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 2014/04/16 02:34:32 UTC
Review Request 20383: If a non-strict registry is in use,
do not inform frameworks that slaves are lost if they do not re-register
after a failover.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/20383/
-----------------------------------------------------------
Review request for mesos, Benjamin Hindman and Vinod Kone.
Repository: mesos-git
Description
-------
We would like a non-strict registry to be "write-only" in the sense that we do not take "actions" based on the state of the registry.
What this means for recovery is that we should ensure that we don't inform frameworks of lost slaves when they do not re-register after recovery. This preserves the old "stateless" master semantics.
Diffs
-----
src/master/master.cpp 3c3c989543167afb7d368a19a16457ed00e6be0c
Diff: https://reviews.apache.org/r/20383/diff/
Testing
-------
Will follow up with an integration test as I make the LogStorage updates to the testing abstractions.
Unfortunately, this case will be a bit tricky to test since we're expecting the *absence* of a lost message.
Thanks,
Ben Mahler
Re: Review Request 20383: If a non-strict registry is in use,
do not inform frameworks that slaves are lost if they do not re-register
after a failover.
Posted by Vinod Kone <vi...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/20383/#review40582
-----------------------------------------------------------
Ship it!
Ship It!
- Vinod Kone
On April 16, 2014, 12:34 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/20383/
> -----------------------------------------------------------
>
> (Updated April 16, 2014, 12:34 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> We would like a non-strict registry to be "write-only" in the sense that we do not take "actions" based on the state of the registry.
>
> What this means for recovery is that we should ensure that we don't inform frameworks of lost slaves when they do not re-register after recovery. This preserves the old "stateless" master semantics.
>
>
> Diffs
> -----
>
> src/master/master.cpp 3c3c989543167afb7d368a19a16457ed00e6be0c
>
> Diff: https://reviews.apache.org/r/20383/diff/
>
>
> Testing
> -------
>
> Will follow up with an integration test as I make the LogStorage updates to the testing abstractions.
>
> Unfortunately, this case will be a bit tricky to test since we're expecting the *absence* of a lost message.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 20383: If a non-strict registry is in use,
do not inform frameworks that slaves are lost if they do not re-register
after a failover.
Posted by Mesos ReviewBot <de...@mesos.apache.org>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/20383/#review40591
-----------------------------------------------------------
Patch looks great!
Reviews applied: [19857, 20097, 20383]
All tests passed.
- Mesos ReviewBot
On April 16, 2014, 12:34 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/20383/
> -----------------------------------------------------------
>
> (Updated April 16, 2014, 12:34 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> We would like a non-strict registry to be "write-only" in the sense that we do not take "actions" based on the state of the registry.
>
> What this means for recovery is that we should ensure that we don't inform frameworks of lost slaves when they do not re-register after recovery. This preserves the old "stateless" master semantics.
>
>
> Diffs
> -----
>
> src/master/master.cpp 3c3c989543167afb7d368a19a16457ed00e6be0c
>
> Diff: https://reviews.apache.org/r/20383/diff/
>
>
> Testing
> -------
>
> Will follow up with an integration test as I make the LogStorage updates to the testing abstractions.
>
> Unfortunately, this case will be a bit tricky to test since we're expecting the *absence* of a lost message.
>
>
> Thanks,
>
> Ben Mahler
>
>