Re: Broadcast on commit and log file sync

  • From: Lei Zeng <leizeng2003@xxxxxxxxx>
  • To: "saurabhmanroy@xxxxxxxxx" <saurabhmanroy@xxxxxxxxx>, oracle-l <oracle-l@xxxxxxxxxxxxx>
  • Date: Fri, 24 Feb 2012 15:04:09 -0800 (PST)

 
For RAC environement with BOC SCN, the 'log file sync' time involves two 
pieces:  redo buffer write to log file, and SCN broadcast acknowledge back time 
redo write time -  Total elapsed time of the write from the redo log buffer to 
the current redo log file in 10s of milliseconds
redo size   -  Total amount of redo generated in bytes
 
Lei 
DBspeed  http://www.dbspeed.com/product.html

 

________________________________
From: saurabh manroy <saurabhmanroy@xxxxxxxxx>
To: oracle-l <oracle-l@xxxxxxxxxxxxx> 
Sent: Friday, February 24, 2012 1:56 AM
Subject: Broadcast on commit and log file sync

Hello All,
I am seeing the following message in lgwr tracefile along with 'log write
time wait time' warnings:

Warning: log write broadcast wait time 3000ms

This is a RAC database (Version 11.1.0.7.9) We are using default BOC SCN scheme.

My question is: In RAC, Can this wait time exacerbate problems of log
file sync waits ? Any known conditions ?

There are few bugs (e.g. 7716356) on MOS pointing to increase in log
file sync waits due to BOC scheme, but i am unable to draw a clear
picture in my mind about the whole mechanism of BOC and how can it
impact log file sync.


Any comments are welcome.


Regards,

Saurabh Manroy


--
//www.freelists.org/webpage/oracle-l





 
A description of the life cycle of 'log file sync' wait :
1. A user sessions issues a commit or rollback and starts waiting on log file 
sync.
2. LGWR gather redo information to be written to redo log files, issues IO and 
queues BOC to an LMS process and posts LMS process.
3. LGWR waits for redo buffers to be flushed to disk and SCN to be ACK'd
4. Completion of IO and receiving ACK from LMS signal write complete. LGWR then 
posts foreground process to continue.
5. Foreground process wakes up and log file sync wait ends.
 
You can check the AWR or v$SYSSTAT for the following stat names.
DBspeed is more easy to examin how their values change over a time frame.
http://www.dbspeed.com/index.html
'AWR instance & service statistics' section, 'AWR instance activity statistics 
by stat name' report
 
redo write broadcast ack count  -  Number of times a commit broadcast 
acknowledgment has not been received by the time when the corresponding log 
write is completed. 
redo write broadcast ack time -  Total amount of the latency associated with 
broadcast on commit beyond the latency of the log write (in microseconds). 

--
//www.freelists.org/webpage/oracle-l


Other related posts: