Conversation
- release every new tag on github - two docker images with single entrypoint offered - We will use two separate docker images as gaiacli can refer to a remote gaiad - gaiali can scale and eventually be attacked while gaiad will keep live Signed-off-by: Karoly Albert Szabo <szabo.karoly.a@gmail.com>
|
@sabau this is failing tests |
Codecov Report
@@ Coverage Diff @@
## develop #3694 +/- ##
===========================================
+ Coverage 61.15% 61.17% +0.01%
===========================================
Files 190 190
Lines 14008 14008
===========================================
+ Hits 8567 8569 +2
+ Misses 4905 4903 -2
Partials 536 536 |
|
Ok, rerunning fixed that |
|
|
||
| # Push the tag also for the version, so clients can build up on top of it | ||
| GAIAD_VERSION=`/tmp/workspace/bin/gaiad version` | ||
| docker push tendermint/gaia:$GAIAD_VERSION |
There was a problem hiding this comment.
I removed the hashed version, in favour of the git tag version that will run only when a new tag will be added
There was a problem hiding this comment.
Why not using git describe's output?
There was a problem hiding this comment.
It seemed to me more readable to use $CIRCLE_TAG instead of getting the tag from a command.
In here the specific reason to change the behaviour was also to trigger the new job only when the tag is pushed.
There was a problem hiding this comment.
It seemed to me more readable to use $CIRCLE_TAG instead of getting the tag from a command.
The risk is that versions may diverge. Did you check that gaiad version kinda looks like the image's actual version?
cwgoes
left a comment
There was a problem hiding this comment.
Thanks; a few notes:
- Needs
PENDING.md - Do we really need separate images for
gaiacli/gaiad? Why not just one? - Can we add instructions in documentation on how to use the Docker images?
|
Sure, documentation and pending will be added today, I'll remove the R4R until It's there. The goal was to use them as commands so clients can get just the piece they need and use them as a single command. It should be easier to handle logs and output when it runs on cloud solution (but feasible also with the current container definition). If it's safer we can try a two step delivery:
|
|
If you think two images is more convenient, I don't have a strong opinion. |
Signed-off-by: Karoly Albert Szabo <szabo.karoly.a@gmail.com>
Signed-off-by: Karoly Albert Szabo <szabo.karoly.a@gmail.com>
|
In this moment I just need tagged docker images to spin off nodes on AWS, when we will need to scale we can decide if splitting or not to handle logs etc. |
cwgoes
left a comment
There was a problem hiding this comment.
ACK although I haven't tested the images
|
I don't think we need 2 images, but this work is great! |
I worked together with @greg-szabo into this process:
Targeted PR against correct branch (see CONTRIBUTING.md)
Linked to github-issue with discussion and accepted design OR link to spec that describes this work.
Wrote tests
n/aUpdated relevant documentation (
docs/)Added entries in
PENDING.mdwith issue #rereviewed
Files changedin the github PR explorerFor Admin Use: