search

soabase / exhibitor

ZooKeeper co-process for instance monitoring, backup/recovery, cleanup and visualization.
https://groups.google.com/forum/#!topic/exhibitor-users/PVkcd88mk8c
Apache License 2.0
1.68k stars 444 forks source link

This project appears to be abandoned. Is anyone paying attention? #368

Open pdehlke opened 6 years ago

pdehlke commented 6 years ago

No PR has been merged or even commented on in two and a half years. Is anyone paying attention here?

Randgalt commented 6 years ago

The project was opened for community maintenance last year. These are the people that signed up to maintain it: https://github.com/orgs/soabase/teams/exhibitor-committers/members - we can always add more. If not, sadly, the project will die.

gitrc commented 5 years ago

I reached out to @xiaochuanyu in hopes of assisting with maintaining this project and merging PR's.

xiaochuanyu commented 5 years ago

Unfortunately I no longer have time nor much motivation to maintain this project anymore. We moved off of Exhibitor where I work since much of what you get with Exhibitor can be currently implemented with other tools or newer versions of ZK itself. We currently use newer version of ZK with auto purge (since 3.4.0) for cleaning up data directory, supervisord for supervision and zookeeper CLI for viewing data. In terms of cluster-wide config changes, we've been just doing that semi-manually with scripts.

clehene commented 5 years ago

We had been using Exhibitor at Adobe while building a distributed operating system based on Mesos, compatible with DC/OS (https://github.com/cell-os). We needed something to bootstrap the infrastructure "microkernel" - basically once VMs were provisioned we had Exhibitor manage Zookeeper to support Mesos workloads and everything else was then deployed on top of Mesos using cluster schedulers.

It's been useful, but a bit hard to solve some of the issues. I've went through the codebase and started refactoring it a while ago and had some notes (still have them somewhere) about what I think should be done. I didn't get a chance to do all the work eventually, although I wish I had.

Since then we moved on. both as a company as well as personally :). I'm currently using Kubernetes on GKE and wouldn't go back TBH. I think it could be useful to put the project in "maintenance" - something like ASF Attic (https://attic.apache.org/) and have people know about the current status.