Performance issue (may be background checkpoint related)

 Hi,

We have a 2-node RAC, version 11202, on Solaris x86. Apps are connected to one of the nodes only (so no cache fusion related issues). 

Everyday, early morning, at the same time, performance drops for about 30-40 seconds and it is the IO that gets choked at that time. log file sync shoots up to 100ms+. And after moving the redo logs to another diskgroup a few days back, now its db file sequential read that shoots to 100ms+. Both of  which is otherwise in single digits, typically, less than 5ms.

We have confirmed that there are no end-of-day jobs or backups running at that time. 5minute and 1minute statspack reports confirm the same i.e. the reports do not show any unusual SQL with heavy IO.