Thanks John :) Mainly to you remember about my problem and suggest me something ..
I'll try it, and post results here in sequence... Can I use "time" to mensure copy time ? # time dd if=/dev/sda of=/dev/nst0 bs=1M count=128 -- Flávio do Carmo Júnior On Nov 23, 2007 4:27 PM, John Drescher <[EMAIL PROTECTED]> wrote: > > Basically, there is one way to do this: Use migration. You do your > > backups to disk based volumes, and later migrate to tape. > > > > You'll probably need more disk space, though, as that pool has to hold > > at least one complete backup. No automatic spooling/despooling sequence... > > > Maybe he could fake this by getting more disk space and making the > spool size several hundred GB. Possibly this way he can have spooling > all night and then he will be able to swap in a few tapes during the > day. However I admit this would be difficult to optimize. > > I would still work very hard to try to figure out what is causing the > tape to write at 1/4 to 1/2 of its native speed. > > One thing you can to to track that down is to use get a new blank tape > and use dd. Something like > > dd if=/dev/sda of=/dev/nst0 bs=1M count=128 > > And see the speed you get here. I used /dev/sda here as the source > instead of /dev/zero because the hardware compression will make that > next to nothing and you will get a false result. Also /dev/random will > not work because I find the random numbers will not be generated fast > enough to keep up with the drive. > > John > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
