UNIX for Advanced & Expert Users

Hi all,
Is there a command that allow the user to know that he reach the end of a backup tape (last portion used)?
Thanks and Regards

Are you asking about when the tape is being written to? or when you are seeking on the tape to find the end?

I don't think you can do that when it is writing, but you can do it while querying a tape.

_____________________________________
There is a command "mt" that manipulates the tape. It has many functions. You can set marks on the tape and many other things as well.

From the man page for mt.

mt accepts the following commands:

eof ...... Write count EOF marks.
fsf ...... Forward space count files.
fsr ...... Forward space count records.
bsf ...... Backward space count files.
bsr ...... Backward space count records.
rew ...... Rewind tape.
offl ...... Rewind tape and go offline.
eod ...... Seek to end of data (DDS and QIC drives only).
smk ...... Write count setmarks (DDS drives only).
fss ...... Forward space count setmarks (DDS drives only).
bss ...... Backward space count setmarks (DDS drives only).
status ...... Print status information about the tape drive.
res ...... Reserve tape drive for sole use by the host issuing the mt command (stape driver only).

rel ...... Release tape drive from sole use by the host issuing
the mt command (stape driver only).

How can setmarks be utilized in a backup, can you backup a group of files and create a "setmark" and use that setmark "count" in the restore command.

Basically, if I utilize the "smk" command, does it give me the option to bypass counting individual file marks, and group together multiple files.

Thanks for any help in this regards.

The question regards our current procedure for backing up a Progress Database (Version 9.1c); with the implementation of storage areas, the fsf command will no longer search to the right area of the tape to restore the database.

We run a script that backs up multiple databases, and during the progress restore we run the

mt [device] rew
mt [device] fsf 0
prorest [database] [device]
mt [device] rew
mt [device] fsf 1
prorest [database] [device]

--- every database after the first will not restore.


Prior to the implementation of storage areas this worked, now it doesn't. Does anybody have any ideas?

Should we add an EOF or SMK command to return the record or setmark counts. And if we do, how do we use these in the restores.


Thanks in advance.

comatt1, what os are you using? What device file? What do you mean by "implementation of storage areas"? It sound to me like you are just using a rewind device. But if you can't get filemarks to work, you shouldn't be increasing the complexity of the problem by adding in setmarks.

You should switch to -f, but -t still works, so that it not the problem.

A device file is something like /dev/rmt/0m or /dev/rmt/0mn. The "n" at the end means no-rewind.

On hp-ux, you can use lssf to decode the minor number. Run "lssf /dev/rmt/0mn" and make sure that it says "no rewind". Make sure that you are using /dev/rmt/0mn when you run prorest.

Bear in mind that the restore script is not the only player here. The backup script must be correct as well.

If this all looks good, you will need to talk with the Progress folks.