Asynch Descriptor Resize explanation

Question

What do asynch descriptor resize wait events mean on an Oracle 11g database running on Windows 2008 R2? According to My Oracle Support, this event applies to any platform, but the description provided includes the following information in Doc ID 1081977.1:

This event is set when the number of asynch descriptor reserved inside the OS kernel has to be readjusted. It is signaled when the number of asynch I/O's submitted by a process has to be increased. The Linux kernel does not allow the limit to be increased when there are outstanding I/O's inside the kernel. Hence, all outstanding I/O's are reaped before the limit is increased. The wait to reap all the outstanding I/O's when the limit is increased uses this event. _{[Emphasis mine]}

This isn't causing a significant problem on our systems, but it would be nice to know what it means on a Windows platform.

This seems similar to a comment on this blog post, particularly because it mentions Oracle 11g R2 with Windows 2008 and VMware. While this indicates that the problem isn't unique to our systems, it does not help identify the cause or possible solutions. The other comment about a Microsoft fix applies to Windows 2008 R1 rather than R2.

Answer 1

Just so there is some sort of answer to this, here is some information I have found.

On MOS there is a document called "High Numbers of 'Asynch Descriptor Resize' Wait Events Seen" (Doc ID 1273748.1) This seems to indicate that the problem is really another wait event slowing the system down and in turn causing these, so they aren't really the source wait event.

Oracle-L has a thread on the wait event, but no mention of it on Windows.

OakTable has some interesting info on the problem, and says,

...the resize apparently sucks on Linux. Perhaps that’s why other ports also suffer and have seen the same wait event.