Project

General

Profile

Bug #6405

Failure to reconnect on libvirtd restart

Added by   Zordrak almost 9 years ago. Updated over 4 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
Compute resources - libvirt
Target version:
-
Difficulty:
Triaged:
No
Bugzilla link:
Pull request:
Fixed in Releases:
Found in Releases:
1.5.1
Red Hat JIRA:

Description

In Compute Resources with a valid, working libvirtd resource; if the libvirtd daemon on the remote host is restarted, all foreman operations with the resource fail:

There was an error listing VMs: Call to virConnectNumOfDefinedDomains failed: internal error client socket is closed

This remains until the foreman rails app is restarted.

Here is an equivalent bug in the OpenStack Compute (Nova) project for reference purposes:
https://bugs.launchpad.net/nova/+bug/1154473

Related issues

Related to Foreman - Bug #14854: Libvirt connection leaksClosed

History

#1 Updated by   Zordrak almost 9 years ago

All machines in use for testing: CentOS 6.5 x86_64

#2 Updated by Ohad Levy almost 9 years ago

we currently rescue ::Libvirt::RetrieveError upon connection failures and try to reconnect, maybe libvirt changed their exception code.

https://github.com/theforeman/foreman/blob/develop/app/models/compute_resources/foreman/model/libvirt.rb#L148

#3 Updated by Dominic Cleal over 8 years ago

  • Category changed from Compute resources to Compute resources - libvirt

#4 Updated by Andy Taylor over 8 years ago

This might not be the same issue but we are experiencing this on Foreman 1.6.1, but not after a libvirt restart, it just seems to occasionally fail. Only a full restart of Foreman gets it working again.

#5 Updated by Lukas Zapletal over 4 years ago

  • Related to Bug #14854: Libvirt connection leaks added

Also available in: Atom PDF