You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@ambari.apache.org by Nahappan Somasundaram <ns...@hortonworks.com> on 2016/03/23 23:45:36 UTC
Review Request 45253: AMBARI-15544: Creating multinode cluster using
Blueprints fails.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/
-----------------------------------------------------------
Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
Bugs: AMBARI-15544
https://issues.apache.org/jira/browse/AMBARI-15544
Repository: ambari
Description
-------
AMBARI-15544: Creating multinode cluster using Blueprints fails.
** Issue **:
This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
** Fix **:
Check the returned Timestamp object for null before accessing it.
Diffs
-----
ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
Diff: https://reviews.apache.org/r/45253/diff/
Testing
-------
** 1. mvn clean install -DskipTests **
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary:
[INFO]
[INFO] Ambari Main ....................................... SUCCESS [5.199s]
[INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
[INFO] Ambari Web ........................................ SUCCESS [31.250s]
[INFO] Ambari Views ...................................... SUCCESS [1.140s]
[INFO] Ambari Admin View ................................. SUCCESS [5.628s]
[INFO] ambari-metrics .................................... SUCCESS [0.355s]
[INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
[INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
[INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
[INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
[INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
[INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
[INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
[INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
[INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
[INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
[INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
[INFO] Ambari Agent ...................................... SUCCESS [22.491s]
[INFO] Ambari Client ..................................... SUCCESS [0.061s]
[INFO] Ambari Python Client .............................. SUCCESS [1.008s]
[INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
[INFO] Ambari Shell ...................................... SUCCESS [0.058s]
[INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
[INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 5:16.311s
[INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
[INFO] Final Memory: 261M/1167M
[INFO] ------------------------------------------------------------------------
** 2. Manual tests **
Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
** 3. Unit tests **
-------------------------------------------------------
T E S T S
-------------------------------------------------------
Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
Running org.apache.ambari.server.agent.TestHeartbeatHandler
Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
Results :
Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1:43.413s
[INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
[INFO] Final Memory: 55M/1019M
[INFO] ------------------------------------------------------------------------
Thanks,
Nahappan Somasundaram
Re: Review Request 45253: AMBARI-15544: Creating multinode cluster
using Blueprints fails.
Posted by Sumit Mohanty <sm...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/#review125147
-----------------------------------------------------------
Ship it!
Ship It!
- Sumit Mohanty
On March 23, 2016, 10:45 p.m., Nahappan Somasundaram wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45253/
> -----------------------------------------------------------
>
> (Updated March 23, 2016, 10:45 p.m.)
>
>
> Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
>
>
> Bugs: AMBARI-15544
> https://issues.apache.org/jira/browse/AMBARI-15544
>
>
> Repository: ambari
>
>
> Description
> -------
>
> AMBARI-15544: Creating multinode cluster using Blueprints fails.
>
> ** Issue **:
>
> This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
>
> ** Fix **:
>
> Check the returned Timestamp object for null before accessing it.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
>
> Diff: https://reviews.apache.org/r/45253/diff/
>
>
> Testing
> -------
>
> ** 1. mvn clean install -DskipTests **
>
> [INFO] ------------------------------------------------------------------------
> [INFO] Reactor Summary:
> [INFO]
> [INFO] Ambari Main ....................................... SUCCESS [5.199s]
> [INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
> [INFO] Ambari Web ........................................ SUCCESS [31.250s]
> [INFO] Ambari Views ...................................... SUCCESS [1.140s]
> [INFO] Ambari Admin View ................................. SUCCESS [5.628s]
> [INFO] ambari-metrics .................................... SUCCESS [0.355s]
> [INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
> [INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
> [INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
> [INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
> [INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
> [INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
> [INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
> [INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
> [INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
> [INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
> [INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
> [INFO] Ambari Agent ...................................... SUCCESS [22.491s]
> [INFO] Ambari Client ..................................... SUCCESS [0.061s]
> [INFO] Ambari Python Client .............................. SUCCESS [1.008s]
> [INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
> [INFO] Ambari Shell ...................................... SUCCESS [0.058s]
> [INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
> [INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 5:16.311s
> [INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
> [INFO] Final Memory: 261M/1167M
> [INFO] ------------------------------------------------------------------------
>
> ** 2. Manual tests **
>
> Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
>
> ** 3. Unit tests **
>
>
> -------------------------------------------------------
> T E S T S
> -------------------------------------------------------
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.agent.TestHeartbeatHandler
> Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
> Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
>
> Results :
>
> Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 1:43.413s
> [INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
> [INFO] Final Memory: 55M/1019M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nahappan Somasundaram
>
>
Re: Review Request 45253: AMBARI-15544: Creating multinode cluster
using Blueprints fails.
Posted by Nate Cole <nc...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/#review125232
-----------------------------------------------------------
Ship it!
ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java (lines 150 - 152)
<https://reviews.apache.org/r/45253/#comment188022>
Is there a unit test that can be added to make sure we don't accidentally revert this fix somehow?
- Nate Cole
On March 23, 2016, 6:45 p.m., Nahappan Somasundaram wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45253/
> -----------------------------------------------------------
>
> (Updated March 23, 2016, 6:45 p.m.)
>
>
> Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
>
>
> Bugs: AMBARI-15544
> https://issues.apache.org/jira/browse/AMBARI-15544
>
>
> Repository: ambari
>
>
> Description
> -------
>
> AMBARI-15544: Creating multinode cluster using Blueprints fails.
>
> ** Issue **:
>
> This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
>
> ** Fix **:
>
> Check the returned Timestamp object for null before accessing it.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
>
> Diff: https://reviews.apache.org/r/45253/diff/
>
>
> Testing
> -------
>
> ** 1. mvn clean install -DskipTests **
>
> [INFO] ------------------------------------------------------------------------
> [INFO] Reactor Summary:
> [INFO]
> [INFO] Ambari Main ....................................... SUCCESS [5.199s]
> [INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
> [INFO] Ambari Web ........................................ SUCCESS [31.250s]
> [INFO] Ambari Views ...................................... SUCCESS [1.140s]
> [INFO] Ambari Admin View ................................. SUCCESS [5.628s]
> [INFO] ambari-metrics .................................... SUCCESS [0.355s]
> [INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
> [INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
> [INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
> [INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
> [INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
> [INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
> [INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
> [INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
> [INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
> [INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
> [INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
> [INFO] Ambari Agent ...................................... SUCCESS [22.491s]
> [INFO] Ambari Client ..................................... SUCCESS [0.061s]
> [INFO] Ambari Python Client .............................. SUCCESS [1.008s]
> [INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
> [INFO] Ambari Shell ...................................... SUCCESS [0.058s]
> [INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
> [INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 5:16.311s
> [INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
> [INFO] Final Memory: 261M/1167M
> [INFO] ------------------------------------------------------------------------
>
> ** 2. Manual tests **
>
> Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
>
> ** 3. Unit tests **
>
>
> -------------------------------------------------------
> T E S T S
> -------------------------------------------------------
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.agent.TestHeartbeatHandler
> Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
> Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
>
> Results :
>
> Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 1:43.413s
> [INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
> [INFO] Final Memory: 55M/1019M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nahappan Somasundaram
>
>
Re: Review Request 45253: AMBARI-15544: Creating multinode cluster
using Blueprints fails.
Posted by Sebastian Toader <st...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/#review125220
-----------------------------------------------------------
Ship it!
Ship It!
- Sebastian Toader
On March 23, 2016, 11:45 p.m., Nahappan Somasundaram wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45253/
> -----------------------------------------------------------
>
> (Updated March 23, 2016, 11:45 p.m.)
>
>
> Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
>
>
> Bugs: AMBARI-15544
> https://issues.apache.org/jira/browse/AMBARI-15544
>
>
> Repository: ambari
>
>
> Description
> -------
>
> AMBARI-15544: Creating multinode cluster using Blueprints fails.
>
> ** Issue **:
>
> This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
>
> ** Fix **:
>
> Check the returned Timestamp object for null before accessing it.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
>
> Diff: https://reviews.apache.org/r/45253/diff/
>
>
> Testing
> -------
>
> ** 1. mvn clean install -DskipTests **
>
> [INFO] ------------------------------------------------------------------------
> [INFO] Reactor Summary:
> [INFO]
> [INFO] Ambari Main ....................................... SUCCESS [5.199s]
> [INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
> [INFO] Ambari Web ........................................ SUCCESS [31.250s]
> [INFO] Ambari Views ...................................... SUCCESS [1.140s]
> [INFO] Ambari Admin View ................................. SUCCESS [5.628s]
> [INFO] ambari-metrics .................................... SUCCESS [0.355s]
> [INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
> [INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
> [INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
> [INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
> [INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
> [INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
> [INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
> [INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
> [INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
> [INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
> [INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
> [INFO] Ambari Agent ...................................... SUCCESS [22.491s]
> [INFO] Ambari Client ..................................... SUCCESS [0.061s]
> [INFO] Ambari Python Client .............................. SUCCESS [1.008s]
> [INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
> [INFO] Ambari Shell ...................................... SUCCESS [0.058s]
> [INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
> [INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 5:16.311s
> [INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
> [INFO] Final Memory: 261M/1167M
> [INFO] ------------------------------------------------------------------------
>
> ** 2. Manual tests **
>
> Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
>
> ** 3. Unit tests **
>
>
> -------------------------------------------------------
> T E S T S
> -------------------------------------------------------
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.agent.TestHeartbeatHandler
> Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
> Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
>
> Results :
>
> Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 1:43.413s
> [INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
> [INFO] Final Memory: 55M/1019M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nahappan Somasundaram
>
>
Re: Review Request 45253: AMBARI-15544: Creating multinode cluster
using Blueprints fails.
Posted by Nate Cole <nc...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/#review125295
-----------------------------------------------------------
Ship it!
Ship It!
- Nate Cole
On March 24, 2016, 12:42 p.m., Nahappan Somasundaram wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45253/
> -----------------------------------------------------------
>
> (Updated March 24, 2016, 12:42 p.m.)
>
>
> Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
>
>
> Bugs: AMBARI-15544
> https://issues.apache.org/jira/browse/AMBARI-15544
>
>
> Repository: ambari
>
>
> Description
> -------
>
> AMBARI-15544: Creating multinode cluster using Blueprints fails.
>
> ** Issue **:
>
> This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
>
> ** Fix **:
>
> Check the returned Timestamp object for null before accessing it.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
> ambari-server/src/test/java/org/apache/ambari/server/configuration/RecoveryConfigHelperTest.java f7c85b6d514cee908f07513e62874ed9ed393fa4
>
> Diff: https://reviews.apache.org/r/45253/diff/
>
>
> Testing
> -------
>
> ** 1. mvn clean install -DskipTests **
>
> [INFO] ------------------------------------------------------------------------
> [INFO] Reactor Summary:
> [INFO]
> [INFO] Ambari Main ....................................... SUCCESS [5.199s]
> [INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
> [INFO] Ambari Web ........................................ SUCCESS [31.250s]
> [INFO] Ambari Views ...................................... SUCCESS [1.140s]
> [INFO] Ambari Admin View ................................. SUCCESS [5.628s]
> [INFO] ambari-metrics .................................... SUCCESS [0.355s]
> [INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
> [INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
> [INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
> [INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
> [INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
> [INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
> [INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
> [INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
> [INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
> [INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
> [INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
> [INFO] Ambari Agent ...................................... SUCCESS [22.491s]
> [INFO] Ambari Client ..................................... SUCCESS [0.061s]
> [INFO] Ambari Python Client .............................. SUCCESS [1.008s]
> [INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
> [INFO] Ambari Shell ...................................... SUCCESS [0.058s]
> [INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
> [INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 5:16.311s
> [INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
> [INFO] Final Memory: 261M/1167M
> [INFO] ------------------------------------------------------------------------
>
> ** 2. Manual tests **
>
> Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
>
> ** 3. Unit tests **
>
>
> -------------------------------------------------------
> T E S T S
> -------------------------------------------------------
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.agent.TestHeartbeatHandler
> Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
> Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
> Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
> Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
>
> Results :
>
> Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 1:43.413s
> [INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
> [INFO] Final Memory: 55M/1019M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nahappan Somasundaram
>
>
Re: Review Request 45253: AMBARI-15544: Creating multinode cluster
using Blueprints fails.
Posted by Nahappan Somasundaram <ns...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45253/
-----------------------------------------------------------
(Updated March 24, 2016, 9:42 a.m.)
Review request for Ambari, Jonathan Hurley, Nate Cole, Sumit Mohanty, Sebastian Toader, and Sid Wagle.
Changes
-------
Added a unit test to simulate scenario that caused the NPE.
Bugs: AMBARI-15544
https://issues.apache.org/jira/browse/AMBARI-15544
Repository: ambari
Description
-------
AMBARI-15544: Creating multinode cluster using Blueprints fails.
** Issue **:
This issue happens when there are multiple agents running before deployment happens. During registration, there is no cluster, so the recovery configuration is not obtained. Subsequently, when hosts become a part of a cluster, agents attempt to get the recovery configuration during the heartbeats. The first agent successfully gets the configuration because the timestamp map is empty. When the next agent heartbeats, it checks to see if the configuration is stale. While there is an entry for the cluster name in the timestamp map created by the previous agent, there is no hostname entry for the current agent which causes the timestamp returned for that hostname to be null.
** Fix **:
Check the returned Timestamp object for null before accessing it.
Diffs (updated)
-----
ambari-server/src/main/java/org/apache/ambari/server/agent/RecoveryConfigHelper.java dca4a9b9a32377a2d7d620d6f939e7250cc40590
ambari-server/src/test/java/org/apache/ambari/server/configuration/RecoveryConfigHelperTest.java f7c85b6d514cee908f07513e62874ed9ed393fa4
Diff: https://reviews.apache.org/r/45253/diff/
Testing
-------
** 1. mvn clean install -DskipTests **
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary:
[INFO]
[INFO] Ambari Main ....................................... SUCCESS [5.199s]
[INFO] Apache Ambari Project POM ......................... SUCCESS [0.037s]
[INFO] Ambari Web ........................................ SUCCESS [31.250s]
[INFO] Ambari Views ...................................... SUCCESS [1.140s]
[INFO] Ambari Admin View ................................. SUCCESS [5.628s]
[INFO] ambari-metrics .................................... SUCCESS [0.355s]
[INFO] Ambari Metrics Common ............................. SUCCESS [0.476s]
[INFO] Ambari Metrics Hadoop Sink ........................ SUCCESS [1.067s]
[INFO] Ambari Metrics Flume Sink ......................... SUCCESS [0.562s]
[INFO] Ambari Metrics Kafka Sink ......................... SUCCESS [0.595s]
[INFO] Ambari Metrics Storm Sink ......................... SUCCESS [1.437s]
[INFO] Ambari Metrics Collector .......................... SUCCESS [6.724s]
[INFO] Ambari Metrics Monitor ............................ SUCCESS [2.089s]
[INFO] Ambari Metrics Grafana ............................ SUCCESS [0.862s]
[INFO] Ambari Metrics Assembly ........................... SUCCESS [1:17.652s]
[INFO] Ambari Server ..................................... SUCCESS [2:31.754s]
[INFO] Ambari Functional Tests ........................... SUCCESS [1.281s]
[INFO] Ambari Agent ...................................... SUCCESS [22.491s]
[INFO] Ambari Client ..................................... SUCCESS [0.061s]
[INFO] Ambari Python Client .............................. SUCCESS [1.008s]
[INFO] Ambari Groovy Client .............................. SUCCESS [2.175s]
[INFO] Ambari Shell ...................................... SUCCESS [0.058s]
[INFO] Ambari Python Shell ............................... SUCCESS [0.694s]
[INFO] Ambari Groovy Shell ............................... SUCCESS [1.028s]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 5:16.311s
[INFO] Finished at: Wed Mar 23 15:30:45 PDT 2016
[INFO] Final Memory: 261M/1167M
[INFO] ------------------------------------------------------------------------
** 2. Manual tests **
Deployed a cluster with 3 nodes, registered a blueprint and template. Noticed that the second agent now gets a **true** value for **isConfigStale(clusterName, hostName, timestamp)**, allowing it to get the recovery configuration.
** 3. Unit tests **
-------------------------------------------------------
T E S T S
-------------------------------------------------------
Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
Running org.apache.ambari.server.agent.TestHeartbeatHandler
Tests run: 24, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 54.831 sec - in org.apache.ambari.server.agent.TestHeartbeatHandler
Picked up _JAVA_OPTIONS: -Xmx2048m -XX:MaxPermSize=512m -Djava.awt.headless=true
Running org.apache.ambari.server.configuration.RecoveryConfigHelperTest
Tests run: 7, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 23.43 sec - in org.apache.ambari.server.configuration.RecoveryConfigHelperTest
Results :
Tests run: 31, Failures: 0, Errors: 0, Skipped: 0
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1:43.413s
[INFO] Finished at: Wed Mar 23 15:44:44 PDT 2016
[INFO] Final Memory: 55M/1019M
[INFO] ------------------------------------------------------------------------
Thanks,
Nahappan Somasundaram