You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@mesos.apache.org by Joerg Schad <jo...@mesosphere.io> on 2015/08/28 18:39:44 UTC
Review Request 37894: Renamed cgroups::kill to cgroups::signal.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/
-----------------------------------------------------------
Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
Bugs: MESOS-3086
https://issues.apache.org/jira/browse/MESOS-3086
Repository: mesos
Description
-------
As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
Diffs
-----
src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
Diff: https://reviews.apache.org/r/37894/diff/
Testing
-------
Thanks,
Joerg Schad
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Jie Yu <yu...@gmail.com>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
IMO, that naming is confusing, and should be 'signal' and 'install'.
- Jie
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Timothy Chen <tn...@apache.org>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
>
> Jie Yu wrote:
> IMO, that naming is confusing, and should be 'signal' and 'install'.
>
> Joerg Schad wrote:
> This is actually answering a discussion here https://reviews.apache.org/r/36620/#comment152376. And I agree that the original names are weird, but you are right in that we should have an open discussion about this. I will send a mail to dev later on...
>
> Cong Wang wrote:
> Like it or not, it is already a part of POSIX (if not just UNIX), pthread_kill() has the same "problem". It's just too late to change, as it simply breaks every programmer's expectation.
>
> Jie Yu wrote:
> This is cgroups::signal and cgroups::kill (has nothing to do with Posix kill and signal).
>
> It reads more natural: signal a cgroup, kill a cgroup. I don't think it'll confuse the reader.
>
> Are you suggesting that any thing suffixed with 'kill' should be linked to kill(2) in posix? I have to disagree with that.
I think I personally don't think we need to tie with POSIX, just need some comments to clarify what the intent is and I personally like to see this rename as well. Don't want this to sit and get stale again, Jie/Cong is there more to add or can we agree on this change?
- Timothy
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Jie Yu <yu...@gmail.com>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
>
> Jie Yu wrote:
> IMO, that naming is confusing, and should be 'signal' and 'install'.
>
> Joerg Schad wrote:
> This is actually answering a discussion here https://reviews.apache.org/r/36620/#comment152376. And I agree that the original names are weird, but you are right in that we should have an open discussion about this. I will send a mail to dev later on...
>
> Cong Wang wrote:
> Like it or not, it is already a part of POSIX (if not just UNIX), pthread_kill() has the same "problem". It's just too late to change, as it simply breaks every programmer's expectation.
This is cgroups::signal and cgroups::kill (has nothing to do with Posix kill and signal).
It reads more natural: signal a cgroup, kill a cgroup. I don't think it'll confuse the reader.
Are you suggesting that any thing suffixed with 'kill' should be linked to kill(2) in posix? I have to disagree with that.
- Jie
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Ben Mahler <be...@gmail.com>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
>
> Jie Yu wrote:
> IMO, that naming is confusing, and should be 'signal' and 'install'.
>
> Joerg Schad wrote:
> This is actually answering a discussion here https://reviews.apache.org/r/36620/#comment152376. And I agree that the original names are weird, but you are right in that we should have an open discussion about this. I will send a mail to dev later on...
>
> Cong Wang wrote:
> Like it or not, it is already a part of POSIX (if not just UNIX), pthread_kill() has the same "problem". It's just too late to change, as it simply breaks every programmer's expectation.
>
> Jie Yu wrote:
> This is cgroups::signal and cgroups::kill (has nothing to do with Posix kill and signal).
>
> It reads more natural: signal a cgroup, kill a cgroup. I don't think it'll confuse the reader.
>
> Are you suggesting that any thing suffixed with 'kill' should be linked to kill(2) in posix? I have to disagree with that.
>
> Timothy Chen wrote:
> I think I personally don't think we need to tie with POSIX, just need some comments to clarify what the intent is and I personally like to see this rename as well. Don't want this to sit and get stale again, Jie/Cong is there more to add or can we agree on this change?
What's the plan for other utilities, like `os::killtree` and `os::kill`? Jie, what does it mean to "kill" a cgroup, send SIGKILL? Or destroy it completely? I thought that's why we have `cgroups::destroy`..?
- Ben
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Cong Wang <cw...@twopensource.com>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
>
> Jie Yu wrote:
> IMO, that naming is confusing, and should be 'signal' and 'install'.
>
> Joerg Schad wrote:
> This is actually answering a discussion here https://reviews.apache.org/r/36620/#comment152376. And I agree that the original names are weird, but you are right in that we should have an open discussion about this. I will send a mail to dev later on...
Like it or not, it is already a part of POSIX (if not just UNIX), pthread_kill() has the same "problem". It's just too late to change, as it simply breaks every programmer's expectation.
- Cong
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Joerg Schad <jo...@mesosphere.io>.
> On Aug. 28, 2015, 4:45 p.m., Cong Wang wrote:
> > Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
>
> Jie Yu wrote:
> IMO, that naming is confusing, and should be 'signal' and 'install'.
This is actually answering a discussion here https://reviews.apache.org/r/36620/#comment152376. And I agree that the original names are weird, but you are right in that we should have an open discussion about this. I will send a mail to dev later on...
- Joerg
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Cong Wang <cw...@twopensource.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96883
-----------------------------------------------------------
Why? Everyone knows kill(2) sends a signal while signal(2) installs a signal handler...
- Cong Wang
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>
Re: Review Request 37894: Renamed cgroups::kill to cgroups::signal.
Posted by Mesos ReviewBot <re...@mesos.apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37894/#review96897
-----------------------------------------------------------
Patch looks great!
Reviews applied: [37894]
All tests passed.
- Mesos ReviewBot
On Aug. 28, 2015, 4:39 p.m., Joerg Schad wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37894/
> -----------------------------------------------------------
>
> (Updated Aug. 28, 2015, 4:39 p.m.)
>
>
> Review request for mesos, Jie Yu, Till Toenshoff, and Timothy Chen.
>
>
> Bugs: MESOS-3086
> https://issues.apache.org/jira/browse/MESOS-3086
>
>
> Repository: mesos
>
>
> Description
> -------
>
> As this function is actually just sending the signal we renamed it. Additionally we want to introduce an actual cgroups::kill() with MESOS-3086.
>
>
> Diffs
> -----
>
> src/linux/cgroups.hpp 204c53038a1ccfa693f4f2293488cff8cdd60835
> src/linux/cgroups.cpp 6ef42ed1bc719f334d1ac6e90919a1bc1840d31f
> src/tests/containerizer/cgroups_tests.cpp 0b171eeb53037f26b7e952830e88e59f1278e7c6
>
> Diff: https://reviews.apache.org/r/37894/diff/
>
>
> Testing
> -------
>
>
> Thanks,
>
> Joerg Schad
>
>