Fault tolerance for multicore architecture

Some, more expensive, server systems have multiple CPUs and one 'backup' or 'failover' CPU in case one of the others dies. I think that's what the OP is referring to.

I'm not sure if FreeBSD would not crash, I presume it does. You can however run an SMP kernel on a single CPU or single core system.
 
gordon@ and SirDice, thank you.

Yes, that's what I was referring to. I want to know in the context of multi-core architecture on a desktop computer, suppose a hardware fault occurs in one of the cores, does the FreeBSD OS do anything to schedule the tasks of the faulty core to the other cores? If not, how can I make the OS achieve this (ie not by using redundant hardware)? I'm not sure if this is the right place to ask this question, could you please guide me about where to ask/resources to get started? I'm sure there's a lot I have to know first.

SirDice, I read about SMP. With SMP on multi-core computers, if one CPU fails, the entire SMP system is down. Clusters of two or more SMP systems can be used to provide high availability (fault resilience). If one SMP system fails, the others continue to operate. Is this what I have to do?
 
I'm not entirely sure but I think FreeBSD will simply panic(9) if one of the cores dies. At least that's what happened to the dual AMD Athlon I had a few years ago. But after taking out the faulty CPU the system booted without any issues or changes.

Your question is probably a bit too 'technical' for this forum. I think you'll have more luck getting good answers on the @freebsd-questions or @freebsd-hackers mailing lists.
 
You must log in or register to reply here.
Back
Top