I've installed Andrea's BIGMEM patch and got 2.2.14 to recognize all 4GB
of RAM. 2.3.40 does not have the same "panic on PCI probe" problem that
2.3.39 had. Intel's e1000 gigabit driver seems to be working with both
2.2.14 and 2.3.40. This is a smokin' machine!
The machine is going to be used to see how high a load we can get with a
certain piece of software we're going to be running on it. The software
will spawn a new process for each user that connects to it. We would
like to be able to get 5-10K users running from this machine, and that
would obviously mean that we'll have 5-10K processes running on this
box. Is this going to cause us any problems? I've paid attention to
the past issues like the "thundering herd" problem. Are those still an
issue with 2.2 kernels? Are there patches we could apply to 2.2 to
better handle thousands of processes? If we deploy these machines,
we'll almost certainly have to be using a "stable" line of kernels if
only because we won't be able to bring the machines down periodically to
install a new, more stable kernel release so we'll probably have to
stick with 2.2, although I imagine we'll be able to get away with a few
patches like the BIGMEM patch and any possible scheduler patches. If it
comes down to it, we'll take better performance with thousands of
processes over better performance with only a few since these machines
will be dedicated to just their one job.
-- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- With Microsoft products, failure is not Derek Glidden an option - it's a standard component. http://3dlinux.org/ Choose your life. Choose your http://www.tbcpc.org/ future. Choose Linux. http://www.illusionary.com/- 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/
This archive was generated by hypermail 2b29 : Mon Jan 31 2000 - 21:00:13 EST