Summary : ufsdump problem

From: Aggeliki Karabassi - SYSTEM ADMINISTRATOR (
Date: Wed Nov 05 1997 - 05:40:43 CST

Original Question :
> I' ve already posted a similar question sometime ago, but although I have
> received some answers , the problem remains.
> I have a Sun DDS2 and I am trying to backup a SunOS 4.1.3 machine,
> while the DDS2 is attached to a Solaris 2.5.1 machine. I have some serious
> problems determining the settings required to perform the dump. I had a
> pretty old DumpTool version which couldn't identify the DDS2 automaticaly
> and I hoped a more recent version would help. I must thank Phil Kao from
> Artecon Inc. for sending me the new version, but I am afraid it still does
> not support DDS2. Some people send me the settings they use , but all ends
> to the following situation :
> ufsrestore ifv /dev/rmt/0lbn
> produces :
> Record size (512) is not a multiple of dump block size (1024)
> I thought specifying block size 1024 would help , but
> ufsrestore ifvb /dev/rmt/0lbn 1024
> produces :
> partial block read: 512 should be 524288
> The last error is not described in man pages and I really don't know what
> it means. The weird thing is ( at least it sounds strange to me ) that the
> second number ( 524288 ) is always the same, even if I change the
> settings for tape density, tape size or even the total size of data I am
> trying to dump.
> I am sure I must be doing something pretty stupid, but I really don't know
> what... so if anybody can help I would appreciate it.
> Thanks in advance

Since nobody could find again any obvious error in the restore process,
I have examined again the dump script and I found the most weird thing :
Before dumping the filesystems, DumpTool tars the script itself and stores
it to the tape using dd. This works fine on Exabyte and DAT streamers we
are using, but it didn't work with our DDS2. Even when trying to read the
tape with dd we would get an I/O error. I really don't know why this
happens, but after removing this line from the script everything works
In any case I 'd like to thank everybody who replied and especially
Casper Dik who reminded me to try using dd to read the tape, and
helped me realize that there was something seriously wrong with the data
stored .

Aggeliki Karabassi
Department of Informatics tel: 7217941 - 121
University of Athens

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:08 CDT