Skip to main content
Question

Autodiscovery job stuck in Running state

  • April 20, 2026
  • 2 replies
  • 16 views
N
Forum|alt.badge.img+1

Hello,

I’m facing an issue with discovery jobs stuck in Running state on my Centreon platform.

Context

  • Multiple discovery jobs configured per poller

  • Providers used:

    • VMware (ESXi & VM discovery, ~20 jobs with filters)

    • AWX (2 jobs: pre-prod and prod)

  • Pollers are sized according to Centreon recommendations

  • Jobs are scheduled sequentially with intervals, running once per day

Issue

Recently, some discovery jobs are getting stuck in Running state indefinitely.

  • On the central server, I get errors indicating that the job cannot be launched because it is already in Running state

  • It is impossible to restart the job from the UI (even after waiting a long time and after restarting gorgoned service )

Checks already performed

  • Reviewed gorgoned logs on central and pollers

  • Checked VMware daemon logs on pollers

  • No explicit errors found explaining the blockage

  • Verified that jobs are not overlapping by design (sequential scheduling)

Observations

  • The issue appeared recently (no major known change)

  • Affected jobs never exit Running state

  • New executions are blocked because of this status

Questions

  • Has anyone already faced discovery jobs stuck in Running state?

  • Are there known issues with VMware or AWX providers causing this behavior?

Any help or feedback would be greatly appreciated.

Best regards,

2 replies

lpinsivy
Centreonian
Forum|alt.badge.img+21
  • lpinsivy
  • Centreonian
  • April 21, 2026

Hi ​@nourraseydou, in the gorgoned part for auto discovery, there is a global_timeout variable set to 60 (1 hour).

Do you think that some job can take more time than 60 minutes?

Can you try to enable DEBUG mode from /etc/sysconfig/gorgoned and check gorgoned logs on poller.

N
Forum|alt.badge.img+1
  • nourraseydou
  • Author
  • Steward *
  • April 21, 2026

Hello ​@lpinsivy,

When you refer to the global_timeout parameter, are you talking about the configuration in the file /etc/centreon-gorgone/config.d/41-autodiscovery.yaml, or is it something configured via the web interface?

On our side, our jobs do not require a 60-minute timeout. The longest one runs for about 2-3 minutes

Note that our Ansible Tower jobs are configured on the infrastructure poller (Engine). The AAP PRD job is the one that fails most often: it returns around 1200 hosts, which is more than the other discovery jobs given the filters we have applied (  in this discovery job, we configured a 120-second timeout to see if it makes any difference; it was previously set to 60 seconds )

Would you like me to send you the gorgoned logs from a poller where the job is in a Running state with debug mode enabled?

 

 

Terms & ConditionsCookie settingsAccessibility statement