Re: [Openstack] "There are not enough hosts available."
On Thu, Jul 28, 2016 at 2:34 PM, Ken D'Ambrosio wrote: > Trying to orchestrate a bunch of VMs on our ~500 core cloud, and we're > getting the "not enough hosts" problem. Only issue is that the dashboard > seems to imply we've got the elbow room needed to instantiate. Furthermore, > it would be *really handy* if I knew what the bottleneck it thought was: > disk, RAM, CPU, whatever. Here's a sample of the error: > > --- begin --- > 2016-07-28 18:40:47.961 1080 WARNING nova.scheduler.utils > [req-977badf4-a4a2-4766-87fd-c4e3ceb2e293 a4c895a6fe294ae3a81e3c6bb9209f52 > 33c2cf3cb61c425abcb0e0d35fdf21b7 - - -] [instance: > ee175a40-11d9-4a43-b6e1-c4d7935acff1] Setting instance to ERROR state. > 2016-07-28 18:40:51.892 1023 WARNING nova.scheduler.utils > [req-e2a65200-1662-4cd7-aa9e-db5e225adfa1 a4c895a6fe294ae3a81e3c6bb9209f52 > 33c2cf3cb61c425abcb0e0d35fdf21b7 - - -] Failed to > compute_task_build_instances: No valid host was found. There are not enough > hosts available. > Traceback (most recent call last): > > File "/usr/lib/python2.7/dist-packages/oslo_messaging/rpc/server.py", line > 142, in inner > return func(*args, **kwargs) > > File "/usr/lib/python2.7/dist-packages/nova/scheduler/manager.py", line > 84, in select_destinations > filter_properties) > > File > "/usr/lib/python2.7/dist-packages/nova/scheduler/filter_scheduler.py", line > 90, in select_destinations > raise exception.NoValidHost(reason=reason) > > NoValidHost: No valid host was found. There are not enough hosts available. > --- begin --- > > Any suggestions on how to track down exactly what's going wrong? Try taking a look at nova-scheduler log on your controller (usually in /var/log/nova/nova-scheduler.log). I do not remember any time I have seen the "not enough hosts" when the reason was not obvious from the scheduler log. > > Thanks! > > -Ken > > ___ > Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > Post to : openstack@lists.openstack.org > Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack --- Mikhail Medvedev IBM ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] [Openstack-operators] [openstack] Unable to launch an instance
On Tue, Feb 28, 2017 at 11:49 AM, Amit Kumar wrote: > Hi All, > > I have installed Openstack Newton using Openstack-Ansible. While creating an > instance, it is failing with following error: > > MessageNo valid host was found. There are not enough hosts The "not enough hosts" could be due to any number of reasons. To know exactly why, check your /var/log/nova/nova-scheduler.log on controller. The error means that nova scheduler was not able to find any suitable hosts to boot the VM. It could be for example because you are using a flavor that does not fit, or because your compute node appears dead to controller. In any case, nova scheduler log should make it a bit clearer. ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] Deleting instance in error state does not free resources
On Wed, Mar 8, 2017 at 10:57 AM, Georgios Dimitrakakis wrote: > Hi all! > > I am setting up an Ocata installation from scratch and my problem is when > some instances come up in ERROR state (for various reasons). > > Since they are not usable when they are deleted they still continue to > consume resources (VCPUs, Number of Instances, RAM etc). This obviously > affects the availability to new VMs. > > The only way is to clean them up using the "nova-manage project > quota_usage_refresh" > > > What could be wrong? Han anyone else encountered this? I might have seen something similar while manually running tests on devstack deployment. So it was a bit newer than Ocata. My test was creating / deleting VMs over and over, and after awhile the default quota of 10 instances got filled up, even though no VMs did exist. I wrote it off as me doing something wrong, and just increased quota. I suspect it was manifestation of the same problem with quotas not being cleaned up. I had deleted quite a few VMs that where in error state. I did not investigate this any further. ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] Why am I getting 2 IP's assigned to every instance?
On Thu, Mar 30, 2017 at 4:04 PM, Sterdnot Shaken wrote: > Openstack version: Mitaka > Number of external bridges: 1 > > Whenever I spawn a new cirros image, I'm get 2 ip's out of the same ip block > assigned to each instance. > > Anyone know why? I saw this happen on Mitaka cloud when an instance was rescheduled to a different compute node. E.g. initial scheduling puts it on compute1, it fails to boot there for some reason, and is being rescheduled to a compute2. It already has a port assigned from first attempt, and gets second port attached on reschedule, leaking the first port ip. To confirm if this is also what's happening in your case you can check nova scheduler log on your controller. > > Thanks! > > ___ > Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > Post to : openstack@lists.openstack.org > Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > --- Mikhail Medvedev (mmedvede) IBM ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack