Re: Redo log size on a 4 node RAC cluster.

From: Powell, Mark <mark.powell2_at_dxc.com>
Date: Wed, 23 Aug 2017 13:24:15 +0000
Message-ID: <DF4PR84MB0124582B5494B3666A1C4C97CC850_at_DF4PR84MB0124.NAMPRD84.PROD.OUTLOOK.COM>

Just because you have defined the redo logs at 1.5G does not mean you use all the space before you switch depending on how you have configured your system and if RAC, what are your database settings for log_checkpoint_interval and log_checkpoint_timeout?

How many redo log groups do you have? From experience we have normally needed 4 to 6 redo log groups.

You appear to have Data Guard in use. With Data Guard you normally want to use smaller redo logs than when DG is not in use, but would likely need more groups.

Mark Powell
Database Administration
(313) 592-5148

---

From: oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org> on behalf of Zabair Ahmed <dmarc-noreply_at_freelists.org> Sent: Wednesday, August 23, 2017 6:58:53 AM To: oracle-l_at_freelists.org
Subject: Redo log size on a 4 node RAC cluster.

Hello,

11.2.0.4 EE on 4 node RAC Redhat 6.6 Cluster.

We are constantly seeing "Checkpoint not complete" message in the alert log:

Thread 2 cannot allocate new log, sequence 109658 Checkpoint not complete
  Current log# 26 seq# 109657 mem# 0: +ARCH01/prd/onlinelog/group_26.393.904013409   Current log# 26 seq# 109657 mem# 1: +ARCH01/prd/onlinelog/group_26.634.904013411 Thread 2 advanced to log sequence 109658 (LGWR switch) Tue Aug 22 00:03:06 2017
LNS: Standby redo logfile selected for thread 2 sequence 109658 for destination LOG_ARCHIVE_DEST_2   Current log# 3 seq# 109658 mem# 0: +ARCH01/prd/onlinelog/group_3.263.877611157   Current log# 3 seq# 109658 mem# 1: +ARCH01/prd/onlinelog/group_3.273.877680615 Tue Aug 22 00:03:11 2017
Archived Log entry 510017 added for thread 2 sequence 109657 ID 0x1b7b9f46 dest 1: Tue Aug 22 00:11:39 2017
Thread 2 cannot allocate new log, sequence 109659 Checkpoint not complete
  Current log# 3 seq# 109658 mem# 0: +ARCH01/prd/onlinelog/group_3.263.877611157   Current log# 3 seq# 109658 mem# 1: +ARCH01/prd/onlinelog/group_3.273.877680615 Thread 2 advanced to log sequence 109659 (LGWR switch)   Current log# 22 seq# 109659 mem# 0: +ARCH01/prd/onlinelog/group_22.744.904009995   Current log# 22 seq# 109659 mem# 1: +ARCH01/prd/onlinelog/group_22.931.904009999 Tue Aug 22 00:11:44 2017
LNS: Standby redo logfile selected for thread 2 sequence 109659 for destination LOG_ARCHIVE_DEST_2 Tue Aug 22 00:11:50 2017
Archived Log entry 510019 added for thread 2 sequence 109658 ID 0x1b7b9f46 dest 1: Tue Aug 22 00:21:55 2017
Thread 2 cannot allocate new log, sequence 109660 Checkpoint not complete
  Current log# 22 seq# 109659 mem# 0: +ARCH01/prd/onlinelog/group_22.744.904009995   Current log# 22 seq# 109659 mem# 1: +ARCH01/prd/onlinelog/group_22.931.904009999 Thread 2 advanced to log sequence 109660 (LGWR switch)   Current log# 4 seq# 109660 mem# 0: +ARCH01/prd/onlinelog/group_4.265.877611161   Current log# 4 seq# 109660 mem# 1: +ARCH01/prd/onlinelog/group_4.274.877680623 Tue Aug 22 00:22:00 2017
LNS: Standby redo logfile selected for thread 2 sequence 109660 for destination LOG_ARCHIVE_DEST_2 Tue Aug 22 00:22:09 2017
Archived Log entry 510026 added for thread 2 sequence 109659 ID 0x1b7b9f46 dest 1: Tue Aug 22 00:30:36 2017
Thread 2 cannot allocate new log, sequence 109661 Checkpoint not complete
  Current log# 4 seq#

Also ASH Report listing "log file switch (checkpoint incomplete)" as the Top User Event.

All this is indicating that we have redo logs which are incorrectly sized.

My question is how do you determine the correct redo log size? Currently the redo log size 1.5GG in all 4 instances.

Can you mix and match different redo log size for different instances? Is this recommended? We are getting the "Checkpoint not complete" message only on Instance 2!

Are there any disadvantages in having different redo log sizes for different instances? Do other people do this? Or should we keep the redo log size the same across all 4 instances?

We are also running this in a DataGuard environment.

Appreciate any feedback.

Thanks

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

Received on Wed Aug 23 2017 - 15:24:15 CEST