RAC - 'gc cr failure' and 'cr request retry'

From: "McLeod, Riley" <riley.mcleod@xxxxxxxxxx>
To: <oracle-l@xxxxxxxxxxxxx>
Date: Fri, 17 Apr 2009 16:45:51 -0400

Hi List,

Environment:

Oracle RAC 10.2.0.4
Redhat Linux x86_64 2.6.18-92.1.6.el5

I'm trying to troubleshoot an issue one of our customers is
experiencing.  I got a trace file from them and there are a couple of
RAC-related wait events in the trace that I can't find much information
about.  Here's a snippet of the trace file:

WAIT #49: nam='gc current block 2-way' ela= 319 p1=14 p2=898365 p3=1
obj#=90107 tim=1210676435708838
WAIT #49: nam='gc cr block 2-way' ela= 317 p1=15 p2=464260 p3=1
obj#=90114 tim=1210676435716631
WAIT #49: nam='gc current block 2-way' ela= 298 p1=18 p2=989016 p3=1
obj#=90113 tim=1210676435717061
 .
<~500 more 'gc current block 2-way' waits>
 . 
WAIT #49: nam='gc cr failure' ela= 208 p1=12 p2=823093 p3=1 obj#=90107
tim=1210676436002591
WAIT #49: nam='cr request retry' ela= 976838 file#=12 block#=823093 p3=0
obj#=90107 tim=1210676436979481
WAIT #49: nam='gc cr block 2-way' ela= 597 p1=12 p2=823093 p3=1
obj#=90107 tim=1210676436980260
WAIT #49: nam='gc current block 2-way' ela= 396 p1=14 p2=928948 p3=1
obj#=90107 tim=1210676436980778
 .

What are the 'gc cr failure' and 'cr request retry' waits?  These two
events happen twice more before the fetch occurs.  Does the 'gc cr
failure' indicate a problem with their interconnect?  From what I see in
the trace file these two events seem to happen only on queries that
involve a high number of 'gc cr block 2-way' waits.  

Any explanation of these events is greatly appreciated.

Thanks,
Riley


Riley McLeod
Morris Technology Services
706-828-2944
--
//www.freelists.org/webpage/oracle-l

RAC - 'gc cr failure' and 'cr request retry'

Other related posts: