stack.sh interrupted because error when configuring subnet pool --default in neutron

When executing stack.sh the installation can't finish because it breaks in the point the script is configuring the subnet pool --default .
seems that this command :
openstack --os-cloud devstack-admin --os-region RegionOne subnet pool create shared-default-subnetpool-v4 --default-prefix-length 26 --pool-prefix 10.0.0.0/22 --share --default -f value -c id

don't work, but it would work withoud "--default" option. I can say that because after failing and interrupting the execution i can sent the command without --default :
openstack --os-cloud devstack-admin --os-region RegionOne subnet pool create shared-default-subnetpool-v4 --default-prefix-length 26 --pool-prefix 10.0.0.0/22 --share -f value -c id

and it is successfull . Then when i try to set default the previous created subnetpool, it rises the same error (unknown error) .

In case that i repeat the installation before unstacking and cleaning , then the problem seems that doesn't occur. it's very strange , but is what i have checked. In that case, if the installation continues, then it breaks again when trying to create a network for a tenant:
 openstack --os-cloud devstack-admin --os-region RegionOne network create --project d65d9cba237c4efd90eba58c7664b2b2 private .
The error tells that no tenant available ... Seems that is due to vni_ranges that are incorrect. If i modify ml2.conf file and set the vni_ranges = 1:1000 , then i can successfully execute the command. But anyway i need to configure that from the local.conf and i have seen that adding following parameters (Q_ML2_PLUGIN_VXLAN_TYPE_OPTIONS=(vni_ranges=1:1000) ) in that file, has no effect (the ml2.conf resulting has the same values(1001:2000) .

How i could bypass that steps and do it manually later on? or they are necessary for the next steps?