I jsut upgraded all my servers to redhats updated kernel, 2.0.35, from
their errata lists. One server, with dual scsi, is now screwing up ,adn
not booting properly.
it has a DPT2144UW, and a Adaptec 2940UW bios 1.23
so far, using initrd to probe aic7xx, and eata_dma.
both modules load, adaptec finds all devices (most of the time), sometimes
it misses one 9GB drive, probably due to slow spin-up.
the eata_dma,, reports. that the card isn't there at its Address that the
bios had assigned it (i tried several addresses with the same results)
it finds all the drives on its channels, and then:
it goes thru the device listing for the drives with the following info:
SCSI device sdc: hdwr sector= 512 bytes, Sectors = 8388314 [4095 MB] [4.1 GB]
it dies this for sdc, and sdd only, but there are SIX DRIVES on this
controller. it used to do it for sdc,sdd,sde,sdf,sdg and sdh
so my system only seems to find FOUR scsi HARD drives. needless to say
this is causing a major problem..
the ID layout for the controller is:
aic7xxx, id0 = IBM OEM 9 gb boot driver (sda)
id8 = Seagate 9.1 gb (sdb)
eata_dma id0 = cdrom
id1 = Seagate 4.1 gb (sdc)
id2 = Seagate 4.1 gb (sdd)
id3 = Seagate 4.1 gb (sde) <-- Can't access
id4 = Seagate 4.1 gb (sdf) <-- Can't access
id5 = Seagate 4.1 gb (sdg) <-- Can't access
id6 = Seagate 4.1 gb (sdh) <-- Can't access
ANY ideas welcomed.
I am also trying a monolitic kernel as i type (its still compiling)
Dave
