ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alejandro Fernandez" <afernan...@hortonworks.com>
Subject Re: Review Request 28753: Rolling Upgrade: ClusterStackVersion should display only one repository version
Date Mon, 08 Dec 2014 18:19:41 GMT


> On Dec. 5, 2014, 7:51 p.m., Alejandro Fernandez wrote:
> > ambari-server/src/main/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProvider.java,
line 373
> > <https://reviews.apache.org/r/28753/diff/3/?file=783802#file783802line373>
> >
> >     The initial state should be UPGRADING since a stack version cannot jump directly
into INSTALLED.
> >     
> >     If the UI is using INSTALLED as a way to create dummy data, I suggest folks
insert the data manually through the DB so that this code can lock-down.
> 
> Dmitro Lisnichenko wrote:
>     It has been done to help UI folks. 
>     Also, according to current design, new cluster stack version should be in INIT state.

I just commented on the design doc. We can discuss the states since I believe that page 11
of the doc used them as an example of the endpoint and not as an exhaustive definition of
what they were. In my opinion, only CURRENT, INSTALLING, should be init states.
Exactly one stack version must be in a CURRENT state at all times. A new stack version should
transition from INSTALLING->INSTALLED as part of the repo distribution and yum install,
then followed by the actual rolling upgrade through UPGRADING->UPGRADED, and once the user
confirms the entire operation, transition to CURRENT.


- Alejandro


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/28753/#review64038
-----------------------------------------------------------


On Dec. 8, 2014, 6:11 p.m., Dmitro Lisnichenko wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/28753/
> -----------------------------------------------------------
> 
> (Updated Dec. 8, 2014, 6:11 p.m.)
> 
> 
> Review request for Ambari, Alejandro Fernandez and Nate Cole.
> 
> 
> Bugs: AMBARI-8549
>     https://issues.apache.org/jira/browse/AMBARI-8549
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> for now cluster stack version resource contains all created repository versions: (repo
version filtering does not work correctly)
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/api/services/ClusterStackVersionService.java
97d3bde 
>   ambari-server/src/main/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProvider.java
a12bd29 
>   ambari-server/src/main/java/org/apache/ambari/server/controller/internal/HostStackVersionResourceProvider.java
1b3eef8 
>   ambari-server/src/main/java/org/apache/ambari/server/controller/internal/RepositoryVersionResourceProvider.java
c031ac1 
>   ambari-server/src/main/java/org/apache/ambari/server/state/Cluster.java 125b71b 
>   ambari-server/src/main/java/org/apache/ambari/server/state/cluster/ClusterImpl.java
5a9731a 
>   ambari-server/src/test/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProviderTest.java
0e8269f 
> 
> Diff: https://reviews.apache.org/r/28753/diff/
> 
> 
> Testing
> -------
> 
> [INFO] Reactor Summary:
> [INFO] 
> [INFO] Ambari Views ...................................... SUCCESS [3.225s]
> [INFO] Ambari Server ..................................... SUCCESS [27:27.658s]
> [INFO] Ambari Agent ...................................... SUCCESS [8.468s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 27:40.036s
> [INFO] Finished at: Fri Dec 05 20:29:41 FET 2014
> [INFO] Final Memory: 33M/323M
> [INFO] ------------------------------------------------------------------------
> 
> 
> Thanks,
> 
> Dmitro Lisnichenko
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message