You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Ben Mahler <be...@gmail.com> on 2013/06/12 21:45:53 UTC
Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/
-----------------------------------------------------------
Review request for mesos, Benjamin Hindman, Vinod Kone, and Jiang Yan Xu.
Description
-------
os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
Diffs
-----
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
Diff: https://reviews.apache.org/r/11848/diff/
Testing
-------
This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
Thanks,
Ben Mahler
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/
-----------------------------------------------------------
(Updated June 20, 2013, 3:27 a.m.)
Review request for mesos, Benjamin Hindman and Vinod Kone.
Changes
-------
Rebase, NNFR.
Description
-------
os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
Diffs (updated)
-----
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
Diff: https://reviews.apache.org/r/11848/diff/
Testing
-------
This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
Thanks,
Ben Mahler
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/
-----------------------------------------------------------
(Updated June 19, 2013, 6:49 p.m.)
Review request for mesos, Benjamin Hindman and Vinod Kone.
Changes
-------
Rebase. NNFR.
Description
-------
os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
Diffs (updated)
-----
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
Diff: https://reviews.apache.org/r/11848/diff/
Testing
-------
This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
Thanks,
Ben Mahler
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/
-----------------------------------------------------------
(Updated June 19, 2013, 6:32 p.m.)
Review request for mesos, Benjamin Hindman and Vinod Kone.
Changes
-------
Rebase, NNFR.
Description
-------
os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
Diffs (updated)
-----
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
Diff: https://reviews.apache.org/r/11848/diff/
Testing
-------
This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
Thanks,
Ben Mahler
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/
-----------------------------------------------------------
(Updated June 18, 2013, 4:42 a.m.)
Review request for mesos, Benjamin Hindman and Vinod Kone.
Changes
-------
NNFR.
Description
-------
os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
Diffs (updated)
-----
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
Diff: https://reviews.apache.org/r/11848/diff/
Testing
-------
This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
Thanks,
Ben Mahler
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Jiang Yan Xu <ya...@jxu.me>.
> On June 13, 2013, 7:32 p.m., Ben Mahler wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp, line 1118
> > <https://reviews.apache.org/r/11848/diff/1/?file=304301#file304301line1118>
> >
> > Is there the need to check for zombie processes? The semantics of "alive" would suggest that zombies are not considered alive ;)
> >
> > I'd be happy to add another function should we require different semantics at some point, but it seems at the moment we simply want an "alive"ness check?
>
> Jiang Yan Xu wrote:
> Sorry I overlooked this comment. In Reaper::monitor(), if exists(pid) is false we can directly return None() because there is nothing to monitor; if exists(pid) is true but alive(pid) is false we can still potentially reap the exit status. I guess this is where the semantical differences can be leveraged.
>
> Jiang Yan Xu wrote:
> So would you prefer a Try<bool> exits(pid) and have Try<bool> alive(pid) return an error if the process doesn't exist (then we won't be able to differentiate that from other errors, e.g. no permission, during liveness check)?
>
> On Linux the other errors for os::alive should rarely happen.
>
> Vinod Kone wrote:
> +1 to exists().
>
> @Yan: Not sure I understand the unable to differentiate part, when the functions returns a Try. Can you expand on your comment?
We have these states for alive():
- process alive: return true
- process terminated but not reaped: return false
- process nonexistent: return Error()
- We don't have permission to check the status of the pid: return Error()
So the bottom two are not differentiable right?
Now since reaper is the only user of the functions perhaps removing them and merging the logic into reaper itself is reasonable too. What do you think?
- Jiang Yan
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21871
-----------------------------------------------------------
On June 20, 2013, 3:27 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 20, 2013, 3:27 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Vinod Kone <vi...@gmail.com>.
> On June 13, 2013, 7:32 p.m., Ben Mahler wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp, line 1118
> > <https://reviews.apache.org/r/11848/diff/1/?file=304301#file304301line1118>
> >
> > Is there the need to check for zombie processes? The semantics of "alive" would suggest that zombies are not considered alive ;)
> >
> > I'd be happy to add another function should we require different semantics at some point, but it seems at the moment we simply want an "alive"ness check?
>
> Jiang Yan Xu wrote:
> Sorry I overlooked this comment. In Reaper::monitor(), if exists(pid) is false we can directly return None() because there is nothing to monitor; if exists(pid) is true but alive(pid) is false we can still potentially reap the exit status. I guess this is where the semantical differences can be leveraged.
>
> Jiang Yan Xu wrote:
> So would you prefer a Try<bool> exits(pid) and have Try<bool> alive(pid) return an error if the process doesn't exist (then we won't be able to differentiate that from other errors, e.g. no permission, during liveness check)?
>
> On Linux the other errors for os::alive should rarely happen.
+1 to exists().
@Yan: Not sure I understand the unable to differentiate part, when the functions returns a Try. Can you expand on your comment?
- Vinod
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21871
-----------------------------------------------------------
On June 20, 2013, 3:27 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 20, 2013, 3:27 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Jiang Yan Xu <ya...@jxu.me>.
> On June 13, 2013, 7:32 p.m., Ben Mahler wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp, line 1118
> > <https://reviews.apache.org/r/11848/diff/1/?file=304301#file304301line1118>
> >
> > Is there the need to check for zombie processes? The semantics of "alive" would suggest that zombies are not considered alive ;)
> >
> > I'd be happy to add another function should we require different semantics at some point, but it seems at the moment we simply want an "alive"ness check?
Sorry I overlooked this comment. In Reaper::monitor(), if exists(pid) is false we can directly return None() because there is nothing to monitor; if exists(pid) is true but alive(pid) is false we can still potentially reap the exit status. I guess this is where the semantical differences can be leveraged.
- Jiang Yan
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21871
-----------------------------------------------------------
On June 20, 2013, 3:27 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 20, 2013, 3:27 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Jiang Yan Xu <ya...@jxu.me>.
> On June 13, 2013, 7:32 p.m., Ben Mahler wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp, line 1118
> > <https://reviews.apache.org/r/11848/diff/1/?file=304301#file304301line1118>
> >
> > Is there the need to check for zombie processes? The semantics of "alive" would suggest that zombies are not considered alive ;)
> >
> > I'd be happy to add another function should we require different semantics at some point, but it seems at the moment we simply want an "alive"ness check?
>
> Jiang Yan Xu wrote:
> Sorry I overlooked this comment. In Reaper::monitor(), if exists(pid) is false we can directly return None() because there is nothing to monitor; if exists(pid) is true but alive(pid) is false we can still potentially reap the exit status. I guess this is where the semantical differences can be leveraged.
So would you prefer a Try<bool> exits(pid) and have Try<bool> alive(pid) return an error if the process doesn't exist (then we won't be able to differentiate that from other errors, e.g. no permission, during liveness check)?
On Linux the other errors for os::alive should rarely happen.
- Jiang Yan
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21871
-----------------------------------------------------------
On June 20, 2013, 3:27 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 20, 2013, 3:27 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21871
-----------------------------------------------------------
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp
<https://reviews.apache.org/r/11848/#comment45170>
Why? If the process does not exist, then the Try will be an error, also we don't log warnings in stout.
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp
<https://reviews.apache.org/r/11848/#comment45172>
Is there the need to check for zombie processes? The semantics of "alive" would suggest that zombies are not considered alive ;)
I'd be happy to add another function should we require different semantics at some point, but it seems at the moment we simply want an "alive"ness check?
- Ben Mahler
On June 12, 2013, 7:45 p.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 12, 2013, 7:45 p.m.)
>
>
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Jiang Yan Xu.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Benjamin Hindman <be...@berkeley.edu>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21950
-----------------------------------------------------------
Ship it!
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp
<https://reviews.apache.org/r/11848/#comment45266>
This shouldn't be here, lets make this an Error please.
- Benjamin Hindman
On June 12, 2013, 7:45 p.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 12, 2013, 7:45 p.m.)
>
>
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Jiang Yan Xu.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Jiang Yan Xu <ya...@jxu.me>.
> On June 12, 2013, 11:12 p.m., Vinod Kone wrote:
> > 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp, line 1118
> > <https://reviews.apache.org/r/11848/diff/1/?file=304301#file304301line1118>
> >
> > Would it make sense for os::process to return a Result instead of a Try?
> >
> > With the current semantics, I'm not sure how to differentiate between a zombie and a reaped process.
Maybe we need an os::exists(pid_t) and an os::alive(pid_t)? Then in order to differentiate the two states we need to call two different methods.
- Jiang Yan
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21823
-----------------------------------------------------------
On June 12, 2013, 7:45 p.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 12, 2013, 7:45 p.m.)
>
>
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Jiang Yan Xu.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request: Added a zombie bit to os::Process,
improved os::alive to handle zombie processes.
Posted by Vinod Kone <vi...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/11848/#review21823
-----------------------------------------------------------
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp
<https://reviews.apache.org/r/11848/#comment45067>
A warning would be nice here.
3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp
<https://reviews.apache.org/r/11848/#comment45073>
Would it make sense for os::process to return a Result instead of a Try?
With the current semantics, I'm not sure how to differentiate between a zombie and a reaped process.
- Vinod Kone
On June 12, 2013, 7:45 p.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/11848/
> -----------------------------------------------------------
>
> (Updated June 12, 2013, 7:45 p.m.)
>
>
> Review request for mesos, Benjamin Hindman, Vinod Kone, and Jiang Yan Xu.
>
>
> Description
> -------
>
> os::alive currently considers zombie processes to be alive, which is less than optimal, as semantically they are "dead" and waiting to be reaped.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/3rdparty/stout/include/stout/os.hpp 1b3fb47d7567b5467fef2a2bb15d5c4a2ea42aa5
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/linux.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/osx.hpp PRE-CREATION
> 3rdparty/libprocess/3rdparty/stout/include/stout/os/process.hpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/11848/diff/
>
>
> Testing
> -------
>
> This was tested through my killtree change, although it would be nice to have tests once we have an abstraction for creating arbitrary process trees.
>
>
> Thanks,
>
> Ben Mahler
>
>