You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Anand Mazumdar <ma...@gmail.com> on 2015/09/30 05:39:45 UTC
Review Request 38875: Moved the framework/executor exists check before
Subscribe call in Agent
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/
-----------------------------------------------------------
Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
Bugs: MESOS-3515
https://issues.apache.org/jira/browse/MESOS-3515
Repository: mesos
Description
-------
Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
Diffs
-----
src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
Diff: https://reviews.apache.org/r/38875/diff/
Testing
-------
make check
Thanks,
Anand Mazumdar
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Anand Mazumdar <ma...@gmail.com>.
> On Sept. 30, 2015, 5 a.m., Guangya Liu wrote:
> > src/slave/http.cpp, line 304
> > <https://reviews.apache.org/r/38875/diff/1/?file=1087515#file1087515line304>
> >
> > Can you please add some comments for when this if condition will be hit? I see that the scheduler API is not handling such logic.
It should be : https://github.com/apache/mesos/blob/master/src/master/http.cpp#L465
The only difference here being that we also need to explicitly check for the state to be `REGISTERING` as the agent creates an executor object when it receives a `RunTaskMessage` from the master
- Anand
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review101074
-----------------------------------------------------------
On Sept. 30, 2015, 3:39 a.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated Sept. 30, 2015, 3:39 a.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Guangya Liu <gy...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review101074
-----------------------------------------------------------
src/slave/http.cpp (line 292)
<https://reviews.apache.org/r/38875/#comment158388>
Can you please add some comments for when this if condition will be hit? I see that the scheduler API is not handling such logic.
- Guangya Liu
On 九月 30, 2015, 3:39 a.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated 九月 30, 2015, 3:39 a.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review103959
-----------------------------------------------------------
Ship it!
Ship It!
- Ben Mahler
On Oct. 24, 2015, 6:43 p.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated Oct. 24, 2015, 6:43 p.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp 80bda34e2b344b96d5d2f9c1fecd36046f67ab49
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Anand Mazumdar <ma...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/
-----------------------------------------------------------
(Updated Oct. 24, 2015, 6:43 p.m.)
Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
Changes
-------
Review comments
Bugs: MESOS-3515
https://issues.apache.org/jira/browse/MESOS-3515
Repository: mesos
Description
-------
Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
Diffs (updated)
-----
src/slave/http.cpp 80bda34e2b344b96d5d2f9c1fecd36046f67ab49
src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
Diff: https://reviews.apache.org/r/38875/diff/
Testing
-------
make check
Thanks,
Anand Mazumdar
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Anand Mazumdar <ma...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/
-----------------------------------------------------------
(Updated Oct. 24, 2015, 6:37 p.m.)
Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
Changes
-------
Modified to use the `SchedulerDriver` instead of manually sending a `KillTaskMessage`
Bugs: MESOS-3515
https://issues.apache.org/jira/browse/MESOS-3515
Repository: mesos
Description
-------
Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
Diffs (updated)
-----
src/slave/http.cpp 80bda34e2b344b96d5d2f9c1fecd36046f67ab49
src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
Diff: https://reviews.apache.org/r/38875/diff/
Testing
-------
make check
Thanks,
Anand Mazumdar
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Anand Mazumdar <ma...@gmail.com>.
> On Oct. 5, 2015, 11:57 p.m., Isabel Jimenez wrote:
> > src/slave/http.cpp, line 310
> > <https://reviews.apache.org/r/38875/diff/1/?file=1087515#file1087515line310>
> >
> > Could you please change the syntax for every case to be consistent with `: {` through this switch?
>
> Guangya Liu wrote:
> Does the "{" needed here?
I can make the change. Bit of an overkill though, as these switches would not be stubs quite soon and have a `{`:
There is already a patch out for Call::MESSAGE: https://reviews.apache.org/r/38899
MESOS-3476 is already being worked on.
- Anand
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review101562
-----------------------------------------------------------
On Sept. 30, 2015, 3:39 a.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated Sept. 30, 2015, 3:39 a.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Guangya Liu <gy...@gmail.com>.
> On 十月 5, 2015, 11:57 p.m., Isabel Jimenez wrote:
> > src/slave/http.cpp, line 310
> > <https://reviews.apache.org/r/38875/diff/1/?file=1087515#file1087515line310>
> >
> > Could you please change the syntax for every case to be consistent with `: {` through this switch?
Does the "{" needed here?
- Guangya
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review101562
-----------------------------------------------------------
On 九月 30, 2015, 3:39 a.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated 九月 30, 2015, 3:39 a.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>
Re: Review Request 38875: Moved the framework/executor exists check
before Subscribe call in Agent
Posted by Isabel Jimenez <co...@isabeljimenez.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38875/#review101562
-----------------------------------------------------------
src/slave/http.cpp (line 298)
<https://reviews.apache.org/r/38875/#comment158997>
Could you please change the syntax for every case to be consistent with `: {` through this switch?
- Isabel Jimenez
On Sept. 30, 2015, 3:39 a.m., Anand Mazumdar wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38875/
> -----------------------------------------------------------
>
> (Updated Sept. 30, 2015, 3:39 a.m.)
>
>
> Review request for mesos, Ben Mahler, Isabel Jimenez, and Vinod Kone.
>
>
> Bugs: MESOS-3515
> https://issues.apache.org/jira/browse/MESOS-3515
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Refactored the `api/v1/executor` endpoint to do executor/framework exists(`NULL`) checks before `Subscribe` since the slave explicitly knows the `frameworkID/executorID` when it launches it. Also modified tests to first spawn a dummy `cat` process and then make a `Subscribe` call to ensure that the Agent knows about the Framework/Executor and not reject it via `BadRequest` due to the `NULL` checks.
>
>
> Diffs
> -----
>
> src/slave/http.cpp f9cf7bbe81b7fe9637de9a8d66329c16a7e1a89b
> src/tests/executor_http_api_tests.cpp 31938c295367686d444f8a8aa2c43d2696b481ec
>
> Diff: https://reviews.apache.org/r/38875/diff/
>
>
> Testing
> -------
>
> make check
>
>
> Thanks,
>
> Anand Mazumdar
>
>