söndagen den 6 oktober 2002 14.03 skrev John Allen: > Well I'm getting something similar. On my setup I have 2 drives, and old > 13G IBM, and a new 123G IBM. I keep getting hdb: dma_intr errors as per > yourself, but then it disables DMA on hda, and I no longer get the hdb > dma_intr errors.
One of my raid disks has fallen out too (one year old ibm disk bought in a batch), so I don't really know what has happened in my case... I have to investigate this further this week. > Anybody care to explain why DMA on the master is being disabled when the > dma_intr errors are apparently occuring on the slave? > > Oden Eriksson wrote: > >Hi. > > > >After some continous degradation in hdd performance I finally found what > > was causing it... > > > >---[snip] --- > >Oct 4 04:15:01 hostname kernel: hda: dma_intr: status=0x51 { DriveReady > >SeekComplete Error } > >Oct 4 04:15:01 hostname kernel: hda: dma_intr: error=0x10 { > > SectorIdNotFound }, LBAsect=4105991, sector=9328 > > > > > >Setting "USE_DMA=0" in "/etc/sysconfig/harddisks" made these errors go > > away... > > > >Has there been a crucial change in the kernel build config since the last > > 3 months or so that I have missed? I have had this working before without > > these problems... I'm curious...