On Thursday 17 August 2006 14:07, Andy Green wrote: > Yes they tell you 'it's broken' and then there is a silence while you > wait to hear why they think it's broken, and they wait for you to fix it. Yep. Amazing how they think we can read minds huh? > > Well you are firing on a lot of cylinders there the only thin difference > is are you running a kernel from the last 6-8 weeks or so? I guess not > on those unrebooted Opterons. I don't have any basis for my suspicion > except that the behaviours on my box of knocking two HDDs to death from > different manufacturers and the fact it was fine for a good long while > beforehand. And I guess I know there is libata and kernelside work to > do with ATA in kernels around that time. But it can as easily, well, > more easily, be a powersupply problem in my case. > Actually, I do have 3 FC5 x86_64 boxes that I just updated a week or so ago and rebooted to the latest kernel. So far, I've need so trouble with them. 2 have SATA drives, one is an IDE boot drive but 2 3Ware SATA RAID cards holding about 2.5TB data out of 4TB usable that we use as an archive location for the last 7 days' worth of NWS radar data for our friends at NCDC. They have the largest repository of weather data on the planet and if they have a problem, they can get any missing data from us. (We provide them the primary feeds as well.) These boxes don't get hammered like the others, but they aren't eating drives either. -- Fere libenter homines id quod volunt credunt. Mark Haney Sr. Systems Administrator ERC Broadband (828) 350-2415
