Re: Wrong Hanoi backup scheme

> I have completely moved the calculation to the shell calculator "bc" =

> now.   In case you want to use my patch, please use the newer one that I =

> have attached to this email.

Thanks for that. I'll consider it in the next release.

> I have also noticed another, more general problem in the Tower of Hanoi =

> backup scheme: the first ever backup (A on the first cycle) is deleted =

> after only two backup runs.  This is not so good, but can be easily =

> overcome by shifting the last backup of a cycle (E for five backups) to =

> the front.

What you say here is well known and the reason why SaraB comes with a GFS s=
chedule by default.
It is also documented in the FAQs. Shifting the cycle will only let you go =
for the first round and the problem actually persists.

> The computer that is being backed =

> is only online at random intervals and sometimes switched off for days, =

> so I'll have to work with pinging and backup timestamps. =


Have you considered using anacron to trigger backups instead of cron? It is=
 designed to exactly do what you need here.

> Moreover, my =

> backup server is an embedded system and runs with a write-protected /etc =

> directory, which would need quite a few changes in the sarab script.

SaraB just installs it's configuration to /etc by default. You can leave th=
at directory anywhere on your system. Then call sarab with the full path to=
 sarab.config on the command line:

sarab.sh /home/martin/.sarab/sarab.conf

Other paths in that file may be relative or absolute and are taken relative=
 to the current directory the script is running in. So when using relative =
paths, you may want to first cd to a specific directory.
 =

> Finally, I want to have a relatively complicated grandfather-father-son =

> backup scheme, where the grandfathers are cycled using a Towers of Hanoi =

> scheme.  This could be done with sarab, but then it's almost faster to =

> write my own Python script -- especially as the complicated work is done =

> by DAR, anyway.

In fact the generations in GFS and TOH look very similar, not to say they a=
re the same.
I doubt you can recreate SaraB faster than just configuring the correct rot=
ation.schedule.

On the other hand, if you like doing it, there is nothing to stop you. ;-)

Have fun!

Hiran
_________________________________________
NEU: Mit WEB.DE DSL =FCber 1000,- =BF sparen!
http://produkte.web.de/go/02/


---------------------------------------------------------------------------=
---
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev