> On Saturday 02 September 2006 00:20, Jesper Krogh wrote: > The cost of doing a Verify -- Catalog is really quite minimal, so in my > opinion, I don't see the benefit in complicatint Backup jobs any more than > they are.
I don't know if they are complicated already.. This feature is far to small to allow to make the codebase be complicated for that. But the cost of a verify is a read-through of the disk. There is an administrative overhead of doing it via verify (since I basically what the check on "all files that has been unchanged according to metadata since last backup". This is pretty hard to describe as a FileSet. > I wouldn't even know what syntax to use to turn it on and > control it. The "assertion" that: "Data may not be altered, without the mtime has been changed" should always be true. Thus making this check and alerting if is fails does not seem to harm anyone. So .. just dont make it configurable seems ok to me. If we dont have the data (a new file) then we cant assume anything, but otherwise.. if the above assertion fails, then we have a disk that is beginning to corrupt data. As another suggested, this can be done via post-analysis of the catalog. I'll try to write some scripts to do that, my first try: select PathId,FilenameId,LStat, count(MD5) as md5s from File group by PathId, FilenameId,LStat where md5s > 1; [1] Is invalid in MySQL :-/ and without the "where-clause" is has currently been running for 49 hours. I'll work a bit on optimizing this. Jesper [1] This is even wrong since it doesn't take the client-hostname into consideration. -- Jesper Krogh ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
