Is this project abandoned?

[lsadehaan]

Is this a BUG REPORT or FEATURE REQUEST?

Choose one: BUG REPORT or FEATURE REQUEST

Versions

MySQL Operator Version:

Environment:

• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• OS (e.g. from /etc/os-release):
• Kernel (e.g. uname -a):
• Others:

What happened?

What you expected to happen?

How to reproduce it (as minimally and precisely as possible)?

Anything else we need to know?

[bweston92]

see #255

[lsadehaan]

So I guess the initial release was just very good!
The reason I asked is because:

1. The git front-page indicates that the latest codebase doesn't compile
2. There's a lot of pull requests that seem to hanging around since 2018
3. There's quite a few issues without replies
4. I will be using Kubernetes 1.17 - it appears to not support anything newer than 1.15? There is a pull request to support 1.16 but it seems to be stuck? Is there a plan to maintain support for newer version of Kubernetes?
5. It is at v0.3.0 - is there a plan/roadmap for getting to v1?

I am leaning towards using Mysql 8 with innodb cluster - and this project, especially since it is from Oracle was my goto option, but I am a bit concerned. I guess I'll just have to give it a go and see how far I get...

[jmordica]

Any updates on supporting newer versions of Kubernetes?

[bweston92]

@owainlewis is there any plans to support newer versions of Kubernetes?

[lsadehaan]

see #255

Note that the this original issue asking whether this project is still alive is more than a year old already and the developer that replied saying it is alive back then has told me now that unfortunately it seems to have died and that it is a real shame since there are no actively maintained operators for innodb cluster at the moment.

[bweston92]

Does anyone know if the license allows us to fork? Seems that #308 by @GLYASAI has done a lot of work to support K8s v1.16+

[bweston92]

There is also https://github.com/presslabs/mysql-operator which looks actively maintained.

[lsadehaan]

There is also https://github.com/presslabs/mysql-operator which looks actively maintained.

That is not based on InnoDB Cluster but rather async replication and Percona DB (mysql 5.7)

[bweston92]

@lsadehaan ahh ok, @owainlewis / @prydie can get some maintainers outside of Oracle to start helping this? I am unable to upgrade clusters pass Kubernetes v1.15 at the minute.

[chinglinwen]

I found a critical issue for this project, even for multi-master mode (of three nodes), if the r/w node down, it can't fail over to new node(even after 40 minutes), and so can't write any new data to the mysql database.

it cost me times to found this. hope this may help others.

after 40 minutes. still can't getcluster.

MySQL  localhost:33060+ ssl  JS > dba.getCluster().status() 
Dba.getCluster: Unable to find a primary member in the cluster: No primaries found in InnoDB cluster group (RuntimeError)

node restarted, after a about 40 minutes, it still can't recover somehow,

I0730 03:39:41.745554       1 cluster_manager.go:341] Found existing InnoDB cluster (metadata exists, but GR is not active)
E0730 03:39:42.236654       1 cluster_manager.go:135] Error bootstrapping cluster: rebooting cluster from complete outage: SystemError: RuntimeError: Dba.reboot_cluster_from_complete_outage: The MySQL instance 'mysql-cluster-1.mysql-cluster:3306' belongs to an InnoDB Cluster and is reachable.

I'd suggest someone do the same test of ha failover ( for r/w node shutdown test ), before using it.