Open roberth opened 2 years ago
Together with the new NixOps way of having explicit network directories this creates confusion and possibility of devastating errors. Consider two deployments, production
and development
, which use different network directories. On NixOps 1 I would have set different configuration paths and use nixops -d production ...
. Currently when I'm in development
network directory I see this:
+--------------------------------------+-------------+----------------------------------------------+------------+------+
| UUID | Name | Description | # Machines | Type |
+--------------------------------------+-------------+----------------------------------------------+------------+------+
| e0a78705-3cff-11ea-85a3-8600003a0125 | development | Development deployment | 1 | none |
| 9c212957-69f2-11ea-88d6-8600003a0125 | production | Development deployment | 1 | none |
+--------------------------------------+-------------+----------------------------------------------+------------+------+
I could by mistake run nixops deploy -d production
here and destroy my servers.
My flake-based deployment contains
which points to a new path. Yet, I see all my usual legacy deployments in
nixops list
.I can't find
databasefile
anywhere but in the NixOps 1->2 upgrade message.