[Haizea] Another bug

[Mehdi Sheikhalishahi]

No it's not working. I've done some debugging but could not figure
out. Here is python traceback.

Traceback (most recent call last):
  File "/home/mehdi/workspace/scheduler/tests/start.py", line 36, in <module>
    MANAGER.start()
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 375, in start
    self.__unexpected_exception(exc)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 642, in __unexpected_exception
    self.__panic()
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 371, in start
    self.clock.run()
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 792, in run
    self.manager.process_requests(self.time)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 442, in process_requests
    self.__unexpected_exception(exc)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 642, in __unexpected_exception
    self.__panic()
  File "/home/mehdi/workspace/scheduler/src/haizea/core/manager.py",
line 438, in process_requests
    self.scheduler.schedule(nexttime)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/lease_scheduler.py",
line 180, in schedule
    self.__schedule_lease(lease, nexttime=nexttime)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/lease_scheduler.py",
line 665, in __schedule_lease
    (vmrr, preemptions) = self.vm_scheduler.schedule(lease,
lease.duration.get_remaining_duration(), nexttime, earliest)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/vm_scheduler.py",
line 111, in schedule
    return self.__schedule_exact(lease, duration, nexttime, earliest)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/vm_scheduler.py",
line 374, in __schedule_exact
    allow_preemption = True)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/mapper.py",
line 194, in map
    avail=aw.get_ongoing_availability(start, pnode, preempted_leases =
preempting)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/slottable.py",
line 1217, in get_ongoing_availability
    return OngoingAvailability(self.changepoints[time].nodes[node],
preempted_leases)
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/slottable.py",
line 1314, in __init__
    if prev_avail != None and available.any_less(prev_avail.available):
  File "/home/mehdi/workspace/scheduler/src/haizea/core/scheduler/slottable.py",
line 260, in any_less
    for i, x in l:
TypeError: 'int' object is not iterable


On Wed, May 19, 2010 at 12:11 AM, Borja Sotomayor <borja at borjanet.com> wrote:
> Hi,
>
>> This is another bug when I want to load site information which
>> contains a multi-instance resource for CPU.
>
> Thanks for pointing this out. The error turned out to be in the
> 'repository node' that gets added by the deployment scheduler, and
> which doesn't have any CPU resources (with single-instance CPUs, this
> doesn't cause a problem since, internally, it's just represented as a
> zero). However, this turned out to cause problems when the site has
> multi-instance CPU resources (where the internal representation is a
> bit more complex). I've committed a fix that should resolve this:
>
> https://phoenixforge.cs.uchicago.edu/repositories/revision/haizea/795
>
> Let me know if it works on your end.
>
> Cheers!
> --
> Borja Sotomayor
> PhD Candidate in Computer Science, University of Chicago
> http://people.cs.uchicago.edu/~borja/
> _______________________________________________
> Haizea mailing list
> Haizea at mailman.cs.uchicago.edu
> https://mailman.cs.uchicago.edu/mailman/listinfo/haizea
>



-- 
Regards,
Mehdi