Resource or resource group is moved to a new node instead of being recovered when two related resources fail together, but the dependent resource failure is detected first, in a RHEL 6 or 7 High Availability cluster
Issue
- I have a
Filesystemresource and another resource that depends on it. If the file system fails, then both resources are going to fail their nextmonitoroperation, but if the dependent resource failure is detected first, thenpacemakertries to recover it and when it can't (because theFilesystemhasn't been recovered yet), it moves the whole resource group to a new node. - If two resources are related to each other and one must be recovered before the other, there's no guarantee with
pacemakerthat the monitor operation will be detected before the other. So, you might have one resource failing, and then it can't be restarted because its "parent" is still in a bad state.
Environment
- Red Hat Enterprise Linux (RHEL) 6 or 7 with the High Availability Add On
pacemaker- One or more resources that depends on another resource, such that if the dependency ("parent") fails, the dependent ("child") will fail as well, and they'd both need to be recovered
- This is often manifested through either:
- Two resources ordered in a specific way within a resource group, such as a
Filesystemand another for the software that uses the file system. Or - Two resources associated with each other through an
ordering constraint
- Two resources ordered in a specific way within a resource group, such as a
- This is often manifested through either:
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase of over 48,000 articles and solutions.
Close
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.