Exclude from autoscale for a machine in drain mode

Avatar
Vishwas Nomula

We have noticed that when we put a machine in drain mode (deactivate), autoscale is taking precedence and making machine active after some time due to which we are unable to perform maintenance just by using "deactivate" button. As a work around we are following 2 step process to get around this. 
1. deactivate machine. 
2. Exclude machine from scale in and out operations. 

so ask is , it would be good have check box for "Exclude machine from scale in and out operation" at the same place when deactivating machine.  

9

Comments (4 comments)

Sort by
1
Avatar
Elizabeth Webb

Yes please!  The whole point of Deactivation / Drain mode is to keep the host out of circulation until you say it's ready to go again.  We often have to wait several hours for all our user sessions to complete before we can start maintenance, so it's imperative the hosts remain in Drain mode.   Having to take an extra step to do this is completely counterintuitive.  At the very least have the check mark to exclude the host available in the dialogue box like you do for the manual Power On / Power Off dialogue boxes. 

Thank you!

1
Avatar
Toby Skerritt

Hi both, thank you for this suggestion. We have captured it for review by the development team. 

0
Avatar
Bjørn-Ingar Trønnes

This morning we had some issues where this feature was deeply missed when a host had issues and was deactivated by our Helpdesk. By the time Helpdesk had reported the issue to our Delivery Services, auto scale had activated the host and we received several more calls from users having issues on the same host.

I would like to see one of two options in future releases:

1. When a host is deactivated, it stays in drain mode until it is explicitly activated

2. The option to exclude from scaling indefinitely or until a specified date/time (same options as if I wanted to shut down the host).

0
Avatar
Dave Clegg

Absolutely agree with this - it almost defeats the point in having drain mode when its just ignored and reversed by auto-scaling.  We too have had more than one incident where Helpdesk have deactivated a host which had issues but auto scale has activated it back into the pool before it can be escalated up to teams to deal with the bad host, resulting in more issues.  

Please can either of the suggested options be looked at being added asap!

Please sign in to leave a comment.