Hybrid application deployments (container/VM/bare metal) in Mesos

Hello, I'm wondering if someone here can help point me some document links about hybrid application deployment in Mesos. The basic idea is that we have some applications in mixed flavors (container, VM, bare metal) and we want to run them on a single cluster. Let me explain by an example.

Re: Slaves not getting registered

sos# tail -f mesos-slave.slave1.invalid-user.log.INFO.20160412-110554.1696 >I0413 03:12:54.532676  1711 group.cpp:519] ZooKeeper session expired >I0413 03:12:58.757953  1715 slave.cpp:4304] Current disk usage 6.44%. Max allowed age: 5.848917453828577days >W0413 03:13:04.539577  17

Re: [Proposal] Remove the default value for agent work_dir

+1 On Tue, Apr 12, 2016 at 9:31 PM, Jie Yu wrote: > +1 > > On Tue, Apr 12, 2016 at 9:29 PM, James Peach wrote: > > > > > > On Apr 12, 2016, at 3:58 PM, Greg Mann wrote: > > > > > > Hey folks! > > > A number of situations have arisen

Re: [Proposal] Remove the default value for agent work_dir

+1 Tim On Wed, Apr 13, 2016 at 5:31 AM, Jie Yu wrote: > +1 > > On Tue, Apr 12, 2016 at 9:29 PM, James Peach wrote: > >> >> > On Apr 12, 2016, at 3:58 PM, Greg Mann wrote: >> > >> > Hey folks! >> > A number of situations have arisen in

Re: Mesos Community Sync 4/7/16 - Notes

Hi Grep, I am located in China. I would love to participate in next meeting. Could you please share me how to join in? appreciated. Regards, Sam On Wed, Apr 13, 2016 at 5:07 AM, Greg Mann wrote: > Hey Mesos community, > Find below the notes from our last Mesos Community

Re: [Proposal] Remove the default value for agent work_dir

+1 On Tue, Apr 12, 2016 at 9:29 PM, James Peach wrote: > > > On Apr 12, 2016, at 3:58 PM, Greg Mann wrote: > > > > Hey folks! > > A number of situations have arisen in which the default value of the > Mesos agent `--work_dir` flag (/tmp/mesos) has caused

Re: [Proposal] Remove the default value for agent work_dir

> On Apr 12, 2016, at 3:58 PM, Greg Mann wrote: > > Hey folks! > A number of situations have arisen in which the default value of the Mesos > agent `--work_dir` flag (/tmp/mesos) has caused problems on systems in which > the automatic cleanup of '/tmp' deletes agent

RE: Slaves not getting registered

ser.log.INFO.20160412-11014 tail: cannot open ‘mesos-master.master1.invalid-user.log.INFO.20160412-11014’ for reading: No such file or directory root@master1:/var/log/mesos# tail -f mesos-master.master1.invalid-user.log.INFO.20160412-11014 mesos-master.master1.invalid-user.log.INFO.20160412-110143.

Re: Slaves not getting registered

Log line format: [IWEF]mmdd hh:mm:ss.uu threadid file:line] msg W0412 11:01:49.024226  3712 authenticator.cpp:511] No credentials provided, authentication requests will be refused   root@master1:/var/log/mesos# tail -f mesos-master.master1.invalid-user.log.INFO.20160412-11014 tail

RE: Slaves not getting registered

cation requests will be refused root@master1:/var/log/mesos# tail -f mesos-master.master1.invalid-user.log.INFO.20160412-11014 tail: cannot open ‘mesos-master.master1.invalid-user.log.INFO.20160412-11014’ for reading: No such file or directory root@master1:/var/log/mesos# tail -f mesos-master.m

Re: [Proposal] Remove the default value for agent work_dir

+1 I have seen this confuse many users of Apache Aurora many times. Eliminating the default will cause operators to select a location with the appropriate persistence properties. On Tue, Apr 12, 2016 at 3:58 PM, Greg Mann wrote: > Hey folks! > A number of situations have

[Proposal] Remove the default value for agent work_dir

Hey folks! A number of situations have arisen in which the default value of the Mesos agent `--work_dir` flag (/tmp/mesos) has caused problems on systems in which the automatic cleanup of '/tmp' deletes agent metadata. To resolve this, we would like to eliminate the default value of the agent

Mesos interconnection among clusters project calico

Hi all i have to set up two mesos clusters. On each cluster i should integrate Project calico in order to distribute tasks among the agents. But these tasks should be sent also from a slave of one cluster to the slave of the other cluster. I know that when i start calico on each slaves, it

[RESULT][VOTE] Release Apache Mesos 0.25.1 (rc4)

Hi all, The vote for Mesos 0.25.1 (rc4) has passed with the following votes. +1 (Binding) -- Benjamin Mahler Vinod Kone Kapil Arya +1 (Non-binding) -- N/A There were no 0 or -1 votes. Please find the release at:

[RESULT][VOTE] Release Apache Mesos 0.24.2 (rc5)

Hi all, The vote for Mesos 0.24.2 (rc5) has passed with the following votes. +1 (Binding) -- * Benjamin Mahler * Vinod Kone * Kapil Arya +1 (Non-binding) -- N/A There were no 0 or -1 votes. Please find the release at:

Re: Package release schedule for Mesosphere Mesos 0.29

As discussed in the previous community sync (notes of which have just been emailed to the list), 0.29.0 is being targeted for mid-may. Mesosphere packages are automatically released (within hours) following the official release. On Tue, Apr 12, 2016 at 1:45 PM, June Taylor wrote:

Re: mesos/kafka issues (org.apache.mesos.Scheduler)

Thanks to those who’ve taken the time to be helpful here, I dug into this a bit more today and verified some things: * When I run the scheduler on the active mesos master and tcpdump traffic to port 5050, I see none. * It does identify its’ IP as the running master at startup. * I do see

Re: Mesos 0.28 SSL in official packages

Mesosphere open-sourced the package-building scripts long ago: https://github.com/mesosphere/mesos-deb-packaging The TeamCity configuration, however, is internal-only, but that wouldn't work for Apache anyway, since Apache doesn't use TeamCity AFAIK. On Tue, Apr 12, 2016 at 11:07 AM, Zameer Manji

Re: Mesos 0.28 SSL in official packages

For the record, I am not a committer on the Apache Mesos project and I do not have the time to contribute packaging tools for the project. I think existing committers who are Mesosphere employees can kick start this effort by asking their employer to contribute the existing tools to the project.

Re: Slaves not getting registered

>Is there a way to find out the number of masters that are present in the environment together through CLI/GUI? There is a command line tool `mesos-resolve` which can show master info in terminal. But I think it may not related to your problem here. If your mesos doesn't depend on zookeeper, it

Re: marathon issue in running a docker container.

>Server running at: http://0.0.0.0:7683 And according to your log, your service is running on 7683 while your configuration use 8080 in portMapping. On Wed, Apr 13, 2016 at 1:25 AM, haosdent wrote: > >If I do telnet or curl, it does not show me any reponse. > Looks wired

Re: marathon issue in running a docker container.

>If I do telnet or curl, it does not show me any reponse. Looks wired here, could you find the status of task is running or other status in mesos webui or marathon webui? And is it possible for you to use `docker ps` to find out the container and use `docker exec` to enter container and check

Re: Mesos 0.28 SSL in official packages

At Mesosphere, we are planning to enable SSL into the nightlies starting sometime later this week. The goal is to have both SSL and non-SSL Mesos packages for Mesos 0.29.0 onwards in the Mesosphere deb/rpm repos. I will send out another email as soon as the stuff is ready for the community. Best,

Re: Mesos 0.28 SSL in official packages

I’d be willing to assist in the effort to have standard packages (and additional packages for modules like net-modules). Steven Borrelli st...@borrelli.org > On Apr 12, 2016, at 11:10 AM, Adam Bordelon wrote: > > We've discussed Apache-built/distributed packages before,

Re: Mesos 0.28 SSL in official packages

We've discussed Apache-built/distributed packages before, and nobody has any objections, but we need somebody to take on the work to get the package builds setup. I believe Vinod had some thoughts on how to get started, but any Apache committer (Zameer?) should have access to builds.apache.org I

Re: Mesos 0.28 SSL in official packages

I heartily agree on both points. While I've found Mesosphere's documentation very helpful, it is often mixed up with the DCOS commercial offering. That may be something we're interested in down the road, but right now we are trying to stand up a relatively small cluster using straight

Re: Slaves not getting registered

Try looking in /var/log/mesos/ at these files: mesos-slave.WARNING, mesos-slave.INFO, mesos-slave.ERROR Thanks, June Taylor System Administrator, Minnesota Population Center University of Minnesota On Tue, Apr 12, 2016 at 4:36 AM, Dick Davies wrote: > There's no

Re: Mesos 0.28 SSL in official packages

FWIW, I quite agree with Zameer's point. That said, I want to make abundantly clear that in my experience the folks at Mesosphere are wonderfully helpful. But what happens if down the road Mesosphere is acquired or there occurs some other event that could represent, if not a conflict of

Re: marathon issue in running a docker container.

I might be wrong here. But I am using marathon-lb package of DCOS which already has haproxy. On Tuesday, 12 April 2016 2:14 PM, Rad Gruchalski wrote: Do you have anything like haproxy for port mappings installed on your Mesos cluster? When using BRIDGE

Re: Slaves not getting registered

There's no mention of a slave there, have a look at the logs on the slaves filesystem and see if it is giving any errors. On 12 April 2016 at 10:17, wrote: > The GUI log shows like this: > > > > I0412 08:45:51.379609 3616 master.cpp:3673] Processing DECLINE

RE: Slaves not getting registered

The GUI log shows like this: I0412 08:45:51.379609 3616 master.cpp:3673] Processing DECLINE call for offers: [ 74f33592-fc48-4066-a59c-977818b4c13c-O282 ] for framework 74f33592-fc48-4066-a59c-977818b4c13c-0001 (chronos-2.4.0) at

Re: SharedFilesystemIsolator (filesystem/shared)

Hi Jie, I tried the new isolator and ran into an issue. Details are available in this ticket: https://issues.apache.org/jira/browse/MESOS-5187? Best Regards, Stephan From: Jie Yu Sent: Monday, April 11, 2016 17:37 To:

RE: Slaves not getting registered

Hi, There is no error while starting the slave machine. I have two masters and one slave configured and the slave isn’t getting registered in the GUI. Previously when I worked on single master and single slave, the GUI showed the activated status of slave. Thank you. From: Abhishek Amralkar

Re: marathon issue in running a docker container.

Do you have anything like haproxy for port mappings installed on your Mesos cluster? When using BRIDGE network, your process inside of the container, say SSH, is running on a standard port 22. Marathon allocates a random port in the default range of 31000 to 32000. However, it is your task to

Re: marathon issue in running a docker container.

>From the stderr log, nothing can be figure out. From stout log, it says server >running at host:port. But If I do telnet or curl, it does not show me any >reponse.output of stderr log : I0412 08:16:12.842341  9909 exec.cpp:134] Version: 0.27.1I0412 08:16:12.844701   9934 exec.cpp:208] Executor

Re: marathon issue in running a docker container.

When marathon launches the task while network : "BRIDGE", I can see from the mesos log that its listening to some host:port Even I can do an ssh to that instance, also saw MONGO_URL as env value over there. Pasting the log below -- 

Re: marathon issue in running a docker container.

anything is sandbox logs, why the tasks are getting killed? `stderr` and `stdout`? On 12-Apr-2016, at 1:35 PM, haosdent wrote: >Its frequently changing the deployment status to Staged Do you find any related log in mesos when marathon lauch the task? On Tue, Apr 12,

Re: Slaves not getting registered

What error you are getting when you are starting the mesos slaves? On 12-Apr-2016, at 1:32 PM, aishwarya.adyanth...@accenture.com wrote: Hi,   I’m unable to get the slave registered with the master node. I’ve configured both the masters and slave machines but am unable

Re: marathon issue in running a docker container.

>Its frequently changing the deployment status to Staged Do you find any related log in mesos when marathon lauch the task? On Tue, Apr 12, 2016 at 3:55 PM, nirmalendu swain < nirmalendu_sw...@yahoo.com> wrote: > Changing the network type to HOST does not work. Its frequently changing > the

Slaves not getting registered

Hi, I’m unable to get the slave registered with the master node. I’ve configured both the masters and slave machines but am unable to get it registered. Thank you. This message is for the designated recipient only and may contain privileged, proprietary, or

Re: marathon issue in running a docker container.

Changing the network type to HOST does not work. Its frequently changing the deployment status to Staged and then to no task. On Tuesday, 12 April 2016 11:49 AM, haosdent wrote: How about change the network type from BRIDGE to HOST? On Tue, Apr 12, 2016 at 2:13 PM,

Re: Jenkins Framework executor failing with peer certificate error

> Error downloading resource: Peer certificate cannot be authenticated with given CA certificates Failed to fetch: https://xxx.xxx.com/-fpci-fpci/jnlpJars/slave.jar Fetcher call libcurl to download the resources here. Your error looks like the cert in that website is invalid.

Jenkins Framework executor failing with peer certificate error

Hi Mesos'ers,     We have  a huge cluster of Jenkins instances that are configured with the mesos plugin. The mesos plugin does the on-demand framework registration with a mesos cluster when a Jenkins build is scheduled and requests for resources for executing the build. We are noticing that

marathon issue in running a docker container.

Hi Mesos user,I am running mesos marathon using dcos for spinning up AWS instances. I have successfully built the mongodb as a docker container, but when I try to deploy my dockerized app, it does n't deploy. My App is dependent upon mongo which is passed as environment variable in the json