Bug 604852
Summary: | when on_crash is set to "coredump-restart", virsh and libvirt API complain unknown lifecycle type and some functions no longer work | ||
---|---|---|---|
Product: | [Community] Virtualization Tools | Reporter: | Tom <xiaohm> |
Component: | libvirt | Assignee: | Libvirt Maintainers <libvirt-maint> |
Status: | CLOSED CURRENTRELEASE | QA Contact: | |
Severity: | high | Docs Contact: | |
Priority: | low | ||
Version: | unspecified | CC: | crobinso, jtomko, xen-maint |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2015-07-28 14:33:24 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Tom
2010-06-16 20:39:22 UTC
There is a workaround for this problem. Set the on_crash="preserve" for a VM. After the VM crashed, a coredump will be generated automatically and the VM will be kept and its status will be "crashed". A deamon process can be used to detect this status and then restart it. (In reply to comment #0) > Description of problem: > After we edited Xen configuration file for a VM to dump core when the VM > crashes, we noticed that some virsh commands, such as "virsh dumpxml" or "virsh > console", start to display following errors and cannot work properly. > # virsh dumpxml linuxvm > error: internal error unknown lifecycle type coredump-restart > We observed this problem in libvirt-0.7.6-1. We believe other versions should > have similiar issue. > How reproducible: > Steps to Reproduce: > 1. create a PV VM named "linuxvm" > 2. edit /etc/xen/linuxvm by adding following line > on_crash = "coredump-restart" > 3. execute "xm create linuxvm" to recreate the VM. > 4. after linux started, run following command, > # virsh dumpxml linuxm > Actual results: > error: internal error unknown lifecycle type coredump-restart > Expected results: > 'dumpxml' should display xml on the screen as expected. > Additional info: > Besides "dumpxml", we also noticed "console" doesn't work properly. We didn't > test other command also we believe some other virsh command could be affected, > too. Also, seems like some underlying libvirt API is broken by this as well. > Is there any workaround solution for this problem? Fixed upstream by: commit b9c10268e14b5f6a8a405813406964ccade67fcb Author: Jim Fehlig <jfehlig> AuthorDate: 2010-08-12 11:15:44 -0600 Commit: Eric Blake <eblake> CommitDate: 2010-08-20 15:06:30 -0600 Add actions to virDomainLifecycle enum Xen supports on_crash actions coredump-{destroy,restart}. libvirt cannot parse config returned by xend that contains either of these actions xen52 # xm li -l test | grep on_crash (on_crash coredump-restart) xen52 # virsh dumpxml test error: internal error unknown lifecycle type coredump-restart This patch adds a new virDomainLifecycleCrash enum and appends the new options to existing destroy, restart, preserve, and rename-restart options. git describe: v0.8.3-65-gb9c1026 contains: v0.8.4~62 |