[Orca-dev] RE[2]: RE: Re[4]: Digest 336

[Rusty Carruth]

"Alan LeGrand" <legrand at inow.com> wrote:
> Rusty,
> 
> If this is the only problem  then simply add
> 
> int GLOBAL_diskinfo_size = 512;
> 
> the size does not matter.  Since the disk_map function is suppose to
> terminate before it reaches the end of the array.

That fixed it.  here's the result as non-root:
rcarruth at msfree> /opt/RICHPse/bin/se /tmp/disks.se
RAW_disk_count:3   GLOBAL_diskinfo_size:512
i=0 j=0 dad0 c0t2d0 atapicd2
i=0 j=1 c0t0d0 c0t0d0 dad0
i=1 j=0 c0t2d0 c0t2d0 atapicd2
i=2 j=0 fd0 c0t2d0 atapicd2
i=2 j=1 fd0 c0t0d0 dad0
Fatal: subscript: 2 out of range for: GLOBAL_disk_info

and as root:

# /opt/RICHPse/bin/se /tmp/disks.se
RAW_disk_count:3   GLOBAL_diskinfo_size:512
i=0 j=0 dad0 c0t2d0 atapicd2
i=0 j=1 c0t0d0 c0t0d0 dad0
i=1 j=0 c0t2d0 c0t2d0 atapicd2
i=2 j=0 fd0 c0t2d0 atapicd2
i=2 j=1 fd0 c0t0d0 dad0
Fatal: subscript: 2 out of range for: GLOBAL_disk_info


> 
> Your fix eliminated searching for the long disk name of the short disk name
> fd0 (floppy disk 0) which is not in the table.   disk_map is suppose to stop
> as soon as it encounters a disk slice   c*s?   apparently your eide disk
> dad0  long disk name differs from a scsi disk  long disk name. For example
> the first scsi disk is normally sd0 and has a long disk name  c0t0d0  with
> slices sd0,{a-f} or c0t0d0s{0-7}

Yeah, and what's strange is that the pointers all look like you'd expect
(see below) - apparently something is different enough to confuse matters.
Hopefully your output above (which looks strange to me, when you look at
what devices I have, below) will help you more than it helps me! ;-)
(Of course, if I *looked* at your program the output would maybe make
more sense ;-)

> What I really wanted to get from running the test program was the short to
> long disk mapping for your disk. If you can't get the test program to run
> you should still be able to identify the mapping from iostat by running it
> with and without the -n option (if solaris 2.5.1 supports iostat -n.)
> Unfortunately,I don't remeber when sun added this feature
> 
> Alan
> 
rcarruth at msfree> iostat
      tty     atapicd2          dad0           fd0          cpu
 tin tout Kps tps serv  Kps tps serv  Kps tps serv  us sy wt id
   0  133   0   0   32   18   2   43    0   0  401   4  1  1 94
rcarruth at msfree> iostat -n
Usage: iostat [-tdDxcI] [-l n] [disk ...] [interval [count]]
rcarruth at msfree> 

Looks like they don't support it in 2.5.1 yet.  Bummer.

This might be slightly useful:

rcarruth at msfree> df -kl
Filesystem            kbytes    used   avail capacity  Mounted on
/dev/dsk/c0t0d0s0     480815   14534  418201     4%    /
/dev/dsk/c0t0d0s4     962134  370228  495696    43%    /usr
/proc                      0       0       0     0%    /proc
fd                         0       0       0     0%    /dev/fd
/dev/dsk/c0t0d0s1     769710  100676  592064    15%    /var
/dev/dsk/c0t0d0s5    4878413 3601648  788925    83%    /export/home
/dev/dsk/c0t0d0s3     240399   76204  140165    36%    /opt
swap                  388200  117880  270320    31%    /tmp
rcarruth at msfree> ls -l /dev/dsk/c0t0d0s{0,1,3,4,5} /dev/fd
lrwxrwxrwx   1 root           46 May 23  2000 /dev/dsk/c0t0d0s0 -> ../../devices/pci at 1f,0/pci at 1,1/ide at 3/dad at 0,0:a
lrwxrwxrwx   1 root           46 May 23  2000 /dev/dsk/c0t0d0s1 -> ../../devices/pci at 1f,0/pci at 1,1/ide at 3/dad at 0,0:b
lrwxrwxrwx   1 root           46 May 23  2000 /dev/dsk/c0t0d0s3 -> ../../devices/pci at 1f,0/pci at 1,1/ide at 3/dad at 0,0:d
lrwxrwxrwx   1 root           46 May 23  2000 /dev/dsk/c0t0d0s4 -> ../../devices/pci at 1f,0/pci at 1,1/ide at 3/dad at 0,0:e
lrwxrwxrwx   1 root           46 May 23  2000 /dev/dsk/c0t0d0s5 -> ../../devices/pci at 1f,0/pci at 1,1/ide at 3/dad at 0,0:f
 
/dev/fd:
total 0
crw-rw-rw-   1 root     176,   0 Oct 10 14:55 0
crw-rw-rw-   1 root     176,   1 Oct 10 14:55 1
crw-rw-rw-   1 root     176,  10 Oct 10 14:55 10
crw-rw-rw-   1 root     176,  11 Oct 10 14:55 11
<etc, for 62 total entries>

Hope this helps!

rc