Front wall in job configuration 😀

[Palooo]

Hello everyone,

In my case, I know it's not a problem with the rundeck core but i've seen some posts in this style so i'll allow myself... 😀

Here's my job, a very simple job, simply, i can't stop the server, make a positive return to rundeck (value 0) and continue executing other tasks after the reboot, in this case a sleep (to rebuild my cluster data) before moving on to the next node.

I tried :

sudo reboot and exit 0 (returns a positive value but does not restart the server)

sudo systemctl reboot (rebooting the serv but returns false value to rundeck so the job stop and don't continue) (I know you can override errors in node configuration, but i'd rather not use this function.)

And many other things that didn't work for me

btw.. Thanks in advance for the help !

[Palooo]

Here the job... not well uploaded in my first message, sorry ! 

[rac...@rundeck.com]

Hi!

Use the shutdown command instead of reboot. shutdown is more flexible and it seems script friendly. You can combine this with the “Refresh Project Nodes” step (configured with a sleep period) and then dispatch any other command. Take a look at this example:

- defaultTab: nodes description: '' executionEnabled: true id: 155c3b77-0b3b-4a91-b637-7cc9c27c13bb loglevel: INFO name: UpdateAndRestartDemo nodeFilterEditable: false nodefilters: dispatch: excludePrecedence: true keepgoing: false rankOrder: ascending successOnEmptyNodeFilter: false threadcount: '1' filter: 'name: remote_node ' nodesSelectedByDefault: true plugins: ExecutionLifecycle: {} scheduleEnabled: true sequence: commands: - autoSecureInput: 'false' passSecureInput: 'false' script: sudo shutdown --reboot +1 - configuration: sleep: '90' nodeStep: false type: source-refresh-plugin - exec: whoami keepgoing: true strategy: sequential uuid: 155c3b77-0b3b-4a91-b637-7cc9c27c13bb

Tested against Rocky Linux, feel free to modify it.

Cheers!

​

[Palooo]

Hello, thank you for you answer. 

Today I tried what you recommended, problem, the server restarts fine and returns a valid code for rundeck, but at the “refresh project nodes” step it moves on to the next node, waits 90 seconds and then sends the “sudo sleep 600” command on the first node, but in the meantime, at step two, it has moved on to node 2. My aim is for it to switch off and send a sudo sleep 600 as soon as it's switched back on, so that Rundeck waits 600 seconds before moving on to the next node. Thanks for your help ! 

I joined my job if ever...

[rac...@rundeck.com]

Hi!

Have you considered running the workflow in parallel on your node set? (Do you have to run that workflow node by node in your use-case?) That should work; edit your job, go to the Nodes page, and enter the number of nodes in the "thread count" field.

Regards!

[Corentin BND]

Hello,

No I can't run in parallel on each node (yes i have to run the workflow node by node), using rundeck on clusters, it would break the data/node sync.

I managed to solve my problem using the following plugin: https://github.com/rundeck-plugins/nixy-step-plugins

This allowed me to wait 60 seconds for the server to restart before executing further commands.

Thanks again for your help Racuna ! 🫶
Rundeck is a really cool piece of software! 

--
You received this message because you are subscribed to a topic in the Google Groups "rundeck-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/rundeck-discuss/7h-7q-KlqYM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rundeck-discu...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/rundeck-discuss/1268c535-1c6e-4cfe-a53f-e8f41d87482en%40googlegroups.com.

[rac...@rundeck.com]

Right, in that case that plugin is the first option. Thanks for sharing the solution!

Greetings!