overall +0 - I'm okay with proceeding with the changelog as the only
deficiency
+1 java tests pass
+1 python test pass
+1 e2e tests pass
+0 changelog is incorrect as wfarner notes
On Tue, May 5, 2015 at 3:36 PM, Bill Farner wrote:
> -1 overall
>
> +1 checksums and signatures are good
> +1 uni
overall +0 - I'm okay with proceeding with the changelog as the only
deficiency
+1 java tests pass
+1 python test pass
+1 e2e tests pass
+0 changelog is incorrect as wfarner notes
On Tue, May 5, 2015 at 3:36 PM, Bill Farner wrote:
> -1 overall
>
> +1 checksums and signatures are good
> +1 unit
-1 overall
+1 checksums and signatures are good
+1 unit tests and end-to-end tests pass
-1 changelog contains tickets that were closed as won't fix. i came up
with a JIRA query [1] that seems to identify a bunch of these
[1]
https://issues.apache.org/jira/issues/?jql=project%20%3D%20AURORA%20AND
Hi,
I have put together a design proposal for improving health-enabled job
update performance. Please, review and leave your comments:
https://docs.google.com/document/d/1ZdgW8S4xMhvKW7iQUX99xZm10NXSxEWR0a-21FP5d94/edit
Thanks,
Maxim
The defaults are reasonable. I would not advise increasing
max_leading_duration dramatically, as the recovery time grows linearly with
the number of log writes due to MESOS-184 [1].
[1] https://issues.apache.org/jira/browse/MESOS-184
-=Bill
On Tue, May 5, 2015 at 12:29 PM, Bhuvan Arumugam wro
What's the recommended setting for following parameters in a 5 scheduler quorum?
-max_leading_duration (default: 24 hours)
-backup_interval (default: 1 hour)
--
Regards,
Bhuvan Arumugam
www.livecipher.com
Yeah, thermos need mesos, svn and apr libraries. If you manage all
docker images used to launch the job, you can install these
dependencies inside the image. If you don't have control over images
used to launch jobs, you can install these dependencies in slave
system and configure aurora-scheduler
there should not be any drift, this is just some cleanup based on us
learning from our first couple releases. Jenkins can not cut release as it
requires committing and signing, but jenkins does currently use the same
build script as this uses to test.
-Jake
On Tue, May 5, 2015 at 2:25 PM, Zamee
To prevent this drift of our release and verification infrastructure,
should we have CI cut a release tarball and try to verify it with this
script?
On Tue, May 5, 2015 at 11:14 AM, Bill Farner wrote:
>
> ---
> This is an automatically gene
I would add that the executor you are running in the docker container should
match the base is it was built for. That is to say a centos built thermos
executor will have issues in a Debian based container and vice versa.
Sent from my iPad
> On May 5, 2015, at 12:45 PM, Steve Niemitz wrote:
>
You need libsvn-dev and libapr1-dev (on ubuntu) installed inside the
container, these are required by the mesos library. Once pesos is ready
these dependencies will no longer be needed in the containers.
On Tue, May 5, 2015 at 12:38 PM, Mauricio Garavaglia <
mauriciogaravag...@gmail.com> wrote:
Hi guys,
What are the dependencies needed in the docker images to run it in aurora?
I added python2.7 but the thermos executor fails for some images with
"thermos_executor.pex: error: Could not load MesosExecutorDriver!" and the
job gets lost.
Are those dependencies documented somewhere? Thanks!
12 matches
Mail list logo
