9.0.0.Alpha2, 8.2.2.Final, 9.0.0.Final > > Attachments: reproducer.zip, start.zip, trace-2.zip, trace-3.zip, trace.zip > > -- This message was sent by Atlassian JIRA (v7.2.3#72005)

7 years, 11 months

1
0
0 / 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

infinispan-issues January 2017