Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

It's actually 2 question in one. First is a bit more theoretically. So when specifying accurate options how does bacula figure out if a file needs to be backed up ? it's a simple AND ?

As in if the options are Accurate = sm5 bacula will not backup the file if
((size = old size) AND (modtime = old modtime) AND (md5 = old md5))
Is that correct ? Do any of the options take precedence ? as in would be a file skipped if modif time is diffreent but it has the same md5sum ? Are there any implied options that you cannot ignore ?

Practical case, ( bacula 5.0.1 ) I have to back-up a svn repo, in order to be able to make incremental backups as simple as posible i am hotcopying (client run before) it to another location, that bacula will backup ( then delete it with client run after). Now in the fileset i have

Accurate = spnd5

This should tell bacula to take into consideration size , permission bits number of links , decreases in size and md5sum. However , an incremental is also including a full copy of the svn. What am i doing wrong ? it seems that it takes into account creation time even tho i have not specified it.

share|improve this question

2 Answers 2

It shouldn't check for this at all given your accurate options, but maybe you can test if setting the option mtimeonly=yes will work, because the ctime will have changed if you add a link to a file.

share|improve this answer
    
I dont not want to check for modification time at all since the files are recreated each time a job starts and then deleted. So modification time aswell as creation time should be ignored. That is since md5sum combined with the size of the file should be enough indication if the file is the same or not. –  Kiss Stefan Aug 11 '10 at 10:13
    
Yeah, but I am not sure if you can really prevent bacula from using the mtime. Also, comparing size and mtime is a cheaper operation than reading the whole file every time and calculate the MD5. –  Sven Aug 11 '10 at 10:21
    
I am aware of the performance impact, but since the files are the same but generated a diffrent time there is no other way to distinguish between them. Other options would be to use svnadmin to make incremental backups, but that really is a way of adding unnecesary complexion - IMO the more simple a system is the more reliable it is. –  Kiss Stefan Aug 11 '10 at 10:49

Here is the main reference PDF for Bacula 5.03. Page 29. It is possible that bacula is using the default options for verify as well as the specified ones for Accurate.

Try changing the options for verify as well as accurate.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.