-
Notifications
You must be signed in to change notification settings - Fork 59
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add sample stream and release metadata, and release index #207
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Few comments, otherwise LGTM
metadata/release-index.json
Outdated
@@ -0,0 +1,17 @@ | |||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
should we include stream name in release-index.json as well? Can be useful to figure out stream while looking at the json file.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The release-index is inside of a stream specific location (e.x.: s3://fcos-builds/prod/streams/<stream>/release-index.json
)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Followup: the release-index is also specific to an architecture. Thus, architecture label should be part of the location, or somewhere in the index.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@lucab: Reading the code again with more coffee leads me to believe that I misread something earlier and that is not the case currently.
Each entry in release-index
points to an endpoint for the release.json
which IS stream specific but is NOT architecture specific.
e.x.:
s3://fcos-builds/prod/streams/foo/release-index.json
is the release-index for the foo stream. An example entry would be something like:
{
"commit": "<hash>",
"version": "30.123.45",
"endpoint": "https://fcos-builds.s3.amazonaws.com/prod/streams/foo/builds/30.123.45/release.json"
}
metadata/release-index.json
Outdated
{ | ||
"releases": [ | ||
{ | ||
"commit": "<hash>", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
From comment #98 (comment), I believe commit and version we are referring here is ostree commit and version. One thing to note that, we will have same ostree version number for a release across all arches but commit hash will be different for different arches.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
None of these fields are unique, we can have different commits with the same version and it won't conflict.
to avoid confusion over naming.
Renamed release index sample to reduce confusion. Will address other comments next week. |
metadata/stream.yaml
Outdated
# slug is the best we can do. | ||
image: fedora_coreos_stable | ||
|
||
updates: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm unsure about the indentation of this. I don't think we want to have different update metadata for each architecture, to avoid exploding RelEng complexity. If so, I think this can become a top-level section on its own.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, in hindsight that choice seems somewhat arbitrary. If we want different update targets for different arches, presumably we'd also want them for different platforms. I don't want to introduce that level of complexity right now, so I'll switch this to a single value.
Might as well keep the document self-contained.
It seems more important to distinguish per-platform targets than per-arch ones, and that would introduce a lot of complexity. So, for now, keep a single update target for the entire stream.
Updated. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
From #98.