Quality of FreeBSD

[Robert Watson]

On Thu, 21 Jul 2005, Nicklas B. Westerlund wrote:

> Although I havn't seen any major problems on our servers, all using u320 
> scsi and smp - I don't feel as secure about my choice of upgrading to 
> 5.x. We still have some 4.x servers in production, and judging by how 
> this is evolving, I think I'll rather skip the 5-branch for those 
> machines and keep testing 6.x. The last thing we need is servers with 
> problems to disturb our sleep at night.
>
> Overall I think we're a few of the lucky ones, as alot of people seem to 
> have huge problems which we havn't encountered, again that is because of 
> different architectures and such.

Actually, I think you're part of the silent majority who find it works 
fine in their environment.  We use RELENG_5 at work on a number of 
machines, and I work with several companies and organizations who do, and 
have no problems at all.  The edge cases seem to be:

- High load environments, or high load testing.

- Hardware that isn't part of the regular testing that FreeBSD developers
   do as part of their work, likely because they don't have the hardware.

- Less commonly deployed features -- i.e., IPX, which has experienced
   serious functional problems in RELENG_5 until a few months ago.
   Interestingly, resulting from a compiler change, not network stack
   changes...

Robert N M Watson