mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jiang Yan Xu" <...@jxu.me>
Subject Re: Review Request 37236: Added the linux filesystem isolator.
Date Tue, 11 Aug 2015 21:06:01 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37236/#review94832
-----------------------------------------------------------



src/slave/containerizer/isolators/filesystem/linux.hpp (line 38)
<https://reviews.apache.org/r/37236/#comment149712>

    To echo Marco's comment, there is too little API level documentation. At least there should
class-level summary on what this isolator does and how it is different from other file system
isolators.



src/slave/containerizer/isolators/filesystem/linux.cpp (line 267)
<https://reviews.apache.org/r/37236/#comment149473>

    s/relative path/relative paths or relative



src/slave/containerizer/isolators/filesystem/linux.cpp (lines 267 - 269)
<https://reviews.apache.org/r/37236/#comment149728>

    mesos.proto documentation on Volume::container_path and Volume::host_path both require
them to be "Absolute path"s, should it be changed?



src/slave/containerizer/isolators/filesystem/linux.cpp (lines 312 - 314)
<https://reviews.apache.org/r/37236/#comment149719>

    Are these constraints documented somewhere else that the users could more easily see?
mesos.proto seems the reasonable place (aside from a user doc, of course) but I didn't find
it there.



src/slave/containerizer/isolators/filesystem/linux.cpp (lines 318 - 335)
<https://reviews.apache.org/r/37236/#comment149759>

    So this is the case where the target is expected to be within the work dir because a relative
path is used. However the user could use something like
    `../a/b/c` to still end up being outside the work dir. Checking it with `os::realpath()`
seems safer.



src/slave/containerizer/isolators/filesystem/linux.cpp (line 389)
<https://reviews.apache.org/r/37236/#comment149682>

    Ian originally commented above this line, which I think is helpful and we can put it back:
    
    ```
      // Cleanup of container volumes mounts is done automatically by the
      // kernel when the mount namespace is destroyed after the last
      // process terminates.
    ```


- Jiang Yan Xu


On Aug. 10, 2015, 3:13 p.m., Jie Yu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37236/
> -----------------------------------------------------------
> 
> (Updated Aug. 10, 2015, 3:13 p.m.)
> 
> 
> Review request for mesos, Lily Chen, Ian Downes, Jojy Varghese, Timothy Chen, Vinod Kone,
and Jiang Yan Xu.
> 
> 
> Bugs: MESOS-2794
>     https://issues.apache.org/jira/browse/MESOS-2794
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added the linux filesystem isolator.
> 
> Note that the persistent volume part (i.e., update) hasn't been implemented yet.
> 
> This review is derived from https://reviews.apache.org/r/36429/
> 
> Tests are in the subsequent review.
> 
> 
> Diffs
> -----
> 
>   src/Makefile.am 942003149b071a322933e2c085d9122903f69713 
>   src/slave/containerizer/isolators/filesystem/linux.hpp PRE-CREATION 
>   src/slave/containerizer/isolators/filesystem/linux.cpp PRE-CREATION 
>   src/slave/containerizer/mesos/containerizer.cpp 2cbb879888baf6aff76fbd7c1e19027300fb86e3

> 
> Diff: https://reviews.apache.org/r/37236/diff/
> 
> 
> Testing
> -------
> 
> sudo make check
> 
> 
> Thanks,
> 
> Jie Yu
> 
>


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