Hi Gary, On Dienstag 18 Dezember 2007 21:43:01 wrote: > I am sure SOMEONE here has come across this scenario. I have a > top-level directory that is ~21 TB. In it are 391 other directories > making up that 21 TB. Unfortunately, one of them alone is 15 TB. My > goal is to split the directories somewhat evenly four ways so that > each tape drive is backing up the same amount of space. > > Has anyone written or found a script that can take a huge directory > full of subdirectories and break it apart evenly?
I haven't a script only few ideas to write one ... 1. imho is the only way to know how big is a subdir 'du -sh /dir/subdir' or a similiar function in perl or whatever. 2. the time to compute this information is imho not evenly to solve. 3. the performance from the backup machine in backup time frame is needed to backup not to compute any additional infos. 4. imho is the dir-size not soooo dymanically that you real need to compute this evenly. If 1..4 true you can setup a 'du -sk /dir/* >sizes.txt' job with a very friendly nice level and then you can generate a couple (drivecount) of fileset definitions wich you can include in your dir.conf. The biggest Problem is that you can then only make full backups after changing the fileset definitions. In other words every time if you change the fileset for a job, the job will be upgraded to a full backup ... I think this is not the right way to do this, only if you make the compute/fileset-change before you plan a regular full backup is it a possibility. The only way to resolve this issue is imho a set of jobs (one per subdir or little group of subdirs) and start all of them with 'maximum concurrent jobs = drivecount' into one pool or one pool per drive/job. With this strategy makes bacula that balancing for you without any computing outside from bacula. The biggest one should be in the front of the job definitions, the others raw ordered by size, he is the first one should be started. i hope you understand my bad english, if not please feel free to ask. Regards Falk ------------------------------------------------------------------------- SF.Net email is sponsored by: Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users