Stats pack report for processing during CTAS logging exercise

sp_8131_8133.txt

And a couple of Spotlight captures:



Notice the RED on log writer - the average log write time can go over 1,000 ms (1 sec) - so is the drive is being swamped? With over 2,500 blocks per second (8K block) that is over to be 20GB/sec transferred. That's high for any single drive.



See the orange block grow as the number of active sessions get seen. This is how I know I affect the online users.

Why would the CTAS saturate the REDO logs and prevent the OLTP sessions from processing. Can it not prioritise the small OLTP sessions?



I feel that doing the CTAS exacerbates a current problem - the system is 'out-of-balance' Very fast CPU / Datafile but slow REDO. However, this is all proper SUN kit!