hubler@tomcat.lerc.nasa.gov (Dale Hubler) (01/02/91)
Does anybody have an answer for an rdump question I have? I tried to use rdump to backup a user's /u filesystem to a QIC 150 tape drive on our server (520). If I take the drive to the workstation I wish to backup, I can easily backup the 75 Meg onto one DC600 cartridge (in QIC 120 format). The same 75 Meg requires 4 cartridges when I use rdump to backup across the network. The documentation refers to the -b option on the rdump command but does not give a default value. I am backing up to /dev/rmt0.6. Does anybody know what blocking factor will allow me to write in a QIC 120 format using rdump? Thanks in advance for any help! Dale Hubler -- Dale A. Hubler *** hubler@galaxy.lerc.nasa.gov *** (216) 977-7014
kaiser@ananke.stgt.sub.org (Andreas Kaiser) (01/06/91)
DH> I can easily backup the 75 Meg onto one DC600 cartridge (in QIC 120
DH> format). The same 75 Meg requires 4 cartridges when I use rdump to
DH> backup across the network.
If you backup on a local tape, the tape will most likely be operated in the
"streaming mode", where there is little or no stop-and-go. When operated via
network, the tape data rate might be about as high as the network data rate
(measured with ftp on an idle 4Mbit token ring). So the tape operates in
stop-and-go mode, significantly reducing its capacity. In the manual of the
tape, there is a picture saying the capacity is reduced from 150 MB to 30 MB
when operated in non-streaming mode. I personally have trouble believing this
relation, especially at large block sizes, but its given in IBM's docs...
Andreas
hubler@galaxy.lerc.nasa.gov (01/08/91)
> DH> I can easily backup the 75 Meg onto one DC600 cartridge (in QIC 120 > DH> format). The same 75 Meg requires 4 cartridges when I use rdump to > DH> backup across the network. > > the tape operates in >stop-and-go mode, significantly reducing its capacity. In the manual of the >tape, there is a picture saying the capacity is reduced from 150 MB to 30 MB >when operated in non-streaming mode. I personally have trouble believing this >relation, especially at large block sizes, but its given in IBM's docs... > Thanks for the reply, I guess I'll just have to carry the tape drive from station to station if I don't want to drastically increase my tape costs. I wonder if I could convince IBM to put the SCSI port on the front of our boxes? -- *********************************************************************** Dale A. Hubler -- Sverdrup Technology -- (216) 977-7014 hubler@galaxy.lerc.nasa.gov **** Whip me, Beat me, Call me Edna *******