[ https://issues.apache.org/jira/browse/MESOS-2845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14581024#comment-14581024 ]
Ian Downes commented on MESOS-2845: ----------------------------------- +1 to removing the hack. In the interim, can we simple match the executor's resource revocable/non-revocable state to that of the task's resources? IIRC the command executor can only run a single task anyway. > Command tasks lead to a mixing of revocable / non-revocable cpus and memory > within the container. > ------------------------------------------------------------------------------------------------- > > Key: MESOS-2845 > URL: https://issues.apache.org/jira/browse/MESOS-2845 > Project: Mesos > Issue Type: Bug > Components: slave > Reporter: Benjamin Mahler > Labels: twitter > > Due to the hack > [here|https://github.com/apache/mesos/blob/9a5788801e7fc95fce99749a23803fc52c67c0ce/src/slave/slave.cpp#L3101], > where we add a small set of resources into the command executor: > {code} > ExecutorInfo Slave::getExecutorInfo( > const FrameworkID& frameworkId, > const TaskInfo& task) > { > if (task.has_command()) { > ... > // XXX: These are always non-revocable. > // Add an allowance for the command executor. This does lead to a > // small overcommit of resources. > executor.mutable_resources()->MergeFrom( > Resources::parse( > "cpus:" + stringify(DEFAULT_EXECUTOR_CPUS) + ";" + > "mem:" + stringify(DEFAULT_EXECUTOR_MEM.megabytes())).get()); > } > ... > } > {code} > The obvious extension here would be to make these revocable, but would be > great to remove this hack entirely. > Seems to originate in [r/22251|https://reviews.apache.org/r/22251/] from > MESOS-1417. > FYI [~idownes] [~jieyu] -- This message was sent by Atlassian JIRA (v6.3.4#6332)