You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Jeffrey Schroeder (JIRA)" <ji...@apache.org> on 2015/05/04 17:05:06 UTC

[jira] [Closed] (MESOS-2686) executor_registration_timeout and container_disk_watch_interval env vars can't be quoted

     [ https://issues.apache.org/jira/browse/MESOS-2686?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jeffrey Schroeder closed MESOS-2686.
------------------------------------
    Resolution: Fixed

The actual issue was that docker wasn't running, but was set as a containerizer.

docker version failed, due to docker not running as a daemon. mesos should be running docker --version instead.

> executor_registration_timeout and container_disk_watch_interval env vars can't be quoted
> ----------------------------------------------------------------------------------------
>
>                 Key: MESOS-2686
>                 URL: https://issues.apache.org/jira/browse/MESOS-2686
>             Project: Mesos
>          Issue Type: Bug
>          Components: slave
>    Affects Versions: 0.22.0
>         Environment: MESOS_master="file:///etc/mesos/zk"
> MESOS_log_dir="/var/log/mesos"
> MESOS_work_dir="/data/mesos/slave"
> MESOS_containerizers="docker,mesos"
> MESOS_isolation="posix/disk,cgroups/cpu,cgroups/mem"
> MESOS_cgroups_root="system.slice/mesos-slave.service"
> MESOS_cgroups_hierarchy="/sys/fs/cgroup"
> MESOS_container_disk_watch_interval="1mins"
> MESOS_executor_registration_timeout="5mins"
> ============ system unit ==================
> # cat /usr/lib/systemd/system/mesos-slave.service
> [Unit]
> Description=Mesos Cluster Manager
> After=network.target
> Wants=network.target
> [Service]
> EnvironmentFile=/etc/sysconfig/mesos-slave
> ExecStart=/usr/sbin/mesos-slave $MESOS_SLAVE_OPTS
> ExecStop=/usr/bin/killall -s 15 mesos-slave
> ExecReload=/bin/kill -HUP $MAINPID
> #######################################
> # NOTE: running as user mesos fails priv switching
> #######################################
> User=root
> Group=mesos
> Restart=always
> RestartSec=20
> LimitNOFILE=16384
> #######################################
> # NOTE: Below are cgroup options
> #######################################
> Slice=mesos
> CPUAccounting=true
> #######################################
> # NOTE: Adjust as needed
> #######################################
> CPUShares=1024
> ##################
> MemoryAccounting=true
> #TBD: MemoryLimit=bytes, MemorySoftLimit=bytes
> ##################
> #BlockIOAccounting=true
> #BlockIOWeight=??
> #BlockIODeviceWeight=??
> #TBD: BlockIOReadBandwidth=bytes, BlockIOWriteBandwidth=bytes
> #DeviceAllow=
> #DevicePolicy=auto|closed|strict
> [Install]
> WantedBy=multi-user.target
> ========================================
> =============== /etc/sysconfig/mesos-slave ===
> # This file contains environment variables that are passed to mesos-slave.
> # To get a description of all options run mesos-slave --help; any option
> # supported as a command-line option is also supported as an environment
> # variable.
> #
> # Also see the "Slave Options" section at:
> #    http://mesos.apache.org/documentation/latest/configuration
> # The mesos master URL to contact. Should be host:port for
> # non-ZooKeeper based masters, otherwise a zk:// or file:// URL.
> MESOS_master="file:///etc/mesos/zk"
> # For a complete listing of options execute 'mesos-slave --help'
> MESOS_log_dir="/var/log/mesos"
> MESOS_work_dir="/data/mesos/slave"
> MESOS_containerizers="docker,mesos"
> # systemd cgroup integration
> MESOS_isolation="posix/disk,cgroups/cpu,cgroups/mem"
> MESOS_cgroups_root="system.slice/mesos-slave.service"
> MESOS_cgroups_hierarchy="/sys/fs/cgroup"
> # Slave specific options
> MESOS_container_disk_watch_interval="1mins"
> MESOS_executor_registration_timeout="5mins"
> ========================================
>            Reporter: Jeffrey Schroeder
>            Priority: Minor
>
> ]# mesos-slave --help
> Failed to load flag 'executor_registration_timeout': Failed to load value '"5mins"': Failed to convert '' to number
> # mesos-slave --help
> Failed to load flag 'container_disk_watch_interval': Failed to load value '"1mins"': Failed to convert '' to number
> Usage: mesos-slave [...]
> The slave should recognize that this is quoted and not try to literally read it in.
> I'm using a backported copy of mesos-0.22.0-4.e890e24 from fedora on RHEL7.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)