Monday, July 19, 2010

asynch descriptor resize

I was playing with a long query running on SSD-PCI drive, and I saw a new wait event "asynch descriptor resize". Never seen that one before.
A small research on Metalink, and I found note 1081977.1.
"...This event is set when the number of asynch descriptor reserved inside the OS kernel has to be readjusted..."
That point to adjusting "/proc/sys/fs/aio-max-nr".
Running the following reduced the issue:
echo 4194304 > /proc/sys/fs/aio-max-nr


discgolfdba said...


Might you have observed such events when using Oracle 11g R2 on an MS Windows Server 2008 guest OS on VMware ESX v4?

In v$session, the following is occurring:
asynch descriptor resize
outstanding #aio
current aio limit
new aio limit



Younes Naguib said...

No, but doing a quick search, I found this:

That can cause this type of events in Oracle.