Bug 979533 - VMs that were running prior to an (unclean?) system or libvirt shutdown sometimes do not appear in virt-manager or virsh list --all on next startup
Summary: VMs that were running prior to an (unclean?) system or libvirt shutdown somet...
Keywords:
Status: CLOSED DUPLICATE of bug 1015246
Alias: None
Product: Fedora
Classification: Fedora
Component: libvirt
Version: 20
Hardware: All
OS: All
unspecified
medium
Target Milestone: ---
Assignee: Libvirt Maintainers
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-06-28 18:23 UTC by Adam Williamson
Modified: 2013-10-03 18:40 UTC (History)
10 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2013-10-03 18:40:43 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)

Description Adam Williamson 2013-06-28 18:23:15 UTC
So I've seen this issue frequently through F19 testing, but it's hard to pin down *precisely* the conditions that cause it.

Roughly, though, I quite often notice that on system or virt-manager startup, at least one of my VMs is missing. It _seems_ like it's always VMs that were running in the last libvirt 'session'. I _think_ it may only happen when the system or at least libvirtd shutdown uncleanly (i.e., crashed).

For instance, just now I had a kernel hang while a VM was running, and on the next boot, that VM is missing from the list in virt-manager. All my other VMs are listed, just that one is missing. I've just checked and it's also missing from the list returned by 'virsh list --all', so this isn't a virt-manager issue.

If I restart libvirtd manually - 'systemctl restart libvirtd.service' - the VM shows up again.

Here's (hopefully) relevant info from checking on current libvirtd 'status', restarting libvirtd (so now it's in 'working' state) and comparing the same info:

[root@adam adamw]# systemctl status libvirtd.service
libvirtd.service - Virtualization daemon
   Loaded: loaded (/usr/lib/systemd/system/libvirtd.service; enabled)
   Active: active (running) since Fri 2013-06-28 11:06:24 PDT; 15min ago
 Main PID: 1245 (libvirtd)
   CGroup: name=systemd:/system/libvirtd.service
           ├─1245 /usr/sbin/libvirtd
           └─1495 /sbin/dnsmasq --conf-file=/var/lib/libvirt/dnsmasq/default....

Jun 28 11:06:25 adam.localdomain dnsmasq[1495]: using nameserver 64.59.144.90#53
Jun 28 11:06:25 adam.localdomain dnsmasq[1495]: using local addresses only f...s
Jun 28 11:06:25 adam.localdomain dnsmasq[1495]: read /etc/hosts - 12 addresses
Jun 28 11:06:25 adam.localdomain dnsmasq[1495]: read /var/lib/libvirt/dnsmas...s
Jun 28 11:06:25 adam.localdomain dnsmasq-dhcp[1495]: read /var/lib/libvirt/dn...
Jun 28 11:06:25 adam.localdomain dnsmasq-dhcp[1495]: not giving name localhos...
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: libvirt version: 1.0.5.2, p...)
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: failed to connect to monito...s
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: cannot lookup default selin...g
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: port 0 must be in range (59...)
[root@adam adamw]# journalctl -ab | grep libvirtd
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: libvirt version: 1.0.5.2, package: 1.fc19 (Fedora Project, 2013-06-12-22:00:57, buildvm-12.phx2.fedoraproject.org)
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: failed to connect to monitor socket: No such process
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: cannot lookup default selinux label for /media/Sea500/images/Test_1.img
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: port 0 must be in range (5900, 65535)
[root@adam adamw]# systemctl restart libvirtd.service
[root@adam adamw]# journalctl -ab | grep libvirtd
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: libvirt version: 1.0.5.2, package: 1.fc19 (Fedora Project, 2013-06-12-22:00:57, buildvm-12.phx2.fedoraproject.org)
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: failed to connect to monitor socket: No such process
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: cannot lookup default selinux label for /media/Sea500/images/Test_1.img
Jun 28 11:06:27 adam.localdomain libvirtd[1245]: port 0 must be in range (5900, 65535)
[root@adam adamw]# systemctl status libvirtd.service
libvirtd.service - Virtualization daemon
   Loaded: loaded (/usr/lib/systemd/system/libvirtd.service; enabled)
   Active: active (running) since Fri 2013-06-28 11:22:20 PDT; 9s ago
 Main PID: 4142 (libvirtd)
   CGroup: name=systemd:/system/libvirtd.service
           ├─1495 /sbin/dnsmasq --conf-file=/var/lib/libvirt/dnsmasq/default....
           └─4142 /usr/sbin/libvirtd

Jun 28 11:22:20 adam.localdomain systemd[1]: Starting Virtualization daemon...
Jun 28 11:22:20 adam.localdomain systemd[1]: Started Virtualization daemon.
Jun 28 11:22:23 adam.localdomain dnsmasq[1495]: read /etc/hosts - 12 addresses
Jun 28 11:22:23 adam.localdomain dnsmasq[1495]: read /var/lib/libvirt/dnsmas...s
Jun 28 11:22:23 adam.localdomain dnsmasq-dhcp[1495]: read /var/lib/libvirt/dn...
Jun 28 11:22:23 adam.localdomain dnsmasq-dhcp[1495]: not giving name localhos...

Comment 1 Fedora End Of Life 2013-09-16 14:18:14 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 20 development cycle.
Changing version to '20'.

More information and reason for this action is here:
https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora20

Comment 2 Cole Robinson 2013-10-03 18:40:43 UTC
Duping to my bug which has an easy reproducer

*** This bug has been marked as a duplicate of bug 1015246 ***


Note You need to log in before you can comment on or make changes to this bug.