#324 closed defect (duplicate)

Cannot start NetVM

Reported by: joanna Owned by: joanna
Priority: major Milestone: Release 1 Beta 2
Component: core Keywords:
Cc:

Description

Sometimes qvm-start netvm hangs, and the following is in the logs:

[joanna@dom0 ~]$ cat /var/log/xen/xl-netvm.log
Waiting for domain netvm (domid 14) to die [pid 13139]
Domain 14 is dead
Action for shutdown reason code 3 is destroy
Domain 14 needs to be cleaned up: destroying the domain
libxl: error: libxl_device.c:513:libxl__wait_for_backend Backend /local/domain/0/backend/pci/14/0 not ready
libxl: error: libxl_device.c:513:libxl__wait_for_backend Backend /local/domain/0/backend/pci/14/0 not ready
Done. Exiting now

}}}}

Change History (3)

comment:1 Changed 22 months ago by rafal

In order to capture dmesg for the crashed domain kernel, please do:
1) in /usr/share/qubes/vm-template.conf, change "on crash=destroy" to "on_crash=preserve"
2) in qubes.py, comment out self.force_shutdown() after failed qrexec_daemon start (around line 974)

then reproduce the crash, connect to the crashed domain via xl console. Collect the logs; then cleanup via xl destroy.

comment:2 Changed 22 months ago by marmarek

How long have you waited and after which message qvm-start hanged?

If dom0 had network connected (via qvm-dom0-network-via-netvm) and netvm was halted, dom0 left with ghost eth0. This causes long DNS timeouts (ex sudo tries to determinate hostname).

comment:3 Changed 22 months ago by joanna

  • Resolution set to duplicate
  • Status changed from new to closed

I think this is a dup of #336.

Note: See TracTickets for help on using tickets.