mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chun-Hung Hsiao <>
Subject Re: Review Request 62336: [WIP] Kicked in disk monitoring early during recovery.
Date Thu, 14 Sep 2017 19:17:05 GMT

This is an automatically generated e-mail. To reply, visit:

(Updated Sept. 14, 2017, 7:17 p.m.)

Review request for mesos, Benjamin Mahler, Jie Yu, and Vinod Kone.

Summary (updated)

[WIP] Kicked in disk monitoring early during recovery.

Bugs: MESOS-7939

Repository: mesos


This patch calls `Slave::checkDiskUsage` after `Slave::recoverFramework`
so terminated containers are scheduled for GC, and before checkpointing
anything so there would be enough disk for checkpointing.

To avoid introducing a one-time disk monitoring function,
`Slave::checkDiskUsage` no longer contains a delayed recursion. Instead,
we use `process::loop` to monitor disk ugase periodically.


  src/slave/slave.cpp 6d1516a5d5b5db684f79385e60d892ff75fd00fd 



sudo make check


Chun-Hung Hsiao

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message