Hi guys ,

I would like to know if there are plans to integrate sidekiq queue soon ? That would be really awesome .

best,
shashi

The make release target should verify the release version is correct. This is to ensure the correct image gets pushed to DockerHub and that the metadata is accurate.

Hi,
I recently came into a problem using docker 1.12.1 services and swarm mode.
When I want to start a replicated service for a local built image, I get "image not found" errors inside docker logs on the swarm workers. I believe this happens due to the fact that the image was locally built only on master, thus the other workers can't find the local image in order to start the service.
Is there a solution to automatically build an image on all swarm workers ? I know docker compose can build an image and start a service only on the nodes the image was built on.
There is also the solution with image repository, but I want to avoid this, since I am making many image builds and it doesn't make sense to populate a docker repository with all kind of junk data.
Thanks,

I was planning to use this with Swarm for a hackday project I was working on when I realized that it doesn't really support it yet. I ended up writing a quick implementation in python - but I'd rather use this instead because it supports multiple schedulers. My implementation uses swarm that runs haproxy as global front-end services. It uses telegraf to push haproxy connection stats to influxdb. My scaler code fetches these stats from influxdb and compares against service-declared total capacity to determine if the service needs to be scaled up or down. It uses service labels for determining which services need scaling and what's the declared service capaity.

Once swarm support is added, I could start adding haproxy/influxdb support to it.

There are problems starting containers with Docker version 1.12.0-rc2 used by the current Docker for Mac and Docker for Windows betas.

HostConfig has been removed from the Container Start API call. So we need to call Container Create with the HostConfig and then call Container Start. This error is also occurring upstream in the Golang Docker API client we're using.

microscaling_1  | ERRO 09:12:09.807:  Couldn't start container ID e1a7c15c2859 for task remainder: API error (400): {"message":"starting container with HostConfig was deprecated since v1.10 and removed in v1.12"}

While following these build section instructions I get the below main.go errors. I don't intend on using the visualization portion of the queue demo so is there a way to build without the websocket feature? My environment won't have connectivity to that endpoint anyway.

# make build
# _/home/foo/github.com/microscaling/microscaling
./main.go:119: cannot use ws (type *"github.com/microscaling/microscaling/vendor/golang.org/x/net/websocket".Conn) as type *"golang.org/x/net/websocket".Conn in argument to getDemandEngine
./main.go:153: cannot use ws (type *"github.com/microscaling/microscaling/vendor/golang.org/x/net/websocket".Conn) as type *"golang.org/x/net/websocket".Conn in argument to getMonitors
./settings.go:174: cannot use ws (type *"golang.org/x/net/websocket".Conn) as type *"github.com/microscaling/microscaling/vendor/golang.org/x/net/websocket".Conn in argument to serverEngine.NewEngine
./settings.go:185: cannot use ws (type *"golang.org/x/net/websocket".Conn) as type *"github.com/microscaling/microscaling/vendor/golang.org/x/net/websocket".Conn in argument to monitor.NewServerMonitor
make: *** [microscaling] Error 2
#

Would be great to have this available

Hi, with auto-scaling what I mean is, scale the containers alone and not provisioning of host VM infrastructure. Say, we run various different spring boot micro-services in the docker containers. Can this tool auto-scale and schedule of our docker containers based on the basis of CPU or API performance Latency ?

mesos seems to provide load/pending offers metrics so it seems to be possible as well, plus for our case we need to elasticise only some specific racks in mesos

just targeting marathon seems to be a half-measure

yes, this is a FYI anyway, not that someone expects these to be implemented

Hey team,

Very new to DC/OS and ACS, and am wanting to see this demo; however the producer application never launches. This is in a newly created ACS (I have redeployed multiple times with the same issue). Stderr gives the error message below. Let me know what additional information is required to troubleshoot or where I can start?

I1205 08:59:20.094852 44087 exec.cpp:161] Version: 1.0.1
I1205 08:59:20.128491 44093 exec.cpp:236] Executor registered on agent 97657c69-5fa5-4466-9a27-a9e418003f78-S1
I1205 08:59:20.129650 44093 docker.cpp:815] Running docker -H unix:///var/run/docker.sock run --cpu-shares 51 --memory 104857600 -e AZURE_LOGGING_QUEUE_TYPE=AzureStorageQueue -e ANALYZER_KEEP_RUNNING=true -e MARATHON_APP_VERSION=2016-12-05T08:53:13.366Z -e HOST=10.32.0.4 -e MARATHON_APP_RESOURCE_CPUS=0.05 -e AZURE_STORAGE_QUEUE_NAME=microscaling-demo -e MARATHON_APP_RESOURCE_GPUS=0 -e MARATHON_APP_DOCKER_IMAGE=rgardler/acs-logging-test-simulate:latest -e PORT_10001=24755 -e MESOS_TASK_ID=producer.1bcd97b8-bac9-11e6-ba43-70b3d5800001 -e PORT=24755 -e MARATHON_APP_RESOURCE_MEM=100.0 -e SIMULATION_ACTIONS=0 -e SIMULATION_DELAY=1 -e PORTS=24755 -e AZURE_STORAGE_ACCOUNT_NAME=XXXX -e AZURE_STORAGE_ACCOUNT_KEY=XXXX -e MARATHON_APP_RESOURCE_DISK=0.0 -e MARATHON_APP_LABELS= -e SLACK_WEBHOOK=XXXX -e ANALYZER_SLEEP_TIME=0.1 -e MARATHON_APP_ID=/producer -e PORT0=24755 -e LIBPROCESS_IP=10.32.0.4 -e MESOS_SANDBOX=/mnt/mesos/sandbox -e MESOS_CONTAINER_NAME=mesos-97657c69-5fa5-4466-9a27-a9e418003f78-S1.477f2fff-2127-4ae1-bf83-5e60c5bd24dc -v /var/lib/mesos/slave/slaves/97657c69-5fa5-4466-9a27-a9e418003f78-S1/frameworks/97657c69-5fa5-4466-9a27-a9e418003f78-0000/executors/producer.1bcd97b8-bac9-11e6-ba43-70b3d5800001/runs/477f2fff-2127-4ae1-bf83-5e60c5bd24dc:/mnt/mesos/sandbox --net host --name mesos-97657c69-5fa5-4466-9a27-a9e418003f78-S1.477f2fff-2127-4ae1-bf83-5e60c5bd24dc rgardler/acs-logging-test-simulate:latest
WARNING: Your kernel does not support swap limit capabilities, memory limited without swap.
Traceback (most recent call last):
File "src/simulatedLogs.py", line 95, in
simulate()
File "src/simulatedLogs.py", line 40, in simulate
notify.info(msg)
File "/src/notify.py", line 19, in info
send(msg, config.SLACK_INFO_CHANNEL)
File "/src/notify.py", line 13, in send
"channel": "#" + channel,
TypeError: Can't convert 'NoneType' object to str implicitly

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "src/simulatedLogs.py", line 99, in
log.error("Unable to simulate logging", exc_info=True)
TypeError: error() got an unexpected keyword argument 'exc_info'

Hi,

I am doing a microscailing test on DCOS and it seems like the container image on quay.io has a problem.

Pulling repository quay.io/rossf7/microscaling
docker: Error: Status 403 trying to pull repository rossf7/microscaling: "{"error": "Permission Denied"}".

Would you help to fix the issue?

Thanks,

Microbadger works fine for all of my projects, except: https://microbadger.com/images/devilbox/php-fpm

The above project has 44 base tags on push for each version and microbadger does not seem to update it. It ususally said forever that it was creating the project, but now it just shows an empty page.

Otherwise /var/lib/docker will eventually fill up. Set RemoveVolumes to true in RemoveContainerOptions in stopTask.

Any plan to implement AMQP (rabbitmq) as Queue Type?