Application Management should have an option to automatically retry failed deployments. We are managing a non-persistent environment using a scheduled nightly reimage, and with application management steps sporadically failing and unwilling to run again for 24 hours, this could be come a major issue as we continue to migrate users to this platform. There should be options to resume, cleanup and rerun, and possibly even to reimage the host in order to ensure that VMs are ready for the users at the start of the next workday.
Automatic retry for Application Management Errors
1
yes, and the worst is task of adding host is reported as successful, but without successful app installation actually it is not:
Hi Erik Langeland - thank you for this suggestion, I will raise this with the product team. I agree, waiting 24 hours is not a good result. I will see what options we have for configuration.
Marek Sokół - thank you for your comment, what would your preferred outcome be? In your scenario there may be a terminal issue with the application, therefore (assuming auto-scale is enabled) if the desired behavior is to delete the host, or otherwise make it inaccessible, auto scale would engage in a loop of creating new hosts and deleting them, potentially meaning that no hosts would be available to users, which seems like a worse default outcome.
It would be reasonable to have a limit on the number of retry attempts, similar to options for fixing AVD Agent failure states. It would also be good if the system could fire us a status notification to let us know when there are unresolved failures that may need to be addressed.
Thanks for the suggestion!
Most important is session host can not disable drain mode (can not be activated). Second adding session host task can not be marked as successful. It will be a nightmare if I schedule rebuilding hostpool over night and next day I have broken environment next day (no apps = broken environment)
Im my environment second attempt can be successful of installing the apps.
If second (or third) attempt of installing application is also not successful, session host should be deleted. Then scaling plan will kik in and try tu build new session host from scratch.
additionally, issue with app installation, should stop rebuilding hostpool process according to what is set in re-image wizard:
These are all great ideas, thank you! I will discuss all with the development team.
Hello Toby,
What is the status if this request?
Today when I logged in I found 1/3 session host in host pool with issue during application installation. But session host was normally power on by scaling plan and assigned to user.
Please sign in to leave a comment.
Comments (8 comments)