> I'm interested in creating a Linux Kernel Beta testing site, with number
> of hosts and QA people to do an extensive test for new ( and old ) beta
> versions of the kernel.
>
> I would appreciate opinions about (a) if this is a good idea at all, and
> (b) how should such an operation work ?
Good idea actually, although yes, it is probably going to be expensive ;-)
However, may I suggest that besides your own lab, you should also try to
provide a more useful QA/bugtracking system for normal users, where they get
guided through some sort of "Bug/Performance Wizard". This is a highly
sophisticated project, since it needs to divide mere stupidity from hardware
errors and real software bugs, which requires a good understanding of Linux
and statistics (at least I would think so). On the other hand, I think this
might be the only viable option to gather useful data about a large and varied
hardware base at all.
If you would ask me what else to focus on, I would suggest that a good first
step would be an extensive system call testsuite and regression test. I would
recommend staying with the kernel first, the testsuite could later be extended
to also verify glibc. (Yes, this is a huge task)
It will also require a lot of manpower to arrive at useful conclusions from
all the raw performance and works/does not work data.
Sincerely,
Lars Marowsky-Brée
-- Lars Marowsky-Brée Network Managementteuto.net Netzdienste GmbH
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/