On Fri, Jan 17, 2003 at 12:55:04PM +0100, zeist wrote: > Greetings > > I'm facing serious problems with a Raid partition on UATA133 bus. > The box is a dual athlon MP with asus MB and REDHAT 8.0. > The entire system except /home stand on two IBM 80Gb disks (on uata 100 > bus) in Raid 1 with ext3. > The /home partition stand on 3 Maxtor 120Gb "diamondmax plus 9" uata133 in > Raid 5 with ext3. > The system ran fine with 2.4.18-18.8.0smp official redhat kernel and the 3 > Maxtor disks attached to and external Adaptec controller based on > highpoint hpt370 chipset (note, i used just the ide bus functionality, > since Raid was software). I think you're starting to realize why we like the Red Hat kernels. Red Hat does a *lot* of work testing and patching the kernels so that they work well in different configurations. Many patches don't appear in vanilla kernels until a future release. > When i switched to 2.4.19 vanilla kernel i started to have problems, first May I ask why you switched? Although the release number went up, you lost a lot of patches that Red Hat had applied. What problem were you trying to solve? My suggesting at this point would be to restore your system back to the way it was before you had problems. Depending on the level of corruption you've already created, you may need to re-init the disks and restore the files from backups. > Theres nothing to do, problems persist, not only during high i/o > throughtput, but also when i start to store large data on partition. > I fell back to 2.4.18-18.8.0smp I didn't install the promise patch since > it seem to me that is already included in rh kernel) without succes, it > seem that when the data on disks reach about 30% of capacity troubles and corruptions You should also go back to your original controller. After all, it worked! > I also thinking about basic design errors, since i have 3 uata 133 disks > attached to a single pci controller all with dma access activated, > somebody can confirm that this could represent a bandwith problem? Even if you have a bandwidth problem, you should not have corruption. At worst you should slow down, not corrupt data unless you have have a faulty motherboard or controller. -- Ed Wilts, Mounds View, MN, USA mailto:ewilts@ewilts.org Member #1, Red Hat Community Ambassador Program -- Psyche-list mailing list Psyche-list@redhat.com https://listman.redhat.com/mailman/listinfo/psyche-list
