Just finished installing Newton on Centos 7.2 using the current RDO packstack.

I made sure there was plenty of space by setting up a cinder-volumes VG and all seemed to be ok until i tried to launch a third instance and I got the following error:

"Error: Failed to perform requested operation on instance "dii", the instance has an error status: Please try again later [Error: No valid host was found. There are not enough hosts available.]".

The messages I'm getting in the log are:

2016-10-23 23:32:50.290 1507 WARNING nova.scheduler.host_manager [req-7fff9103-75ac-4578-8c11-0957545b7295 f68c4ddce4b24061ababda71362b7324 4752b726f6e84141b386e8f0adacae13 - - -] Host anesidora has more disk space than database expected (130 GB > 19 GB)
2016-10-23 23:32:50.291 1507 INFO nova.filters [req-7fff9103-75ac-4578-8c11-0957545b7295 f68c4ddce4b24061ababda71362b7324 4752b726f6e84141b386e8f0adacae13 - - -] Filter DiskFilter returned 0 hosts
2016-10-23 23:32:50.292 1507 INFO nova.filters [req-7fff9103-75ac-4578-8c11-0957545b7295 f68c4ddce4b24061ababda71362b7324 4752b726f6e84141b386e8f0adacae13 - - -] Filtering removed all hosts for the request with instance ID '9d0e6a3e-1401-4f14-bdd1-219dd60315be'. Filter results: ['RetryFilter: (start: 1, end: 1)', 'AvailabilityZoneFilter: (start: 1, end: 1)', 'RamFilter: (start: 1, end: 1)', 'DiskFilter: (start: 1, end: 0)']

Of course there is plenty of disk pace on my cinder-volumes VG where the volumes are actually being created:

[root@anesidora nova]# vgs
  VG             #PV #LV #SN Attr   VSize   VFree
  centos           1   3   0 wz--n- 246.00g  15.86g
  cinder-volumes   1   2   0 wz--n- 684.82g 504.82g

And the volumes are being created on the cinder VG

  LV                                          VG             Attr       LSize   Pool   Origin Data%  Meta%  Move Log Cpy%Sync Convert
  pool00                                      centos         twi-aotz-- 140.00g               10.73  5.67
  root                                        centos         Vwi-aotz-- 140.00g pool00        10.73
  swap                                        centos         -wi-ao----  90.00g
  volume-75aa56f5-03e9-4d04-b7b7-147e44bb5abe cinder-volumes -wi-ao----  80.00g
  volume-eab6405f-3824-456d-bad4-aabfcca87d0d cinder-volumes -wi-ao---- 100.00g

From what Im seeing, the hypervisor local_gb has been set to the size of the pool00 and not to the size of the cinder-volumes VG. And I gather that as I add instances the local_used is increased based on the instance volume size and so it runs out fast as the local_gb is wrong for this install.

Can anyone help with how I can reset the local_gb to the actual size of the cinder-volumes VG?

Thanks in advance,

Randall.