Failure handling mechanism in Octopus

Hi, Request your response for below query on priority basis:
Can you please let me know what happens when a particular step in the Process fails for a specific machine. Does it continue with the deployment steps for the remaining machines or it just halts the overall Deployment process.
So lets say there are 50 App servers and their 200 Client machines (4 clients per App Server).
And as per Project’s process steps,the release first gets deployed on all the machines with role as ‘App Server’ and then followed by deployment on all machines with role as ‘Client’ . In below case, what will happen if step b of the Parent Step 2 (for AppServer) fails on any of the machine for any reason?

Parent Step 1 (preliminary activities for machines with roles App Server or Client)
Step a
Step b
Step c
Parent Step 2 (for all machines with role ‘App Server’)
Step a
Step b
Step c
Parent Step 3 (for all machines with role ‘Client’)
Step a
Step b
Step c

Note: Each environment consists an App Server and its clients. So in this case there are 50 environments and Release will be deployed in 50 environments.

a) Will it halt the over all deployment process?
b) Will it fail the Parent Step 2 and proceed to Parent Step 3? i.e. skip the deployment on App Server and instead continue deployment with Client machines?
c) Complete the deployment process excluding for the impacted machine

If the answer is ‘c’, then I am happy with it, but if it is either ‘a’ or ‘b’ then its a very big impact and contradicts the warrant to use the Octopus for such bulk deployment in our case.

Hi Pratik,

Thanks for getting in touch! Any failure will halt the deployment. When it comes to rolling deployments they will finish on the machines it has started deploying on for the current rolling set (all children within the parent) then halt the deployment entirely.
We suggest that you use guided failure, then you can determine if the error is local to the machine and the rest should progress or if the deployment should stop entirely.

Warm Regards,
Vanessa

Thanks a lot Vanessa.
You highlighted a very useful feature. This will be really helpful in case of any failure handling.
But I just see one challenge . For example, if there are 10 steps (with Guided Failure enabled) that needs to be executed, and one machine is unavailable then if i select ‘ignore’ on 1st step,then i have to keep ignoring on all the remaining 9 steps so that the deployment completes on the other accessible machine. Is there anyway we can address this to exclude the faulty machine from the remaining steps?

Hi Pratik,

Unfortunately there is no way to exclude the machine from the remaining steps. We do have a similar UserVoice suggestion that you should vote on: https://octopusdeploy.uservoice.com/forums/170787-general/suggestions/8723770-automated-continue-on-failure-deployment-mode

Vanessa