If there is not enough room on the data page for the new row, a page split must be performed.
A new data page is allocated on an extent already in use by the table. If there is no free page available, a new extent is allocated.
The next and previous page pointers on adjacent pages are changed to incorporate the new page in the page chain. This requires reading those pages into memory and locking them.
Approximately half of the rows are moved to the new page, with the new row inserted in order.
The higher levels of the clustered index change to point to the new page.
If the table also has nonclustered indexes, all pointers to the affected data rows must be changed to point to the new page and row locations.
In some cases, page splitting is handled slightly differently.
See “Exceptions to page splitting”.
In Figure 9-3, the page split requires adding a new row to an existing index page, page 1007.
Figure 9-3: Page splitting in an allpages-locked table with a clustered index