-
Notifications
You must be signed in to change notification settings - Fork 10
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added initial support for MySQL as a database backend #16
Conversation
@mheese Just realized that I left the DB password in plaintext in |
@maugustosilva awesome! great to see that this is coming together! To come up with good defaults for secrets, I always opted to generate some as a Kubernetes secret with an init job (similarly to what we do with the CA right now). This prevents that there are some default passwords floating around that attackers could try to gain access to instances like this. Additionally, like in the other cases, one should be able to specify a secret though which will prevent it from being created. This allows preprovisioning of it as well. To make things in the helm charts a bit easier, I was already thinking that we might simply want to eliminate the init job subchart and move it to the main chart. In general I used subcharts more for structural reasons and making it easier for others to look at the charts than anything else (well, and of course so that it is easy to enable/disable the deployment of them). As for updating dependencies in the helm chart, I always felt so-so about that. As MySQL isn't the "default" though, I think it is probably okay, as long as that is not a necessary step to begin with. Alternatively, we can just package up the version that we know works - depends on the licensing of the other helm chart of course. Personally I don't like to have an external dependency on availability of another chart, but that might just be me being paranoid. I will review the PR later this week! Unfortunately, this week I'm going to be busy with some other non-keylime things as well. |
@mheese thanks for your answers! So, on my last update, which I just pushed, I tried my very best to have re-usable randomly generated passwords. The code looks syntactically correct (making use of the "lookup" function), but something is not quite right yet (i.e., every time I call with I have attempted to use the trick described at helm/helm-www#1259 (comment), but couldn't finish it. Please take a look, and let me know. Here is what I would like to have:
Where an empty password would become a randomly generated password which is then stored into a |
Ok, managed to get something that works with auto-generated passwords. It survives multiple invocations of |
Since it was added as subchart (from bitnami), the following steps are needed before deployment. ``` cd build/helm helm dependency update keylime ``` In addition to it, tweaked a little bit the messages in NOTES.txt This uses randomly generated password which persist accross upgrades Here we also have a new Makefile, providing options to deploy/undeploy/update/debug Signed-off-by: Marcio Silva <marcio.a.silva@ibm.com>
63916be
to
50763d2
Compare
Since it was added as subchart (from bitnami), the following steps are needed before deployment.
In addition to it, tweaked a little bit the messages in NOTES.txt