Re: Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-02 Thread Meng Zhu
CORRECTION: This is a new behavior that only appears in the current 1.5.x branch. In 1.5.0, Mesos agent still has the old behavior, namely, any reordered tasks (to the same executor) are launched regardless. On Fri, Mar 2, 2018 at 9:41 AM, Chun-Hung Hsiao wrote: >

Re: Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-02 Thread Chun-Hung Hsiao
Gilbert I think you're right. The code path doesn't exist in 1.5.0. On Mar 2, 2018 9:36 AM, "Chun-Hung Hsiao" wrote: > This is a new behavior we have after solving MESOS-1720, and thus a new > problem only in 1.5.x. Prior to 1.5, reordered tasks (to the same executor) >

Re: Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-02 Thread Chun-Hung Hsiao
This is a new behavior we have after solving MESOS-1720, and thus a new problem only in 1.5.x. Prior to 1.5, reordered tasks (to the same executor) will be launched because whoever comes first will launch the executor. Since 1.5, one might be dropped. On Mar 1, 2018 4:36 PM, "Gilbert Song"

Re: Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-01 Thread Gilbert Song
Meng, Could you double check if this is really an issue in Mesos 1.5.0 release? MESOS-1720 was resolved after the 1.5 release (rc-2) and it seems like it is only at the master branch and 1.5.x branch (not 1.5.0). Did I miss anything? - Gilbert

Re: Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-01 Thread Benjamin Mahler
Put another way, we currently don't guarantee in-order task delivery to the executor. Due to the changes for MESOS-1720, one special case of task re-ordering now leads to the re-ordered task being dropped (rather than delivered out-of-order as before). Technically, this is strictly better.

Tasks may be explicitly dropped by agent in Mesos 1.5

2018-03-01 Thread Meng Zhu
Hi all: TLDR: In Mesos 1.5, tasks may be explicitly dropped by the agent if all three conditions are met: (1) Several `LAUNCH_TASK` or `LAUNCH_GROUP` calls use the same executor. (2) The executor currently does not exist on the agent. (3) Due to some race conditions, these tasks are trying to