Some providers do not provide all the information (CPU frequency, GPU availability, ...) through API about a Host template. So in order to allow the Host selection to be more precise, the SafeScale Scanner will investigate a given set of templates to register some useful information in a local database.
SafeScale Scanner used to be a separate binary. Starting with release v21.11, it's now a service included into safescaled
that can be requested using safescale
command (you can find usage here or directly through gRPC API.
To launch the scan, just launch the command safescale tenant scan <tenant_name>
.
To be scanned, a Tenant should have the field Scannable set to true
[[tenants]]
name = "ovh_tenant"
client = "ovh"
[tenants.compute]
Scannable = true
WhitelistTemplateRegexp = "<regex>" # Ex: "s1-*"
BlacklistTemplateRegexp = "<regex>"
...
...
To scan one or more specific templates from a given tenant, you can use the --template <template_name>
(alias: -t
) option.
A database of all templates scanned (with the added details) will then be stored in $HOME/.safescale/ of the safescaled
command, allowing SafeScale to create hosts more precisely. (not implemented yet)
To filter templates to scan (forbid some templates and/or allow some others), you can set the WhitelistTemplateRegexp
and/or BlacklistTemplateRegexp
keywords of tenants file (tenants.{toml|json|yaml}
), with regular expressions.
Please be aware that a scan is specific to a provider and to a region, as templates can vary with regions and providers.