Is RexRay actively maintained? #1338
Comments
|
Hello Catalin,
Have you looked at CSI? CSI was introduced a couple of years ago and is the
focus of numerous storage providers for interop with K8s/Mesos. Generally,
community PRs are welcome across the repo and releases are as needed at
this point. Our focus is on parts of REX-Ray such as the `gocsi` library.
…On Mon, Oct 28, 2019 at 1:19 AM Catalin Jora ***@***.***> wrote:
Summary
As I don't see any activity since January 2019 on this repository, I'm
wandering if the project is still maintained. My use case is the
integration with open-source DC/OS external persistent storage
<https://docs.d2iq.com/mesosphere/dcos/1.12/storage/external-storage/>.
Will new feature be added?
Will new bug be fixed?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#1338?email_source=notifications&email_token=AAKUYMIXH2E2PO6K5AKFSELQQ2OC5A5CNFSM4JFW3ZB2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HUWFR6A>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAKUYMN7O5PXMCV42ENQLPLQQ2OC5ANCNFSM4JFW3ZBQ>
.
|
|
Looking at #1337 you could probably draw your conclusion there. |
|
Sigh. Literally went live with this solution last December for a large open source project. I ended up scaling it back due to #1331 glad I did :-| Anyone interested in forking / maintaining or taking over the project from the original author? (if you can find them) |
|
Sorry for this @kallisti5. If there are community members that are interested in maintaining please do let me know. |
|
@kallisti5 what did you use instead? @clintkitson I would love to help but don't even know what language rexray is written in :/ Looks like you are the last "active" maintainer of this repo, is that correct? |
|
@clintkitson I'm willing to help maintain this project. Kinda sucks most of the CSI available implementations are done mostly for k8s and e.g. in case of AWS ECS we have to use RexRay for ECS <-> EBS integration, so RexRay is great here. |
|
Glad to hear you're making use of it @Trojan295. Do you have any thoughts about the direction you would want to take the project? |
|
If the rexray maintainers are willing to give up control that would be useful, otherwise a fork wouldn't be a bad idea. It's a shame how projects like these get abandoned... probably just the company shifting priorities but still... at the very least hand it over to someone that cares |
|
@clintkitson Given the lack of clear direction from Mirantis and the closed sourcing/deletion of CloudStor documentation I'd really like to see this pick up as it is the only remaining solution for Docker Swarm (especially for AWS and OpenStack).
I have my own fork that brings in several PR's for and a few of my own fixes for the OpenStack Cinder but I'd be happy to work with a few folks bring them and many of the others in. I've noticed that the community (especially the Swarm community) is starting to just take for granted that we are "second-string" despite the fact that most of us are firm believers that it is a better solution (for us anyway) than K8s, mesos etc. I have experienced what Mirantis support looks like, (at least in the short term) and if we want to save Swarm we are going to have to do it ourselves. /rallycry |
|
I am open to having a community Zoom call to further discuss. Please thumbs
up if you would attend and reply if there are any special considerations
for time etc.
…On Thu, Jul 9, 2020 at 2:07 AM Floris van Lint ***@***.***> wrote:
@Enquier <https://github.com/Enquier> yes please. We make use of docker
in swarm mode extensively using Caprover
<https://github.com/caprover/caprover> and have now reached the point
where we really need swarm mode data persistence, which we planned to
achieve using Rexray/s3fs.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#1338 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAKUYMNLCDP63TCPQWY7FX3R2WCGTANCNFSM4JFW3ZBQ>
.
|
|
@Enquier @clintkitson I am willing to throw my hat in to help with keeping this project going. I'm actually using the Hashicorp Nomad Orchestrator on Openstack and this projects Cinder Docker Volume Plugin makes life so much easier. Yeah, Nomad is slowly adding CSI support but, even though it's supposed to be a "platform agnostic" specification, I have yet to come across a single CSI driver that documents its deployment process. The docker volume plugins however are truly orchestrator agnostic and it took me about 3 minutes to find the information I needed to install/configure the plugin and have migratable persistant storage across my cluster. @Enquier Can you share what fixes you made in the cinder driver (as mentioned before, this is what i'm using Rexray for)? I started digging into the Cinder driver to look into adding Application Credentials support but was curious if that may have been one of the things you worked on in your own fork? |
|
@cPanelRicky Check-out logitainer/rexray I wasn't sure if @clintkitson was more interested in handing off or starting a fork so I moved mine over to that as a starting point in the event of option #2. I'm in the process of updating the whole stack to use Go Mod as opposed to Dep so let me know if you encounter any issues (see logitainer/csi-nfs,csi-blockdevice,csi-vfs etc) |
|
@Enquier Perfect, appreciate it! I started attempting to migrate over to Go modules as well but found that there are going to be some modules that have become incompatible over the last year or so (mainly the Gophercloud module) so that's gonna take some work to get fixed up. |
|
Just following up, busy summer. Anyone have a feeling one-way or the other? |
|
@Enquier Yoo hello, Im interested, Im building a PoC using Docker Swarm and I was searching for a persistent storage solution until the algorithm showed me this repository where can I found a list of issues we should solve? how should we prioritize them? |
|
So @Enquier how did this turn out? I'm looking at logitainer/rexray and it look stalled :( |
|
I have the same issue @jostmart. Running a swarm cluster with nodes on VMWare platform. We sadly can't use the VSFD plugin : https://github.com/vmware-archive/vsphere-storage-for-docker. I'm in need of a storage provider that can use vSphere as a provider. Has anyone tried to create a vSphere provider for REX-Ray ? |
|
Apparently, VMware maintains this. https://blogs.vmware.com/cloudnative/2018/06/22/rex-ray-persists-your-questions-answered/
|
That's quite unexpected. But cool! |
|
Note that blog post is nearly two years old. I’m fairly certain that VMWare
has since created a separate CSI based plugin elsewhere.
…On Sat, Oct 24, 2020 at 06:40 Floris van Lint ***@***.***> wrote:
Apparently, VMware maintains this.
https://blogs.vmware.com/cloudnative/2018/06/22/rex-ray-persists-your-questions-answered/
Q. Is REX-Ray still being maintained? A: Yes, by VMware.
Q. To what extent is VMware maintaining REX-Ray? A: VMware has committed
multiple open-source engineers to REX-Ray.
Q. I’ve been a part of the REX-Ray community for a long time, where should
I go for continued support and collaboration?
A: The REX-Ray GitHub repository.
That's quite unexpected. But cool!
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#1338 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABNEWT6FW6R6O4Y425FU7K3SMLDKZANCNFSM4JFW3ZBQ>
.
|
It’s still there. It’s still my intention that it be maintained but I was looking for a signal one way or the other from the original team. It seems to me the most valuable piece here is maintaining the CSI compatible interface for swarm and that should allow other CSI plug-ins for K8s PV to work for swarm. So factoring that out and then identifying working solutions from other sources seems to be the right approach |
Yes but VMWare gave a commitment to the Open Source community. Therefore they are obliged to ensure the project is maintained (nobody is asking for new features, just bugs fixed, maintenance releases and commitment adhered to). |
|
I’d appreciate any support from VMware but at this point all of us using non K8 solutions are on our own. So, I’ve expressed willingness to start working the crank. I have a solution that works for me over at logitainer/rexray and other than working to add native efs support I don’t have a clear feature path. So at this point we have two options, either fracture out and start pushing CSI at our individual orchestration platform level at which point we’d still want to maintain the CSI features that are here but nowhere else (Ie cinder and S3) or continue on to rexray as a platform for CSI. I think the best solution is a hybrid so, I’d like to start by getting some of the PR’s closed out. So I would put out a call to anyone who made a PR to clean them up and reprise them in Logitainer and we can start to triage them. We will need to sort out the build pipelines from there |
|
Hello, I saddened to see the status of this project 😢 . I come here because I just ran into an issue with the dobs integration. I looked at the code and opened a PR only to 5 minutes later realize that someone else had opened another PR for the same issue with a very similar solution. Would it be possible to merge it? #1344 Also, I've tried building the project locally but I'm unable to do it. Help? |
No, they are not obliged to do anything. Making software Opensource does not mean there is a contract to maintain or even pay attention to your issues, you are however free to ignore certain contributors or companies in the future. Depending on the Opensource license you are however free to fork the project and run it as your own. It would be really polite of them to warn others that this project is stale and no longer developed, but again there is no rule that they should. It is up to you to judge them and consider alternatives if you come across a solution of theirs. Not much else you can do. |
|
@clintkitson the website and installer are offline as bintray.com is EOL. Any chance this could be resolved? |
|
is it possible to build and host the binaries in github releases instead of bintray ? Since rexray provided instruction to build the go binaries, that isn't difficult to do. I believe many people are getting Forbidden when they reach out to bintray.com |
|
I've had very poor experiences with Rexray in the past and do not recommend it. Rclone offers a much better (maintained!) volume driver for Docker Swarm: https://rclone.org/docker/ Have been using this volume driver in production with no issues for well over a year. |
|
Is there an alternative to using Rexray? |
We moved to https://github.com/sapk/docker-volume-rclone first, then later migrated our entire Docker Swarm setup to https://github.com/rancher/rancher with https://github.com/longhorn/longhorn, but there is an rclone-CSI as well https://github.com/wunderio/csi-rclone. But as @griffmick said, rclone has a native volume driver, which is probably the best directly comparable alternative to rexray. #1338 (comment) |
|
any of these support EBS? (#1282 (comment)) |
and others
Summary
As I don't see any activity since January 2019 on this repository, I'm wandering if the project is still maintained. My use case is the integration with open-source DC/OS external persistent storage.
Will new features be added?
Will new bugs be fixed?
The question is mainly for the the vmware Rex-Ray project maintainers: @codenrhoden @akutz @clintkitson
The text was updated successfully, but these errors were encountered: