EtcdConfigurationSource

[spoon16]

An Etcd watched configuration source very similar to the ZooKeeperConfigurationSource.

Uses the Boon Etcd client https://github.com/boonproject/boon/blob/master/etcd/README.md. Seems to be the most widely adopted Etcd client.

Tests included.

[cloudbees-pull-request-builder]

NetflixOSS » archaius » archaius-pull-requests #85 FAILURE
Looks like there's a problem with this pull request

[spoon16]

Build failed because of a JVM error... or maybe a bunch of failing unit tests unrelated to the Etcd module.

Unless I am missing something.

[cloudbees-pull-request-builder]

archaius-pull-requests #254 SUCCESS
This pull request looks good

[howardyuan]

There're two pull-requests builds. One of them was not configured with
proper Gradle opts so it runs out Permgen. I'll see how to either disable
the second one or add proper GRADLE_OPTS.

On Thu, Jun 11, 2015 at 6:02 PM, Eric Schoonover notifications@github.com
wrote:

Build failed because of a JVM error... or maybe a bunch of failing unit
tests unrelated to the Etcd module.

Unless I am missing something.

—
Reply to this email directly or view it on GitHub
#318 (comment).

[howardyuan]

Can the action() return null when it's error?

[spoon16]

only if I duplicate the etcd.waitRecursive call on l:89

It'd probably be best to capture the wasError condition, throw an exception and have a finally that calls etcd.waitRecursive. But I'm not sure that really makes things more readable. I'll just duplicate l:89 and return early.

[cloudbees-pull-request-builder]

NetflixOSS » archaius » archaius-pull-requests #86 FAILURE
Looks like there's a problem with this pull request

[cloudbees-pull-request-builder]

archaius-pull-requests #255 SUCCESS
This pull request looks good

[spoon16]

@howardyuan, tried to address your feedback with the following changes.

• fixed bug where delete wouldn't remove the value from the EtcdConfigurationSource
• added a unit test to validate delete behavior
• refactored the updateHandler in EtcdConfigurationSource... tried to improve readability by not always instantiating three maps

[howardyuan]

Looks good to me. Will leave the PR open for a few days to see if anybody else have any more comments.

[spoon16]

@howardyuan merge?

[chandrapersonal]

why there is no case for node action "compareandswap" in etcd UpdateHandler. we have node actions create,set and delete action on UpdateHandler . but no compareandswap. Is there any reason for not having that action.

[spoon16]

@chandrapersonal I don't think "compareandswap" is a type of update notification. It's part of the Etcd update API and the side effect is that an update notification is emitted, either set or create.