On Mon, 29 Oct 2001, Rich C wrote: > 2.) Promise supports Red Hat with PDC20265 drivers in both UP and SMP > flavors.
Keep in mind that if Promise ever goes out of business, stops supporting Red Hat, or just plain drops the ball, you will be unable to upgrade your system any more, because of the lack of drivers. Furthermore, the Linux kernel people generally will not help you, since you are running a system with binary-only drivers. Personally, I avoid anything that does not provide source. It is for that reason that my next video card is probably going to be an ATI Raedon, rather than an NVidia GeForce. > ... the system wouldn't boot afterwards. I don't know if it was because > of the "funny" RAID superblocks ... A hardware RAID controller should not expose any of its internal housekeeping data to the OS (or other software). As far as the OS is concerned, you just have a single, normal drive. All the mirroring happens at a layer lower than that. > ... or if it was because my boot partition was bigger than 1023 > cylinders, but I downloaded the newest version of LILO and that fixed > that problem. Probably the 1023 cylinder thing, huh? I would say so. > Next was the ultimate test. I removed the Primary Master drive from the > machine, the one that would have been "sda" if the RAID wasn't working > ;o) and powered it up. Ah, but will the disk be so nice as to fail when the power is off? :-) For a real test, I would start some intensive task (e.g., kernel compile), and disconnect power from one of the drives. See if it lives through that. If it does, repeatedly, it might be worth trusting. > After rebuilding the array, it was time to get the SMP working. Red Hat > install didn't make an initrd image for the SMP kernel, since there was > only one processor in the machine at install time. I have does Red Hat installs on quite a few SMP boxes with only one processor installed, and in every case, it installed and booted the SMP kernel (with initrd) by default. If I had to guess, I would say the fact that you had to feed it drivers from diskette was the cause. > I was pretty sure the Viper didn't support AGP 4x, so I turned it off. > > Well, I guess that was the problem, because the machine has been solid > for 4 days now. Strange that under one processor it would work, and with > two it wouldn't. Can anyone explain this? SMP involves more components, and pushes other components harder. Perhaps something was just marginal enough to fail with SMP, but work with a single proc. Or perhaps there was some subtle incompatibility between the SMP and AGP components. > To sum up, how do I like having a 2GHz machine ... I am picking nits now, but that should be written as "2 x 1 GHz". One 2 GHz processor is not the same as two 1 GHz processors. Multithreaded applications -- and overall system performance -- will get a bigger boost from SMP. Single-threaded applications will get a bigger boost from a higher clock. Probably not so important to you, but it can be a very significant difference if you are tuning for a particular application. :-) -- Ben Scott <[EMAIL PROTECTED]> | The opinions expressed in this message are those of the author and do not | | necessarily represent the views or policy of any other person, entity or | | organization. All information is provided without warranty of any kind. | ***************************************************************** To unsubscribe from this list, send mail to [EMAIL PROTECTED] with the text 'unsubscribe gnhlug' in the message body. *****************************************************************
