How to restart application rather than stop and recreate a new one?

[newbacai]

How to restart application rather than stop and recreate a new one?
I tried:

k3s kubectl get pod {podname} -n {namespace} -o yaml | kubectl replace --force -f -

But not working, it returns:

zsh: command not found: kubectl

 

[Samuel Tai]

You forgot to prepend k3s in the 2nd invocation of kubectl.

 

[newbacai]

You forgot to prepend k3s in the 2nd invocation of kubectl.

Thanks so much. But I found that the system still remove and build a new one:

Is any possbile to keep the same pod?

 

[stavros-k]

Is any possbile to keep the same pod?

Pod dies the moment the service inside it is stopped/crashed/killed.
It's not designed to be restarted. It's designed to be recreated into a new pod.

It might be better explain what you want to achieve, so people can help you do it right, rather than trying to "hack" a solution

 

[newbacai]

Pod dies the moment the service inside it is stopped/crashed/killed.
It's not designed to be restarted. It's designed to be recreated into a new pod.

It might be better explain what you want to achieve, so people can help you do it right, rather than trying to "hack" a solution

Thanks for your expianation. Here is what I want to achieve:

1. To visit some service I have to run "export http_proxy=xxx" each time when a new container / pod created because I'm in China where many network service including truenas has been forbidden.
2. I need some tools like "iputils-ping" but many images don't have them. I have to install it manually each time.
3. Last but not least, I need to run crontab jobs, cause the images don't support it, I was using "docker exec" in the TrueNAS Scale host. If the name or id changed each time, I have to modify the crontab then.

 

[stavros-k]

To visit some service I have to run "export http_proxy=xxx" each time when a new container / pod created because I'm in China where many network service including truenas has been forbidden.

This is "easy", as for the most part of containers/images you use Environment Variables to set things.
In a normal k8s deployment what would be couple of lines in YAML.
In case of SCALE Apps, app developer, has to expose such fields in the UI.

I don't know if official apps has such fields, but TrueCharts apps have for sure.

I need some tools like "iputils-ping" but many images don't have them. I have to install it manually each time.

This is hard as not all apps allow you to install packages into an image/container. (As containers are not designed to do that)
It might be "ok" to do once in a while for troubleshooting, but if it's needed more often, you might want to ask the upstream container creator to add it to the container.

Last but not least, I need to run crontab jobs, cause the images don't support it, I was using "docker exec" in the TrueNAS Scale host. If the name or id changed each time, I have to modify the crontab then.

This is also able to be implemented in k8s (without the need of cronjobs in Host),
But I don't think there is such an option in official apps Scale GUI, For sure there aren't in TrueCharts apps.

But generally if an app needs a cronjob for core things, TrueCharts apps has CronJobs defined in the background.

eg. Nextcloud. , Firefly etc

If you want to give TrueCharts apps a shot, you can reach us for support in our Discord

TrueCharts

Your Source for Awesome Helm-Charts!

truecharts.org

 

[newbacai]

This is also able to be implemented in k8s

Okay, I'll google it. Thanks a lot.