[rescue] 10GB emulated drive and disk layout.
Mouse
mouse at Rodents-Montreal.ORG
Sat Jan 24 12:50:16 UTC 2026
> I've created a 10GB emulated drive with a ZuluSCSI fitted in a
> SPARCstation 20 which has the latest firmware. I'm attempting to
> partition this and install SunOS 4.1.4. The root filesystem is <1GB
> and all are <2GB.
I don't recall enough details about 4.1.4; it's been a very long time
since I tried to do anything with SunOS.
This sounds a bit like the 6-byte CDB issue, but I doubt it's that
simple, since 6-byte CDBs limit access to slightly over the first 1G of
the disk (they have only 21 bits of address and 8 bits of length, so
they're unable to access anything beyond sector 2097406, and sectors
past 2097151 can be accessed only by doing >1-sector I/O). That seems
to me like a crippling enough limitation that I'm inclined to doubt
4.1.4 would have it.
> This appears to go well until the installer runs newfs on the h
> partition and at which point I get errors [and the label is
> corrupted].
What has it done before then? In particular, has it written any of the
other partitions' filesystems? If so, are they usable?
I'm also wondering if the problem is actually with the partition it
wrote just before h; it could be that that one corrupted the label and
nothing noticed until it went to newfs h. What order does it write
things in?
I massaged your layout text into readbility and had a look. I would
expect it to have written /usr and /var, at least, before doing /home,
but maybe not; as I said, it's been a long time.
> The drive is defined as 19377c, 16h and 63s (10000392192 bytes).
Combined with the layout you gave, I note that h is the only partition
beginning after the 8G point, which is the limit of 24-bit sector
numbers. If it writes the partitions in order, writing the end of g
could have wrapped (eg, at the 8G point) and scribbled on the beginning
of the disk, possibly including overwriting the partition table.
I would not expect 4.1.4 to have a 24-bit limit on sector numbers,
especially not one that fails silently when reached, but I've been
surprised often enough before.
If that's not it, I'm out of suggestions.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
More information about the rescue
mailing list