The Challenges of Space

  • Comments posted to this topic are about the item The Challenges of Space

  • This is a major headache indeed. But are there any solutions for the problems you bring up? Table backup & restore would be great to have. Categorizing tables into priority groups, and choose what needs to be backed up every day, and what once a week, or what need to be re-indexed every day and what not.

    Another approach is to save database data into separate databases, e.g. logs, but this creates a new problem with ending up with very many databases, when you get over 50 or so, the management software doesn't do a great job anymore, e.g. no filter options.

  • In SQL Server, I believe a Differential backup solves some of this problem.


  • Differential backup doesn't give that power to choose what to back up. You can't choose what to restore either. It does speed up your daily backup, which can be useful.

    However, let's say your database is 300 GB, and something went wrong and in a specific 100.000 record table something goes wrong, and you want to restore that from a backup, you will have to restore the entire 300 GB to even take a look at previous data. That is a major pain.

  • Specialized backups certainly have issues. They're really exports of data, and the importer is responsible for ensuring integrity, but having copies of lookup tables, or other specific (and limited data), can help with the "Whoops" issue.

  • The challenge I have had with differential backups with some OLTP systems I have worked on is that the rate of data change is higher than the overall size of the database looked over a period of time... meaning that if I setup differential backups with a full backup once a week (or whatever interval) the differential quickly becomes longer to take... never mind restore than a full backup.

    What frustrates me is that computer systems are using letters instead of numbers for the drives at the logical level.. thereby compelling the use of mount points... which Steve mentioned are not always fully supported. (Believe me.... *sigh*)

    In addition, clusters still wont let me use the A or B drive letters. (From floppy days...) I wish for Server level OS at least, Microsoft would start to use numbers for drives instead. (drive 1 instead of C etc.)

    It would also be easier if from a cluster perspective there was a built in safe online method to extend existing luns from a san.

    ok, so ends my rant more about drive limitations than space challenges 😛

Viewing 6 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic. Login to reply