[Haizea] Haizea suddenly stops with Message: 'sanity-check'

Zaina AFOULKI zaina.afoulki at ensi-bourges.fr
Fri Dec 10 06:42:40 CST 2010


Hi,

I noticed an unexpected behavior while running Haizea branch 1.1:

Whenever I run an AR lease (for a limited time, say 5min), and when It
needs to be stopped, Haizea suddendly stops. This is what the
haizea log says:


[2010-12-09 17:53:09.00] CLOCK   Waking up to manage resources
[2010-12-09 17:53:09.00] CLOCK   Wake-up time recorded as 2010-12-09
17:53:09.00
[2010-12-09 17:53:09.01] VMSCHED Stopped VMs for lease 1 on nodes [1]
[2010-12-09 17:53:09.24] CLOCK   Going back to sleep. Waking up at
2010-12-09 17:53:19.00 to see if something interesting has happened by
then.
[2010-12-09 17:53:19.00] CLOCK   Waking up to manage resources
[2010-12-09 17:53:19.00] CLOCK   Wake-up time recorded as 2010-12-09
17:53:19.00
[2010-12-09 17:53:19.02] CLOCK   Going back to sleep. Waking up at
2010-12-09 17:53:29.00 to handle slot table event.
[2010-12-09 17:53:29.00] CLOCK   Waking up to manage resources
[2010-12-09 17:53:29.00] CLOCK   Wake-up time recorded as 2010-12-09
17:53:29.00
[2010-12-09 17:53:29.00] VMSCHED Lease 1's VMs have shutdown.
[2010-12-09 17:53:29.00] RM      An unexpected exception has happened.
[2010-12-09 17:53:29.04] RM        File
"/usr/local/lib/python2.6/dist-packages/haizea/core/manager.py", line 462,
in process_ending_reservations
    self.scheduler.process_ending_reservations(time)

[2010-12-09 17:53:29.04] RM        File
"/usr/local/lib/python2.6/dist-packages/haizea/core/scheduler/lease_scheduler.py",
line 287, in process_ending_reservations
    self._handle_end_lease(lease)

[2010-12-09 17:53:29.04] RM        File
"/usr/local/lib/python2.6/dist-packages/haizea/core/scheduler/lease_scheduler.py",
line 864, in _handle_end_lease
    if get_config().get("sanity-check"):

[2010-12-09 17:53:29.04] RM        File
"/usr/local/lib/python2.6/dist-packages/haizea/common/config.py", line
142,
in get
    return self._options[opt]

[2010-12-09 17:53:29.04] RM      Message: 'sanity-check'
[2010-12-09 17:53:29.04] RM      Panicking...
[2010-12-09 17:53:29.69] RM      --- Haizea status summary ---
[2010-12-09 17:53:29.69] RM      Done:                    1 (1 AR)
[2010-12-09 17:53:29.69] RM                          ------
[2010-12-09 17:53:29.69] RM                    TOTAL:     1
[2010-12-09 17:53:29.69] RM      ---- End summary ----
[2010-12-09 17:53:29.69] RM      Next change point (in slot table): None

onevm list shows that the VM is correctly shutdown by OpenNebula.

I'm not sure if this is related but I put a shutdown-time of 20 in the
haizea config file.
Anybody could explain the reason for this?

Thank you

Zaina



More information about the Haizea mailing list