mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Srinivas Brahmaroutu <srbra...@us.ibm.com>
Subject Re: Review Request 49348: Added implementation to Appc Runtime Isolator.
Date Tue, 19 Jul 2016 21:12:00 GMT


> On July 7, 2016, 6:38 a.m., Guangya Liu wrote:
> > src/slave/containerizer/mesos/isolators/appc/runtime.cpp, line 265
> > <https://reviews.apache.org/r/49348/diff/6/?file=1437512#file1437512line265>
> >
> >     Seems you are losing the logic to handle logic in row 2 when there are arguments?
> 
> Gilbert Song wrote:
>     The logic seems fine to me here. That case is covered.
> 
> Guangya Liu wrote:
>     Seems this will only cover line 1 but not line 2, comments?
> 
> Gilbert Song wrote:
>     We dont need to do anything special to line 2. Could you verify? I can be wrong.

@gyliu I fixed the comment. When command is absent, we first set exec[0] as the command and
then we replace arguments with exec[1]... only when they are not present. 
I do not want to append the args after exec[1]... or vice versa. More common use case where
first argument to a executable is the ip address then command exec[1] set in image as -host=localhost
which user likely to replace with proper ip.


> On July 7, 2016, 6:38 a.m., Guangya Liu wrote:
> > src/slave/containerizer/mesos/isolators/appc/runtime.cpp, lines 215-219
> > <https://reviews.apache.org/r/49348/diff/6/?file=1437512#file1437512line215>
> >
> >     So for the case of sh=0,value=0,argv=1,Exec=1, what about the value of `Exec[1]...`
etc? Should not it be `./Exec[0] Exec[1] ... argv`
> 
> Gilbert Song wrote:
>     Thanks Guangya. Let's keep this open. 
>     
>     It depends on whether or not users have the ability to overwrite.
> 
> Guangya Liu wrote:
>     Yes, but I thin at least we need to clarify the behaviro in comments and the document
if there are multiple exec.
> 
> Gilbert Song wrote:
>     Guangya, I think it is the right semantic to have the users able to overwrite. Thanks
for deep dive.
>     
>     @Srini, could you add comments on why we do it this way?

I have added more information into the comments above the table.


- Srinivas


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/49348/#review141133
-----------------------------------------------------------


On July 19, 2016, 9:11 p.m., Srinivas Brahmaroutu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/49348/
> -----------------------------------------------------------
> 
> (Updated July 19, 2016, 9:11 p.m.)
> 
> 
> Review request for mesos.
> 
> 
> Bugs: MESOS-4778
>     https://issues.apache.org/jira/browse/MESOS-4778
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added implementation to Appc Runtime Isolator.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/mesos/isolators/appc/runtime.cpp 19c68e8a25d9ceedc5dfd562e287d6b6a56a9d3a

> 
> Diff: https://reviews.apache.org/r/49348/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Srinivas Brahmaroutu
> 
>


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