Skip to:

  1. Skip to Jira Navigation
  2. Skip to Side Navigation
  3. Skip to Main Content

Thread leak in Snmp4JStrategy

Description

Customer reports a persistent apparent leak of about 50-100 threads per day. The thread count seems to jump each day after midnight. These jumps are consistent with Provisiond's daily rescans triggering the leak.

Attaching resource graph of JMX data showing long-term thread count trend, plus thread dumps from consecutive days from a running 20.0.2 install.

Much more background and history dating back to the original 18.0.4 upgrade is available from the support ticket history (see Environment).

Environment

https://mynms.opennms.com/Ticket/Display.html?id=5104 Synopsis: Started seeing a leak after upgrading 18.0.2 -> 18.0.4. Upgrading to 19.1.0 slowed the leak considerably (consistent with fix of NMS-9233 and NMS-8825) but did not stop it. Upgrading to 20.0.1 and later 20.0.2 seems to make no difference in this regard versus 19.1.0.

Acceptance / Success Criteria

None

Attachments

6

Lucidchart Diagrams

Activity

Show:

Alejandro Galue July 24, 2018 at 8:25 PM

Could this be affecting Meridian 2016 ? I'm currently helping a customer with performance issues, and as SNMP4j appears on the analysis thrown by fastthread.io after analyzing a few thread dumps, I'm wondering if if this could be related or not.

Jeff Gehlbach November 22, 2017 at 3:30 PM

Thanks

Jesse White November 22, 2017 at 2:45 PM

It was merged into foundation-2017, so it's present in M2017.1.2. I updated the fix versions accordingly.

Jeff Gehlbach November 22, 2017 at 1:59 PM

Should this fix make it into Meridian 2017?

Jesse White October 30, 2017 at 8:16 PM

Fixed

Details

Assignee

Reporter

Labels

Components

Sprint

Fix versions

Affects versions

Priority

PagerDuty

Created September 8, 2017 at 3:38 PM
Updated July 24, 2018 at 8:25 PM
Resolved November 8, 2017 at 3:12 PM
Configure