Applies to:

Zend Server

Linux Operating system

Problem

In some situations, such as multiple upgrades and changes to the Zend Server installation base, temporary resources might be "stuck" in the temp directory and services which comes from PHP runtime modules, such as monitoring, jobQueue and Statistics might lose communication and break their normal flow. As a result, tasks might stop running, or data in the Admin UI (such as lists and dashboards) might be partial or missing completely.

Solution

The most straight forward solution, which resolves most scenarios of broken services and pulling data, is to clean up temporary resources and initialize a new set by a fresh start of all Zend services on any affected Zend Server instance (or cluster nodes).

The following manual actions, running as root on the machine, will ensure in a fresh run of all Zend services.

  1. Stop Zend services:
    /usr/local/zend/bin/zendctl.sh stop
  2. Verify all services are down:
    ps aux | grep zend
  3. Kill any leftover process so nothing holds temp files
  4. Cleanup temp files (sockets, semaphores):
    cd /usr/local/zend/tmp
    rm -vf *.sock zsemfile_* zshm_* *sem* *.pid
  5. Start Zend Services:
    /usr/local/zend/bin/zendctl.sh start
  6. Verify all are running
    /usr/local/zend/bin/zendctl.sh status

Once done, you should be checking the original problem which led to this article, and see if it has been resolved.

Generating a Support Tool Archive