2927aa2c-bcad-4982-b785-6b9ccc007482 Posted April 26, 2012 Report Share Posted April 26, 2012 Hi, We have three nightly backups running in sequence whose source directories are 1.5TB, 620GB, and 280GB respectively in size. Each directory contains lots of files and there's about 12 million files (not my fault!) in total over the three sets. All directories are stored on our local file server and backed up via the backup server onto one of our offsite file servers (with a very good connection). The network traffic looks like this: local file server <--> backup server <--> remote file server As you may expect based on the source size and number of files, the backups take a long time to run and they are now starting to run into the working day which is not ideal. The backup script uses the following options: media verification data compression match source files against the media set don't add duplicate files to the media set use status modification date when matching (Linux) I was wondering where Retrospect is likely to slow down more so I can figure out how best to admin the file server - is it the sheer number of files or the size of the files? Are MD5 hashing over the network or the checking for duplicate files the costly operations? Incidentally, adding more detailed profile information to the logs would be a great feature for Retrospect 9 unless it's already been implemented. Quote Link to comment Share on other sites More sharing options...
2927aa2c-bcad-4982-b785-6b9ccc007482 Posted May 14, 2012 Author Report Share Posted May 14, 2012 Sorry to bump this thread but does anyone have any ideas about this? Quote Link to comment Share on other sites More sharing options...
2927aa2c-bcad-4982-b785-6b9ccc007482 Posted June 8, 2012 Author Report Share Posted June 8, 2012 In case anyone reads this, I did convince one user to move a large amount of files to a weekly backup which brought our daily backups down from about 11 hours to three and a half hours. The files did not consume a large amount of space, relatively, so my guess would be that just reducing the number of files made a big difference in speed. Quote Link to comment Share on other sites More sharing options...
twickland Posted June 8, 2012 Report Share Posted June 8, 2012 If there are a lot of files on the source volume, the time to scan and match that volume can be considerable, even if few or no files need to be backed up. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.