[UIMA-5098] DUCC Agent does not handle /bin/cgcreate failure properly on re-start when there are left over processes

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.2.0-Ducc
• Component/s: DUCC
• Assignee: Jaroslaw Cwiklik
• Reporter: Lou DeGenaro
• Created: 01/Sep/16 14:29
• Updated: 21/Dec/16 15:58
• Resolved: 01/Sep/16 14:57

Description

Upon (re-)start of DUCC, one machine showed status as "defined" on the Machines page and "unknown" on Daemons page in ducc-mon.

Examination of the Agent's log revealed:

host.domain.net.agent.log:java.io.FileNotFoundException: /cgroup/ducc/327204.327204.28574/cgroup.procs (No such file or directory)

Also, manually attempting to run

/bin/cgcreate -t degenaro:mygroup -a degenaro:mygroup -g memory,cpu:ducc/test

failed with:

/bin/cgcreate: can't create cgroup ducc/test: Cgroup, operation not allowed

This may be due to Agent attempting to clean-up processes from previous lifetime, but is unable to do so.