glb_services - drained nodes become re-enabled

[john275]

When using brocadevtm::glb_services and setting a location to draining in the UI subsequent puppet run removes/overwrites the location_draining property.

10.4r2
puppet 3.8.7

[john275]

Is there any further information I should supply to help understand the issue ?

[TuxInvader]

Hi John,

The glb_service manifest has a setting "basic__location_draining", which needs to contain the name of the location you want to drain. When you mark a location as draining in the vTM UI it sets this key. If you don't want puppet to revert the change, then you need to set the key in your manifest. Otherwise puppet will find the key has been modified on its next run and revert it.

When you have a puppet agent monitoring your vTM cluster, then you should probably avoid making changes in the UI. Alternatively you could disable puppet at times when you are making temporary config changes (such as location/server draining). In this instance, both you and puppet are trying to configure the vTM with conflicting settings.

cheers,
Mark

[john275]

Thanks Mark. Please close this as dumb user.

I may have another issue coming... currently My BSD license is setting a cluster ID but puppet run dislikes that when deep compare on global_settings