On Wed, 2003-11-12 at 14:10, Eric Sammons wrote: > Using the following, which was provided to me, > > # echo "add device range=0206" > /proc/dasd/devices > > I took the liberty of modifing the command to have >> so that I don't blow > away my devices file.
Actually, it won't. It's not really a file, it's an ioctl interface in the /proc pseudo-filesystem. So > is safe in this case. >> works too, though, I think. > I then execute the following dasdfmt command: > > dasdfmt -y -b 4096 -n 0206 > dasdfmt: Unable to open device /dev/dasd/0206/device: No such file or > directory .... > 0206(ECKD) at ( 94: 28) is dasdh : active at blocksize: 4096, 594000 > blocks, 2320 MB > > Anyidea what is happening here? I have noticed that the dasdfmt command > isn't cooperating in SLES8. Where if I execute format at the VM layer all > is well. Yeah, I've seen this before too. Let me guess: this is real, honest-to-goodness virgin disk, that nothing has ever touched before, right? For some reason, Linux won't touch it unless it finds some magic goodie on it that CMS formatting it gives it. I don't know if low-level CP formatting, or just Label formatting, would do the trick as well. I'd say it's a bug in the DASD driver or the dasdfmt utility, but I don't really know; the workaround, as you've found, is to touch all your disks with CMS first. Adam
