the wonders of SCSI
Gregory G. Woodbury
ggw at wolves.uucp
Thu May 31 09:26:51 AEST 1990
gsteckel at diag2.East.Sun.COM (Geoff Steckel - Sun BOS Software) writes:
>>ggw at wolves.UUCP (Gregory G. Woodbury) writes:
>>> mount: /dev/dsk/c1d0s0 no such device
>What partitions do you have set up? What does /etc/partitions say about your
>drive? What does "/etc/mkpart -t vp" say?
That was the weirdness that was throwing me. mkpart kept giving
strange answers. Then I tried it again (when fully awake!) and it
>If you have no entries in your /etc/partitions for the new drive, try
>copying the entries for an existing disk, decode them with the /etc/mkpart
>documentation, edit them to match your real disk, and make a vtoc with
>/etc/mkpart. Also label the volume (and I've forgotten the name of the
final effective sequence:
1. /etc/diskconfig on /dev/dsk/c1d0p0 to get basic geometry
(be careful - some of the /etc/disk* commands will trash
/etc/partitions - make a backup!)
2. run fdisk /dev/dsk/c1d0p0 and make unix partitions
3. put the basic disk "stanza" in /etc/partitions
4. run mkpart -i diskNN to create initial vtoc
5. backup /etc/partitions and run /etc/disksetup
this will fail! but partition stanzas will be placed in
/etc/partitions for your selected configuration.
6. check /etc/partitions to make sure it isn't corrupted (you
did make that backup didn't you?) and add the new
stanzas for the device to the /etc/partitions file.
Probably need to change names from *x to *NN.
7. mkpart -P partition-name diskNN
this is for each of the partitions you configured.
8. now that the partitions are on the drive (verify with mkpart
-t pv diskNN) note the partition numbers.
9. mkfsys /dev/rdsk/c1d0s<part#> ....
this is where the real file system is made.
If you are doing this on a drive that needs remapping,
do a mkpart -v diskNN to perform a surface analysis and
populate the alternates table.
10. labelit /dev/dsk/c1d#s# fsname volname
optional, but highly desireable step, keeps mount from
complaining about mouting <> on /whatever. The fsname
should be the last component of the path that the
section will be mounted on.
11. mount /dev/dsk/c1d#s# /path/name
12. cd /path/name and hand build a lost+found directory!
try to make the lost+found directory inode 3, then cd to
the new lost+found and create a WHOLE LOT of empty
files. This is to make the lost+found directory itself
of a large enough size to be usefull in the case of a
crash. rm all the files just made ;-)
13. edit /etc/fstab to add the new device and its mount point.
This is quite similar to the perigrinations that most normal
unixes go thru in making a new drive, but the specifics of
making partitions and slices and whatever are a bit different
from the ole pdp-11 and vax stuff I used to do, and the Opus
stuff I also do.
This is for ISC 2.0.2, it is also similar to what a friend went
thru for adding a big scsi to ESIX. (Of course, he get back in
town AFTER I get this mess worked out ;-)
Note, that when mounting a 300MB scsi partition on ISC, it will
take a few minutes for the system to read the free list into
memory! It also chews up some of the available memory for the
>I don't thing you want to mount /dev/dsk/c1d0s0...
> slice 's0' is not (according
>to TFM) normally a mountable slice. If you have only mountable one slice,
>/dev/dsk/c1d0s1 is the likeliest name. You should have done mkfs on
>/dev/rds/c1d0s[134...], depending on which partitions are defined in
>If you have defined more than one slice using the default utilities, you will
>probably have `c1d0s1', `c1d0s3', and `c1d0s4' as mountable partitions.
As your note pointed out (in a deleted section) the "default"
utilities are very buggy. Using them, I managed to wipe out
/etc/partitions and almost lost my root drive vtoc! Thanks to
Larry Jones for the info on mkpart that saved me.
Gregory G. Woodbury @ The Wolves Den UNIX, Durham NC
UUCP: ...dukcds!wolves!ggw ...mcnc!wolves!ggw [use the maps!]
Domain: ggw at cds.duke.edu ggw%wolves at mcnc.mcnc.org
[The line eater is a boojum snark! ] <standard disclaimers apply>
More information about the Comp.unix.i386