I have tried to get this script to do what I would expect from
such a script, but either I am missing something obvious or
the script is lacking this function.
I want it to write dump files with successive revision ranges
only, *not* with overlapping ranges, as it does.
Let's say my repository has HEAD rev 1500 and I call the script
with svn-backup-dumps.py -c 1000 <repos> <dumpdir>
It produces the files
repository.000000-000999.svndmp
repository.001000-001500.svndmp
Then I commit something to the repository (HEAD=1501)
and call the script again. It produces:
repository.001000-001501.svndmp
So the last file includes a lot of redundant information.
If, say revision 1001 contains the checkin of a very large
file, every time the script is called and the HEAD of the
repository has not reached 2000, the whole data of that
file is written to the dump files. As it is, it wastes disk space,
CPU time, backup media space and is more complicate to
restore.
Is it not possible to run the script so that it starts with
the *next* revision not dumped yet? In the example
repository.001500-001501.svndmp
Oliver
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Feb 27 08:41:12 2007