Re: CHECKPOINT NOT COMPLETE
From: Mark D Powell <Mark.Powell_at_eds.com>
Date: Mon, 21 Sep 2009 07:23:45 -0700 (PDT)
Message-ID: <b325bfe3-9868-477e-b043-e7c4bfb320ec_at_33g2000vbe.googlegroups.com>
On Sep 21, 9:50 am, ddf <orat..._at_msn.com> wrote:
> On Sep 21, 8:10 am, Michael <mbulch..._at_gmail.com> wrote:
>
> > We have 5 redo logs each of 150M in size and there are 2 members in
> > each group. We have been constantly getting the following error
> > message:
> > Checkpoint not complete
> > Current log# 2 seq# 127927 mem# 0: /u01/NI00/redo_tbs/redo_2a.log
> > Current log# 2 seq# 127927 mem# 1: /u03/NI00/redo_tbs/redo_2b.log
>
> > What can we do to get rid of this checkpoint error? Should we increase
> > the size of my redo logs?
>
> > Thanks in advance for your help.
>
> > Michael
>
> That is one way, presuming you're not executing regularly scheduled
> checkpoints; check the setting on log_checkpoint_interval to see if it
> is non-zero. Another way to address this is to add one or more log
> groups as the issue is usually caused when Oracle has cycled through
> the log groups and attempts to reuse a group and the checkpoint is
> still in progress.
>
> David Fitzjarrell.
Date: Mon, 21 Sep 2009 07:23:45 -0700 (PDT)
Message-ID: <b325bfe3-9868-477e-b043-e7c4bfb320ec_at_33g2000vbe.googlegroups.com>
On Sep 21, 9:50 am, ddf <orat..._at_msn.com> wrote:
> On Sep 21, 8:10 am, Michael <mbulch..._at_gmail.com> wrote:
>
> > We have 5 redo logs each of 150M in size and there are 2 members in
> > each group. We have been constantly getting the following error
> > message:
> > Checkpoint not complete
> > Current log# 2 seq# 127927 mem# 0: /u01/NI00/redo_tbs/redo_2a.log
> > Current log# 2 seq# 127927 mem# 1: /u03/NI00/redo_tbs/redo_2b.log
>
> > What can we do to get rid of this checkpoint error? Should we increase
> > the size of my redo logs?
>
> > Thanks in advance for your help.
>
> > Michael
>
> That is one way, presuming you're not executing regularly scheduled
> checkpoints; check the setting on log_checkpoint_interval to see if it
> is non-zero. Another way to address this is to add one or more log
> groups as the issue is usually caused when Oracle has cycled through
> the log groups and attempts to reuse a group and the checkpoint is
> still in progress.
>
> David Fitzjarrell.
The "Checkpoint not complete" is a fairly reliable indicatior that your online redo logs are size too small. In the absence of the use of Data Guard in a log shipping mode you should likely resize your online redo logs. you can check this by reviewing how many are created and how much time passes between log switches using the alert log or the v$archived_log view.
Generally speaking you probably want only 24 - 48 redo log switches per day. (one every 30 - 60 minutes on average).
HTH -- Mark D Powell -- Received on Mon Sep 21 2009 - 09:23:45 CDT