Skip to content
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

Feature Requests #869

Open
Daxcor69 opened this issue Jun 5, 2022 · 3 comments
Open

Feature Requests #869

Daxcor69 opened this issue Jun 5, 2022 · 3 comments

Comments

@Daxcor69
Copy link

Daxcor69 commented Jun 5, 2022

Hello,
During my noob experience of configuring, installing and operating Talos I had some challenges. I had to ask my self how would have it been easier? Here is what I think would have helped me.

  1. Display kernel boot options
    The very first thing that should be displayed is what Talos got from the kernel boot options. I had an issue where my GW was wrong by a number and I struggled with why. It was there in the boot as an error message but it goes by so quick. I didn't see it.

  2. Boot time verbosity flag
    There are expected errors in the boot and config application phases of Talos. We get hammered in our heads, when looking for support, "what did the error messages say?" Well when I posted them, in some responses I got "oh thats normal". It would have been nice to see a progress report of the different stages of booting, hiding all the details, that at this point are irrelevant. Something like, "Network Setup and Working..................................... True". If we have issues then if we could enable the verbosity flag, and dig deeper. Something like talos.output.verbosity=t/f. To add to this, having the time syc message keep repeating in the console is totally pointless.

  3. BIG message - "In maintenance Mode"
    Your output does say at the end what the next steps are after the iso boots. My ipmi window was so small that I missed it as some wrapped making it hard to read, and some scrolled off the screen. Then when you are trying to sort it out, the time sync message keeps appearing and further scrolls the "whats next message". I would have a message that repeats. "************* You are in maintenance mode apply config now ****************"

  4. File system clean up
    Due to the above facts and other issues, I installed Talos many times. I used the wipe disk feature once and it took 2 hours for the task to complete. I would enhance your kernel param, and have talos.experimental.wipe with a few values.

  • full (wipe all drives with writing zeros to every sector and remove partitions.)
  • reset (only nuke the partitions on all devices)
  • system (only nuke the partitions on only the boot drive)
  1. Big Ask
    I would love to see a interactive configurator. The idea here is when we use talosctl init --interactive (i think) I would love for it to start a process of asking for your input to build your custom configurations. When we get the networking part especially, a configurator could ask some high values questions to ease the complexities of the configuration. You go over all this information in your docs, however, alot of times, the description doesn't make sense because it is out of context for the task you are trying to complete. If this is of interest, I would be better able to explain in a voice session. This sounds alot like a helm chart values.yaml file. But being interactive the next question would be tailored to your previous in put.

After reading this, you are saying we already do that, then I missed it. So that should add values to this post. Why did he miss it?

My perspective on this post is coming from "it should be dead easy to get a cluster up and running" This is not meant for the experienced it pro that charges 200/hr. With all that said, it was ALOT easier than most projects I have tried to get up and running.

The ramblings of a grumpy old fart on a Sunday morning drinking his coffee.

@smira
Copy link
Member

smira commented Jun 6, 2022

Probably points 1 & 2 are siderolabs/talos#4584

Point number 3 is siderolabs/talos#5613

Point number 4 is not clear, but you don't need to set .machine.install.wipe: true unless you need to do a full wipe. Sidero handles wiping on its own via the agent.

Point number 5 would be cool for a higher-level project on top of CAPI/Sidero.

@Daxcor69
Copy link
Author

Daxcor69 commented Jun 7, 2022

On point 4. On more than one occasion I had to load a small iso with gpartd and manually nuke all the partitions. the install would not work if I didn't clear the partitions. so the idea was to have an option to nuke partitions only without the two hour long wipe feature. That was the reason for the request.

@smira
Copy link
Member

smira commented Jun 7, 2022

On point 4. On more than one occasion I had to load a small iso with gpartd and manually nuke all the partitions. the install would not work if I didn't clear the partitions. so the idea was to have an option to nuke partitions only without the two hour long wipe feature. That was the reason for the request.

This issue is in Sidero Metal repository. Sidero Metal wipes the nodes automatically. Talos itself doesn't wipe the partitions by default to prevent accidental wipe, but it has reset API to wipe the node and get it back to initial state.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants