You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by Nate Cole <nc...@hortonworks.com> on 2014/08/15 18:41:30 UTC
Review Request 24742: Flume: stopping agent doesn't work on host with
multiple agents
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24742/
-----------------------------------------------------------
Review request for Ambari, Srimanth Gunturi and Sid Wagle.
Bugs: AMBARI-6876
https://issues.apache.org/jira/browse/AMBARI-6876
Repository: ambari
Description
-------
Due to a refactor, the code that picked up state changes for targeted host components was no longer getting invoked. This patch addresses that.
In the case where you issue a STOP on a host component and you supply information for the host component (as with Flume), the command should be sent down even if the state of that host component is already stopped. The refactor skips this case.
Diffs
-----
ambari-server/src/main/java/org/apache/ambari/server/controller/AmbariManagementControllerImpl.java 59544dd
ambari-server/src/test/java/org/apache/ambari/server/controller/AmbariManagementControllerTest.java fe2023e
Diff: https://reviews.apache.org/r/24742/diff/
Testing
-------
Added test case
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 19:51.784s
[INFO] Finished at: Fri Aug 15 11:50:31 EDT 2014
[INFO] Final Memory: 19M/123M
[INFO] ------------------------------------------------------------------------
Thanks,
Nate Cole
Re: Review Request 24742: Flume: stopping agent doesn't work on host
with multiple agents
Posted by Srimanth Gunturi <sr...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24742/#review50768
-----------------------------------------------------------
Ship it!
Ship It!
- Srimanth Gunturi
On Aug. 15, 2014, 4:41 p.m., Nate Cole wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/24742/
> -----------------------------------------------------------
>
> (Updated Aug. 15, 2014, 4:41 p.m.)
>
>
> Review request for Ambari, Srimanth Gunturi and Sid Wagle.
>
>
> Bugs: AMBARI-6876
> https://issues.apache.org/jira/browse/AMBARI-6876
>
>
> Repository: ambari
>
>
> Description
> -------
>
> Due to a refactor, the code that picked up state changes for targeted host components was no longer getting invoked. This patch addresses that.
>
> In the case where you issue a STOP on a host component and you supply information for the host component (as with Flume), the command should be sent down even if the state of that host component is already stopped. The refactor skips this case.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/controller/AmbariManagementControllerImpl.java 59544dd
> ambari-server/src/test/java/org/apache/ambari/server/controller/AmbariManagementControllerTest.java fe2023e
>
> Diff: https://reviews.apache.org/r/24742/diff/
>
>
> Testing
> -------
>
> Added test case
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 19:51.784s
> [INFO] Finished at: Fri Aug 15 11:50:31 EDT 2014
> [INFO] Final Memory: 19M/123M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nate Cole
>
>
Re: Review Request 24742: Flume: stopping agent doesn't work on host
with multiple agents
Posted by Nate Cole <nc...@hortonworks.com>.
> On Aug. 15, 2014, 12:56 p.m., Sid Wagle wrote:
> > Change looks good, however can you quickly expalin why Flume needs to be STOPPED even though it is already in INSTALLED state?
The issue is that the code that makes flume report as INSTALLED is based on even one single flume agent being down. The way we shutdown a specific flume agent on a host is to provide the name in a special token in the request that gets passed to the agent. So for this patch, if we find this special token (which is really only the component name), we still need to send the request to the host even though the states are the same.
- Nate
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24742/#review50735
-----------------------------------------------------------
On Aug. 15, 2014, 12:41 p.m., Nate Cole wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/24742/
> -----------------------------------------------------------
>
> (Updated Aug. 15, 2014, 12:41 p.m.)
>
>
> Review request for Ambari, Srimanth Gunturi and Sid Wagle.
>
>
> Bugs: AMBARI-6876
> https://issues.apache.org/jira/browse/AMBARI-6876
>
>
> Repository: ambari
>
>
> Description
> -------
>
> Due to a refactor, the code that picked up state changes for targeted host components was no longer getting invoked. This patch addresses that.
>
> In the case where you issue a STOP on a host component and you supply information for the host component (as with Flume), the command should be sent down even if the state of that host component is already stopped. The refactor skips this case.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/controller/AmbariManagementControllerImpl.java 59544dd
> ambari-server/src/test/java/org/apache/ambari/server/controller/AmbariManagementControllerTest.java fe2023e
>
> Diff: https://reviews.apache.org/r/24742/diff/
>
>
> Testing
> -------
>
> Added test case
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 19:51.784s
> [INFO] Finished at: Fri Aug 15 11:50:31 EDT 2014
> [INFO] Final Memory: 19M/123M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nate Cole
>
>
Re: Review Request 24742: Flume: stopping agent doesn't work on host
with multiple agents
Posted by Sid Wagle <sw...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24742/#review50735
-----------------------------------------------------------
Change looks good, however can you quickly expalin why Flume needs to be STOPPED even though it is already in INSTALLED state?
- Sid Wagle
On Aug. 15, 2014, 4:41 p.m., Nate Cole wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/24742/
> -----------------------------------------------------------
>
> (Updated Aug. 15, 2014, 4:41 p.m.)
>
>
> Review request for Ambari, Srimanth Gunturi and Sid Wagle.
>
>
> Bugs: AMBARI-6876
> https://issues.apache.org/jira/browse/AMBARI-6876
>
>
> Repository: ambari
>
>
> Description
> -------
>
> Due to a refactor, the code that picked up state changes for targeted host components was no longer getting invoked. This patch addresses that.
>
> In the case where you issue a STOP on a host component and you supply information for the host component (as with Flume), the command should be sent down even if the state of that host component is already stopped. The refactor skips this case.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/controller/AmbariManagementControllerImpl.java 59544dd
> ambari-server/src/test/java/org/apache/ambari/server/controller/AmbariManagementControllerTest.java fe2023e
>
> Diff: https://reviews.apache.org/r/24742/diff/
>
>
> Testing
> -------
>
> Added test case
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 19:51.784s
> [INFO] Finished at: Fri Aug 15 11:50:31 EDT 2014
> [INFO] Final Memory: 19M/123M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nate Cole
>
>
Re: Review Request 24742: Flume: stopping agent doesn't work on host
with multiple agents
Posted by Sid Wagle <sw...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/24742/#review50737
-----------------------------------------------------------
Ship it!
Ship It!
- Sid Wagle
On Aug. 15, 2014, 4:41 p.m., Nate Cole wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/24742/
> -----------------------------------------------------------
>
> (Updated Aug. 15, 2014, 4:41 p.m.)
>
>
> Review request for Ambari, Srimanth Gunturi and Sid Wagle.
>
>
> Bugs: AMBARI-6876
> https://issues.apache.org/jira/browse/AMBARI-6876
>
>
> Repository: ambari
>
>
> Description
> -------
>
> Due to a refactor, the code that picked up state changes for targeted host components was no longer getting invoked. This patch addresses that.
>
> In the case where you issue a STOP on a host component and you supply information for the host component (as with Flume), the command should be sent down even if the state of that host component is already stopped. The refactor skips this case.
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/controller/AmbariManagementControllerImpl.java 59544dd
> ambari-server/src/test/java/org/apache/ambari/server/controller/AmbariManagementControllerTest.java fe2023e
>
> Diff: https://reviews.apache.org/r/24742/diff/
>
>
> Testing
> -------
>
> Added test case
>
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 19:51.784s
> [INFO] Finished at: Fri Aug 15 11:50:31 EDT 2014
> [INFO] Final Memory: 19M/123M
> [INFO] ------------------------------------------------------------------------
>
>
> Thanks,
>
> Nate Cole
>
>