Skip to content

x/build/cmd/makemac: automatically power cycle dead MacStadium VMware hosts #32033

@bradfitz

Description

@bradfitz

Our MacStadium VMware host machines have a tendency to die or somehow wedge and stop reporting to vSphere.

MacStadium has an API now to hard power cycle machines:

https://www.macstadium.com/features/api

Sever Action (Power) by IP
URI: https://api.macstadium.com/core/api/servers/{id or ip_address}/{action}
Request: GET
Params: Action can be ‘On’,’Off’, or ‘Reboot’ (case insensitive).
Return: String. ‘True’ on success or an error message.
Example:

curl -H "Accept: application/json" -H "Content-Type: application/json" -X GET -u user:pass https://api.macstadium.com/core/api/servers/192.168.1.101/reboot

We should make our x/build/cmd/makemac daemon monitor the healthy status of our 10 VMware nodes (using the govc API or just whether they've registered with the coordinator in the past N minutes) and if we suspect a node is dead, we hard-reboot it with the API.

/cc @andybons @dmitshur @bcmills

Metadata

Metadata

Assignees

No one assigned

    Labels

    Buildersx/build issues (builders, bots, dashboards)FrozenDueToAgeNeedsFixThe path to resolution is known, but the work has not been done.

    Type

    No type

    Projects

    Status

    Done

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions