> On Jan. 17, 2015, 1:34 a.m., Ben Mahler wrote:
> > Just curious, what happens to the orphans if you don't kill them? Was there 
> > a ticket for this?
> 
> Timothy Chen wrote:
>     the orphans remains untouched. there is a jira ticket for adding this 
> flag, i can fimd it later once im next to a computer
> 
> Ben Mahler wrote:
>     I understood that part :)
>     
>     But why is it ok to leave orphans untouched? Sounds like a bug to me.. is 
> there some context I'm missing here?
> 
> Timothy Chen wrote:
>     I think the context is that sometimes it's not desirable to remove all 
> orphans on recovery, especially when the discovery mechanism that a task is 
> launched by Mesos currently is looking for Docker containers with a mesos- 
> prefix (future going to be mesos-{slave_id} which is safer).
>     We want to leave this optionally so if users like to keep the containers 
> and want to have their own recovery or gc plan we can let them do so.

Btw this flag is on  by default so unless users really want to it's always 
killing orphan docker container tasks.


- Timothy


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/29328/#review68532
-----------------------------------------------------------


On Jan. 17, 2015, 1:26 a.m., Timothy Chen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/29328/
> -----------------------------------------------------------
> 
> (Updated Jan. 17, 2015, 1:26 a.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Bernd Mathiske.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Add option to disable docker containerizer killing orphans
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/docker.cpp 5f4b4ce 
>   src/slave/flags.hpp a4498e6 
> 
> Diff: https://reviews.apache.org/r/29328/diff/
> 
> 
> Testing
> -------
> 
> 
> make check
> 
> 
> Thanks,
> 
> Timothy Chen
> 
>

Reply via email to