kewr prefix is for Kernel Workload Repository as far as I understand.
The trace file is most probably from m000 process (MMON_SLAVE).
What are GRFR? maybe something like Global Request to Flash Remotely.
MMON collects information all the time including frequent samplings of
session waits (ASH data). It has cyclical memory buffer. This buffer
is flashed to disk from time to time including times it gets full.
Perhaps, there is also global trigger to flush buffers and the request
has timeout. Maybe during the AWR snapshot it requests all instances
to flush its in-memory AWR data? Can't say for sure but it does make
sense. Can you check timestamp of your AWR snapshots and see it it
matches the errors? Of course, only if you are licensed otherwise you
can't select from those views without license violation even if AWR
causes you troubles.
So it seems timeout occurred waiting for that. Why? Another node busy
or whatever else the reason for not getting it done on time. Do you
experience any issues due to that? Don't know if it affects anything
besides AWR data. I would open an SR if that bothers you.
On 8/2/07, William Wagman wrote:
I am seeing this a lot in trace files in the bdump directory. This is a
two node RAC cluster running 1ogR1 Standard Edition on REHL3. I can't
find anything with google.
*** 2007-07-31 12:15:40.125
*** ACTION NAME:(Auto-Flush Slave Action) 2007-07-31 12:15:40.125
*** MODULE NAME:(MMON_SLAVE) 2007-07-31 12:15:40.125
*** SERVICE NAME:(SYS$BACKGROUND) 2007-07-31 12:15:40.125
*** SESSION ID:(136.58862) 2007-07-31 12:15:40.125
KEWRGRFR: failed waiting for remote flushers, rcode=-125
Univ. of California at Davis
IET Campus Data Center
Alex Gorbachev, Oracle DBA Brewer, The Pythian Grouphttp://www.pythian.com/blogs/author/alex http://www.oracloid.com
BAAG party - www.BattleAgainstAnyGuess.com