ops
ops copied to clipboard
image extensions should be standardized
right now we don't really enforce naming conventions even though all the images are standard disk images - it's prob. worth specifying them as .img or something and I believe currently ops image list -t onprem
and the equivalent run only finds them if they are comosed w/ '.img' extensions
right now on gcloud ->
➜ ~ ops image list
+-----------------------+--------+-------------------------------+
| NAME | STATUS | CREATED |
+-----------------------+--------+-------------------------------+
| bob2 | READY | 2020-04-07T10:40:30.004-07:00 |
+-----------------------+--------+-------------------------------+
| bobrandom-image | READY | 2020-02-26T09:33:57.598-08:00 |
+-----------------------+--------+-------------------------------+
and instances are brought up w/timestamps slapped on them to diff.
and aws has the same done for images
➜ ~ ops image list -t aws -z us-west-1
+------------+-------------------------------+-----------+--------------------------+
| NAME | ID | STATUS | CREATED |
+------------+-------------------------------+-----------+--------------------------+
| node-image | node-image1588609143075943000 | available | 2020-05-04T16:19:03.000Z |
so I think for local we should just keep the .img extension but list w/out the extension and same for cloud deploys
the other thing is that for both (incoming) esx && hyper-v support those will have diff. formats and might? store diff. formats locally
https://github.com/nanovms/ops/issues/1336 - seems like on-prem images are the odd ones out