Re: suggestion for reliable boot: disk-by-id
The cited problems could have solutions / workarounds documented.
In our data center, the problem nr 1 was not heard of for years. It was alerted
at the article  that changing hw changes the id.
Regarding problem nr. 2, we already decided to not directly convert linux
physical machines to vm. Instead, redeploy them from scratch using documentation
and internal custom apps repos, eventually updating them. A good oportunity to
Regarding problem nr 3, we decided to use XenServer, that also uses disk by id,
afaik. But for who uses kvm it could be an issue.
Maybe, a solution for all cases could exist, but until now I guess that disk by
id solves more relevant problems at data centers than causing new ones.
Is there a complete solution?
Is there a way to choose methods at expert mode?
Is there at least a way to alert at d-i screens those who stuck at the boot