Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> RE: RE: RE: controlfile schema global enqueue lock

RE: RE: RE: controlfile schema global enqueue lock

From: Joan Hsieh <>
Date: Thu, 01 Feb 2001 09:42:58 -0800
Message-ID: <>

Thanks, so what you suggest? That's why the other day I asked the list about the asyn i/o on os. If asyn is enabled, is it ok to set dbwr_io_slaves? I am not sure for that part. For checkpoint, You see, before I made the change, we have 8 log switch per hr plus every 30 min and every 3200 block (raw disk is 512b /block) do the checkpoint. I think that's too much. I forgot to set log_checkpoint_to_alert to true. Right now, after the change. the checkpoint just occur at log switch time. I am wondering maybe we still hold the lock too long to finished the checkpoint. Should I increased the redo log size to 500m? What's the impact if we have very large size redo log. I remember Steve said 500m is a good number.(please correct me if I am wrong) The principle dba just won't buy it. he think it might cause other problems.

Thanks again


-----Original Message-----
Sent: Thursday, February 01, 2001 11:44 AM To: Joan Hsieh; Multiple recipients of list ORACLE-L


    I doubt this will be the "silver bullet", but it did reduce this particular
problem on our Operational Data Store where testers are constantly writing data.
 Set DBWR_IO_SLAVES = number_of_drives_with_datafiles/2. This ups the number of
processes that are busy writing data to the datafiles thereby reducing the amount of time it takes to complete a checkpoint & thereby releasing the lock
sooner. You may also want to set LGWR_IO_SLAVES in a similar manner.

    Your basic problem is that there is one and only one enqueue lock for the
control files. Let's face it we really don't want two processes trying to change data therein at the same time. What a mess that would be. And when a
checkpoint does occur we need to update the datafiles and control files as a set. This is where the enqueue lock comes into play. If you can speed up the
time it takes to flush out all of the dirty blocks from the sga & log_buffers to
the redo log files then the checkpoint gets completed faster. Right now it looks like you've got a checkpoint happening only at log switches since log_checkpoint_interval is set larger than your logfile size.

Dick Goulet


Please see the official ORACLE-L FAQ:

Author: Joan Hsieh

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
To REMOVE yourself from this mailing list, send an E-Mail message to: (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). Received on Thu Feb 01 2001 - 11:42:58 CST

Original text of this message