Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
sureshot
Yak Posting Veteran
72 Posts |
Posted - 2008-04-26 : 09:51:44
|
| Due to SQL's proportional fill algorithm I'd like to have the amount of data in my data files as close to equal as possible.File sizes before:dat 482,895 MB 0% free2dat 436,927 MB 3% freeI made the max size of each file 600,000 MB and added a third file 3dat also 600,000 MB. I rebuilt all the clustered indexes (and nonclustered for good measure) and unfortunately the re-balancing wasn't quite right.File sizes after:dat 464,761 MB 77% free2dat 443,234 MB 73% free3dat 309,568 MB 51% freeI only have a handful of heap tables that take up <100MB total so they're not the issue. I did do an ONLINE index rebuild. I'm not sure if an offline rebuild would have been better. I will not be able to try and offline for a few weeks though as it's time consuming and I have other tasks I need to run on this test server now.I did a FULLSCAN rebuild on any column statistics not updated by the index rebuild but that didn't help either. |
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2008-04-26 : 10:25:40
|
| Keep in mind that all system objects are stay in primary file only. |
 |
|
|
sureshot
Yak Posting Veteran
72 Posts |
Posted - 2008-04-26 : 11:58:43
|
| Yeah, but we're talking >100GB difference so it can't be just the system objects. |
 |
|
|
sodeep
Master Smack Fu Yak Hacker
7174 Posts |
Posted - 2008-04-26 : 15:33:28
|
| What? Not sure I understood what you are trying to achieve? |
 |
|
|
|
|
|