Hi all!!

We're currently experimenting an error that's it's pretty weird, the reason
it's that with another user same base image and same flavor we can resize
the instance but with another user with the same base image we cannot with
this error on the logs:

2015-04-21 22:03:37.700 4047516 WARNING nova.scheduler.utils
[req-19a28c0f-ac5e-488b-8722-59135496f8e4 70886ff6b1434e0db5ddefa6180916f8
361f01feb5564973bcffe0a1d75ec795] Failed to compute_task_migrate_server: No
valid host was found.
Traceback (most recent call last):

File "/usr/lib/python2.7/dist-packages/oslo/messaging/rpc/server.py",
line 139, in inner
return func(*args, **kwargs)

File "/usr/lib/python2.7/dist-packages/nova/scheduler/manager.py", line
298, in select_destinations
filter_properties)

File
"/usr/lib/python2.7/dist-packages/nova/scheduler/filter_scheduler.py", line
140, in select_destinations
raise exception.NoValidHost(reason='')

NoValidHost: No valid host was found.

In case that someone ask this options are already on all the nodes and the
controller:

allow_resize_to_same_host=True
allow_migrate_to_same_host=True

Boths are in the default section, as I said before the resize can be done
with other instances and with and instances with the same specs.

On the other hand we tried with both cli and dashboard and both user with
enought quota to do the resize and with all the nodes with ram and cpu for
the instance, as a said we already did it with an instance with the same
flavor, can anyone point us on how to do a more deep debug or any advice
about how to handle the error to do a the resize?

Thanks in advance

jonathan.abdiel at gmail

Apr 21, 2015, 7:54 PM

Post #2 of 2
(464 views)

Permalink

Hi George,

We already did that we try the resize with an instance on the same node and
did the resize succesfully without troubles but with the instance with the
problem it didn't work, we're already working with debug=True and
verbose=True and nothings shows up.

We're now trying modifying the .xml file by hand but doesn't look like a
good solutions

On Tue, Apr 21, 2015 at 11:46 PM George Mihaiescu <lmihaiescu [at] gmail>
wrote:

> Hi Jonathan,
>
> The compute node where the instance that fails to resize is hosted must
> have a different config than the others.
> To confirm this theory, find another instance that runs on that host and
> initiate a resize as the user who owns that instance and it will probably
> fail as well.
>
> Enabling debug and verbose in nova.conf and restarting the services on
> that host before trying should give enough info in the nova-compute.log
> about the cause of the failure.
>
> George
>
> On Tue, Apr 21, 2015 at 10:23 PM, Jonathan Abdiel Gonzalez Valdebenito <
> jonathan.abdiel [at] gmail> wrote:
>
>> Hi all!!
>>
>> We're currently experimenting an error that's it's pretty weird, the
>> reason it's that with another user same base image and same flavor we can
>> resize the instance but with another user with the same base image we
>> cannot with this error on the logs:
>>
>> 2015-04-21 22:03:37.700 4047516 WARNING nova.scheduler.utils
>> [req-19a28c0f-ac5e-488b-8722-59135496f8e4 70886ff6b1434e0db5ddefa6180916f8
>> 361f01feb5564973bcffe0a1d75ec795] Failed to compute_task_migrate_server: No
>> valid host was found.
>> Traceback (most recent call last):
>>
>> File "/usr/lib/python2.7/dist-packages/oslo/messaging/rpc/server.py",
>> line 139, in inner
>> return func(*args, **kwargs)
>>
>> File "/usr/lib/python2.7/dist-packages/nova/scheduler/manager.py", line
>> 298, in select_destinations
>> filter_properties)
>>
>> File
>> "/usr/lib/python2.7/dist-packages/nova/scheduler/filter_scheduler.py", line
>> 140, in select_destinations
>> raise exception.NoValidHost(reason='')
>>
>> NoValidHost: No valid host was found.
>>
>>
>> In case that someone ask this options are already on all the nodes and
>> the controller:
>>
>> allow_resize_to_same_host=True
>> allow_migrate_to_same_host=True
>>
>> Boths are in the default section, as I said before the resize can be done
>> with other instances and with and instances with the same specs.
>>
>> On the other hand we tried with both cli and dashboard and both user with
>> enought quota to do the resize and with all the nodes with ram and cpu for
>> the instance, as a said we already did it with an instance with the same
>> flavor, can anyone point us on how to do a more deep debug or any advice
>> about how to handle the error to do a the resize?
>>
>>
>> Thanks in advance
>>
>>
>> _______________________________________________
>> Mailing list:
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>> Post to : openstack [at] lists
>> Unsubscribe :
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>
>>
>