0

When we deploy an autoscale group in SoftLayer, there are chances that we see an autoscale group has been scaling for hours or even days. Also we saw in other times they were suspended for some unknown reasons. Could you list a few typical reasons that an autoscale group endlessly stays in either "scaling" or "suspend" mode?

Based on my experience, this happens when I used a very large image, or created a very large VM instance. This also happens when our account hits a hourly limit that doesn't allow the current autoscale group to continue to be created. There are also chances that the compute resources, such as cpu, memory or disk has used up, or almost used up in the data center that I am about to create the autoscale group. Is there a list of full reasons that we can refer to, in order to better plan future use?

Fan Zhang
  • 59
  • 7

1 Answers1

0

Check out this link: https://knowledgelayer.softlayer.com/articles/auto-scale-terms#status

Scaling: Used when any members are not yet provisioned or have an active provisioning transaction running. In scaling status, an auto scale group cannot be triggered for scaling and only non-actionable properties can be edited.

Busy: Used when any members have an active transaction but the transaction is not a provisioning or reclaim transaction. In busy status, an auto scale group cannot be triggered for scaling and only non-actionable properties can be edited.

Suspended: Used when a member is not scaling or busy, but has been set to Suspended. In Suspended status, an auto scale group cannot be triggered for scaling; however, any property can be edited. Any actions that would occur as a result of a group create or edit don't occur until the user resumes the auto scale group. An auto scale group can be suspended manually by the user or automatically if there is an error scaling the auto group for any reason. An auto scale group can only be resumed by a user.

Active: An auto scale group is active if it is not scaling, busy, or suspended. When active, an auto scale group value can be edited and all triggers can be invoked.

This is the documentation available about the status, that information could help you to understand better what is going on. If you see scaling is because there are provisioning proccess runing (It could be because your VM is very large and the provisioning takes long time, there are several members which have provisioning proccess running, the provisionings have not been approved yet).

Suepend is for any error, such as when the datacenter does not have enough room, the provisioning has failed or such as the reasons you mentioned, it could be several reasons for an error and there is not any documentation about all the reasons.

Regards

Community
  • 1
  • 1