Re: Can Marathon ensure single instance of a service at any give time?

2016-02-24 Thread Petr Novak
Thanks everybody for the great input. If I understand it correctly it doesn't help in this case, it just blindly restart service somewhere else once it looses heartbeat. Partition doesn't happen only because network failure it can be as simple as JVM "stop the world" with large heap or pretty much

Re: Mesos 0.25 not incresing Staged/Started counters in the UI

2016-02-24 Thread Geoffroy Jabouley
Hi again just checked the /metrics/snapshot endpoint. Staged value is zero. Is this normal? { "allocator\/event_queue_dispatches":0.0, "frameworks\/jenkins\/messages_processed":9.0, "frameworks\/jenkins\/messages_received":9.0, "master\/cpus_percent":0.0958,

limiting nodes to specific frameworks

2016-02-24 Thread Clarke, Trevor
I've got a custom framework running in mesos (0.24.1 for now). It supports failover and I'd like to be able to start the framework daemons (scheduler, etc.) from Marathon so I can automatically handle scaling and restart. I'm running a small cluster where the mesos master is also the primary

Re: Re: Mesos 0.25 not incresing Staged/Started counters in the UI

2016-02-24 Thread Geoffroy Jabouley
Thanks for the clarification. Does staged means "currently in staging state"? In previous versions of Mesos (at least 0.22.1), the Staged value was increased for each staged tasks, to you could tell "X tasks have been executed on the cluster". My point is there is no straightforward way of

Re: Mesos 0.25 not incresing Staged/Started counters in the UI

2016-02-24 Thread haosdent
If you have some tasks which state is not equal to TASK_STAGING, it would become non-zero. On Wed, Feb 24, 2016 at 8:23 PM, Geoffroy Jabouley < geoffroy.jabou...@gmail.com> wrote: > Hi again > > just checked the /metrics/snapshot endpoint. Staged value is zero. Is this > normal? > > { >

Re: Can Marathon ensure single instance of a service at any give time?

2016-02-24 Thread Mauricio Garavaglia
Right, Marathon can't provide uniqueness guarantees. As you said, network partitions are really common in distributed systems and shouldn't be considered edge cases. On Wed, Feb 24, 2016 at 8:49 AM, Petr Novak wrote: > Thanks everybody for the great input. If I understand

Re: limiting nodes to specific frameworks

2016-02-24 Thread Klaus Ma
Static reservation feature will help; but there's a limitation that Marathon can only manage resources from one role. For example, if we use static reservation for two resources group "master" (--resources="cpus(master):16") & "agent" (--resources="cpus(agent):16"), Marathon

Re: Mesos fetcher in dockerized slave

2016-02-24 Thread Shuai Lin
ping @Tim, I think this bug also affects https://issues.apache.org/jira/browse/MESOS-4743 . On Wed, Jan 20, 2016 at 10:20 PM, Shuai Lin wrote: > The testing of this case requires to build a docker image for mesos-slave, > so it seems not practical to add a test case

Re: Re: Mesos 0.25 not incresing Staged/Started counters in the UI

2016-02-24 Thread haosdent
>My point is there is no straightforward way of telling how many tasks had been running on the cluster since it is up. Or am i missing something? I think we could get it from sum up "master/task_*" metrics? On Wed, Feb 24, 2016 at 9:04 PM, Geoffroy Jabouley < geoffroy.jabou...@gmail.com> wrote:

How to deploy a Database cluster

2016-02-24 Thread Alfredo Carneiro
Hello guys, I have been trying to deploy a Galera MariaDB cluster on my Mesos Cluster following this tutorial [1], but I am facing some problems. After I set Mesos-DNS up, I noticed that nodes use their internal cointainer IP addresses to communicate with other nodes, so the other nodes will be

Bangalore Mesos User Group [http://www.meetup.com/Bangalore-Mesos-User-Group/]

2016-02-24 Thread Dhilip Kumar S
Hi All, We are PaaS team from Huawei Technologies Bangalore, India. We are pleased to announce that we have created a Mesos user group for Bangalore. It would be awesome to build an active community around Apache Mesos in Bangalore. Please feel free to join us and spread the word to your

Re: Apache Mesos Community Sync

2016-02-24 Thread Michael Park
Our next community sync will be on Thursday, February 25, 2016 at 3pm PST. To join in person, come to Mesosphere HQ at 88 Stevenson St. and see the reception on the 2nd floor. Please add your agenda items to the Google Doc

Re: How to deploy a Database cluster

2016-02-24 Thread Rad Gruchalski
Alfredo, Not sure how to do this with Calico and Co but you need to investigate LIBPROCESS_ADVERTISE_IP and LIBPROCESS_ADVERTICE_PORT. https://github.com/apache/mesos/blob/master/docs/configuration.md#libprocess-options Basically, what you need to do is, in your container you need to: export