Closed 0xIslamTaha closed 5 years ago
After discussion with @zaibon, we decided to drop the naemspace selfheal where it reinstalls itself on a different zerodb. The selfhealing implemented by @muhamadazmy and @andrewayoub is already enough and if a namespace is not reachable anymore it uninstalls and delete it, then creates a new namespace service and install it
This https://github.com/threefoldtech/0-templates/pull/236/files reverts the namespace selfhealing. And makes sure the s3 keeps trying to delete the dead namespace services.
we found how exactly that happened
we were trying to get a service with
empty name
so it returns all zdbs. I suggest checking thelen
of thename
args to fix the error.In another hand we found that this namespace hasn't a zdb and that may be occured cause we turned off this namespace's zdb many times and at some point all the running zdb has a namespace with the same name of this namespace so there is no way to create it again so it wouldn't have a zdb forever.
This scenario should be healing on the s3_redundant