You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@ambari.apache.org by Jonathan Hurley <jh...@hortonworks.com> on 2017/03/10 22:09:55 UTC
Review Request 57521: HOU wizard wrongly shows hosts with only client
components at the start of 'Upgrade Hosts' group
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57521/
-----------------------------------------------------------
Review request for Ambari, Alejandro Fernandez, Dmytro Grinenko, and Robert Levas.
Bugs: AMBARI-20401
https://issues.apache.org/jira/browse/AMBARI-20401
Repository: ambari
Description
-------
- Deploy HDP-2.5.3 cluster with Ambari-2.5.0.1 (cluster topology is: 2 hosts with only masters, 2 hosts with only slaves, 1 host with only clients)
- Issued below call to start Host Ordered Upgrade to HDP-2.5.4 (note that host named 'bugbash-cluster-2-2.openstacklocal' - third in the list has only client components)
In the HOU wizard in Ambari web UI, when the upgrade reaches 'Upgrade All Hosts' group, there is already an entry for 'Stop on bugbash-cluster-2-2.openstacklocal'
- There is a bug deep in the ambari orchestration logic which causes a request to hang in a {{PENDING}} state if the stage numbers are not directly sequential (ie 1, 2, 3 vs 1, 3)
- HOU is exposing this bug by trying to schedule a stage with no tasks (which is wrong anyway)
It's way to dangerous to fix the first problem. The 2nd problem is now fixed and I'm putting it up for review. Therefore, I think this bug should remain scoped for 2.5.0.1 and I can open a new bug for 2.5.1 for the deeper issue. Sound good?
Diffs
-----
ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderGrouping.java 7734731
ambari-server/src/test/java/org/apache/ambari/server/controller/internal/UpgradeResourceProviderTest.java 62e55c4
Diff: https://reviews.apache.org/r/57521/diff/1/
Testing
-------
PENDING...
Thanks,
Jonathan Hurley
Re: Review Request 57521: HOU wizard wrongly shows hosts with only
client components at the start of 'Upgrade Hosts' group
Posted by Robert Levas <rl...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57521/#review168687
-----------------------------------------------------------
Ship it!
Ship It!
- Robert Levas
On March 10, 2017, 5:09 p.m., Jonathan Hurley wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57521/
> -----------------------------------------------------------
>
> (Updated March 10, 2017, 5:09 p.m.)
>
>
> Review request for Ambari, Alejandro Fernandez, Dmytro Grinenko, and Robert Levas.
>
>
> Bugs: AMBARI-20401
> https://issues.apache.org/jira/browse/AMBARI-20401
>
>
> Repository: ambari
>
>
> Description
> -------
>
> - Deploy HDP-2.5.3 cluster with Ambari-2.5.0.1 (cluster topology is: 2 hosts with only masters, 2 hosts with only slaves, 1 host with only clients)
> - Issued below call to start Host Ordered Upgrade to HDP-2.5.4 (note that host named 'bugbash-cluster-2-2.openstacklocal' - third in the list has only client components)
>
> In the HOU wizard in Ambari web UI, when the upgrade reaches 'Upgrade All Hosts' group, there is already an entry for 'Stop on bugbash-cluster-2-2.openstacklocal'
>
> - There is a bug deep in the ambari orchestration logic which causes a request to hang in a {{PENDING}} state if the stage numbers are not directly sequential (ie 1, 2, 3 vs 1, 3)
> - HOU is exposing this bug by trying to schedule a stage with no tasks (which is wrong anyway)
>
> It's way to dangerous to fix the first problem. The 2nd problem is now fixed and I'm putting it up for review. Therefore, I think this bug should remain scoped for 2.5.0.1 and I can open a new bug for 2.5.1 for the deeper issue. Sound good?
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderGrouping.java 7734731
> ambari-server/src/test/java/org/apache/ambari/server/controller/internal/UpgradeResourceProviderTest.java 62e55c4
>
>
> Diff: https://reviews.apache.org/r/57521/diff/1/
>
>
> Testing
> -------
>
> PENDING...
>
>
> Thanks,
>
> Jonathan Hurley
>
>
Re: Review Request 57521: HOU wizard wrongly shows hosts with only
client components at the start of 'Upgrade Hosts' group
Posted by Jonathan Hurley <jh...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57521/
-----------------------------------------------------------
(Updated March 10, 2017, 8:29 p.m.)
Review request for Ambari, Alejandro Fernandez, Dmytro Grinenko, and Robert Levas.
Bugs: AMBARI-20401
https://issues.apache.org/jira/browse/AMBARI-20401
Repository: ambari
Description (updated)
-------
- Deploy HDP-2.5.3 cluster with Ambari-2.5 (cluster topology is: 2 hosts with only masters, 2 hosts with only slaves, 1 host with only clients)
- Issued below call to start Host Ordered Upgrade to HDP-2.5.4 (note that host named 'bugbash-cluster-2-2.openstacklocal' - third in the list has only client components)
In the HOU wizard in Ambari web UI, when the upgrade reaches 'Upgrade All Hosts' group, there is already an entry for 'Stop on bugbash-cluster-2-2.openstacklocal'
- There is a bug deep in the ambari orchestration logic which causes a request to hang in a {{PENDING}} state if the stage numbers are not directly sequential (ie 1, 2, 3 vs 1, 3)
- HOU is exposing this bug by trying to schedule a stage with no tasks (which is wrong anyway)
It's way to dangerous to fix the first problem. The 2nd problem is now fixed and I'm putting it up for review. Therefore, I think this bug should remain scoped for 2.5.0.1 and I can open a new bug for 2.5.1 for the deeper issue. Sound good?
Diffs
-----
ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderGrouping.java 7734731
ambari-server/src/test/java/org/apache/ambari/server/controller/internal/UpgradeResourceProviderTest.java 62e55c4
Diff: https://reviews.apache.org/r/57521/diff/1/
Testing (updated)
-------
Results :
Tests run: 4943, Failures: 0, Errors: 0, Skipped: 39
Thanks,
Jonathan Hurley
Re: Review Request 57521: HOU wizard wrongly shows hosts with only
client components at the start of 'Upgrade Hosts' group
Posted by Alejandro Fernandez <af...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/57521/#review168682
-----------------------------------------------------------
Ship it!
Ship It!
- Alejandro Fernandez
On March 10, 2017, 10:09 p.m., Jonathan Hurley wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/57521/
> -----------------------------------------------------------
>
> (Updated March 10, 2017, 10:09 p.m.)
>
>
> Review request for Ambari, Alejandro Fernandez, Dmytro Grinenko, and Robert Levas.
>
>
> Bugs: AMBARI-20401
> https://issues.apache.org/jira/browse/AMBARI-20401
>
>
> Repository: ambari
>
>
> Description
> -------
>
> - Deploy HDP-2.5.3 cluster with Ambari-2.5.0.1 (cluster topology is: 2 hosts with only masters, 2 hosts with only slaves, 1 host with only clients)
> - Issued below call to start Host Ordered Upgrade to HDP-2.5.4 (note that host named 'bugbash-cluster-2-2.openstacklocal' - third in the list has only client components)
>
> In the HOU wizard in Ambari web UI, when the upgrade reaches 'Upgrade All Hosts' group, there is already an entry for 'Stop on bugbash-cluster-2-2.openstacklocal'
>
> - There is a bug deep in the ambari orchestration logic which causes a request to hang in a {{PENDING}} state if the stage numbers are not directly sequential (ie 1, 2, 3 vs 1, 3)
> - HOU is exposing this bug by trying to schedule a stage with no tasks (which is wrong anyway)
>
> It's way to dangerous to fix the first problem. The 2nd problem is now fixed and I'm putting it up for review. Therefore, I think this bug should remain scoped for 2.5.0.1 and I can open a new bug for 2.5.1 for the deeper issue. Sound good?
>
>
> Diffs
> -----
>
> ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/HostOrderGrouping.java 7734731
> ambari-server/src/test/java/org/apache/ambari/server/controller/internal/UpgradeResourceProviderTest.java 62e55c4
>
>
> Diff: https://reviews.apache.org/r/57521/diff/1/
>
>
> Testing
> -------
>
> PENDING...
>
>
> Thanks,
>
> Jonathan Hurley
>
>