Applies To:: How To Install Oracle Cluster Health Monitor (Former IPD/OS) On Windows (Doc ID 811151.1)

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Document Display https://support.oracle.com/epmos/faces/SearchDocDisplay?_adf.ctrl-stat...

Copyright (c) 2019, Oracle. All rights reserved. Oracle Confidential.

How to Install Oracle Cluster Health Monitor (former IPD/OS) on Windows (Doc ID 811151.1)

In this Document

Goal
Solution
References

APPLIES TO:

Oracle Database - Enterprise Edition - Version 10.2.0.1 to 11.1.0.7 [Release 10.2 to 11.1]
Microsoft Windows x64 (64-bit)
Microsoft Windows (32-bit)

GOAL

This note describes how to install Oracle Cluster Health Monitor (IPD/OS) on Windows. If you need to know more about the advantages of
the Oracle Cluster Health Monitor (IPD/OS) on Windows please see Note 810915.1 .

SOLUTION

'Cluster Health Monitor' can be downloaded from OTN:

http://download.oracle.com/otndocs/products/clustering/ipd/crfpack-winnt.zip

Be aware that Windows Server 2003 Service Pack 2 or Windows Server 2008 R2 is required to install and operate the tool.
Please note that IPD/OS has issues with Windows Server 2008 R1 and Windows Server 2003 Service Pack 1.

The Berkley DB cannot be installed on the C:\ drive and MUST not be on shared disk.

Make sure that the perl.exe is in your path. Temporarily you can set it for example, as follows:

==============
10gR2 example:
==============
set PATH=%PATH%;D:\oracle\product\10.2.0\db_1\perl\5.8.3\bin\MSWin32-x86-multi-thread

==============
11gR1 example:
==============
set PATH=%PATH%;D:\app\11.1.0\grid\perl\bin

(path for the perl.exe above needs to reflect the path on your server)

Ensure that the user that runs the CHM does not have the ORACLE_HOME variable set as this will inhibit the startup of the CHM. You should
be aware that it is recommended as a best practice for RAC on Windows to NOT have the ORACLE_HOME environment variable set even
when it is set to a null value.

Installation

1. the administrator user should be used to install the tool and no additional user is required for that.

2. make sure that UNC (uniform naming convention, such as \\Server\Volume\File) path is configured on all nodes where OSTool is to be
installed.

3. If you have a previous install of this tool, delete it from all nodes. Login as privileged user (admin on Windows):

1 de 4 13/12/2019 12:44
Document Display https://support.oracle.com/epmos/faces/SearchDocDisplay?_adf.ctrl-stat...

a. Disable the tool

C:\Program Files\oracrf\bin>stopcrf

b. Uninstall

C:\Program Files\oracrf\install>perl crfinst.pl -d

c. Make sure all BDB databases are deleted from all nodes.

d. Manually delete the install home if it still exists.

4. Login as admin user.


Unzip the crfpack.zip file.

5. Run crfinst.pl (see below for usage details) script on a node with desired node list, specified as comma separated list, for cluster-wide
install.
You will find this script in the install subdirectory. An example invocation will look like this:

perl D:\ipdos\crfpack-winnt\install\crfinst.pl -i win1,win2 -b d:\ipdos

(where win1 and win2 are the node names of the cluster and d:\ipdos is location for the Berkeley DB. If you install the Berkeley DB make
sure you precreate the location (e.g d:\ipdos) on all nodes, the location MUST be on a non shared drive (local))

(if you encounter: Perl lib version (v5.8.3) doesn't match executable version (v5.10.0)... set environment variable perl5lib. Example: set
PERL5LIB=%ORACLE_HOME%\perl\lib)

confirm the recommended master node:

Shall I use win1 as master?(y/n)[y]: y

Details regarding Berkley DB:


a Berkeley DB can be created to
store OS metrics. This location MUST be outside of the location
where you unzipped the ZIP file because all the directories
under that location which were created by unzip will be removed.
BDB files can be kept as it is for later usage. The location should
be a path on a volume with at least 2GB per node space available
and writable by privileged user only.This location is required to be same on all
hosts. If that can not be done, please specify a different location
during finalize (-f) operation on each host, following the above
size requirements. The path MUST not be on shared disk. If a shared
BDB path is provided to multiple hosts, BDB corruption will happen.

6. Once the step 5 finishes, it will instruct you to run crfinst.pl script with -f and optionally -b <bdb location> on each node (including the
first / master node) to finalize the install:

C:\Program Files\oracrf\install>perl crfinst.pl -f -b d:\ipdos

(where d:\ipdos is location for the Berkeley DB)

7. Enable the tool on ALL nodes. Once the finalize operation is complete, from windows command prompt run the following while logged in
as privileged user:

c:\runcrf

this starts the OSYSMOND service (OracleOsToolSysmService). Service is created with autostart attribute and runs in REALTIME.

2 de 4 13/12/2019 12:44
Document Display https://support.oracle.com/epmos/faces/SearchDocDisplay?_adf.ctrl-stat...

In case the install fails complaining that msvcp71.dll and msvcr71.dll are missing you can get the missing DLLs if you install Berkeley DB
version 4.6.21 on a spare machine, we only need to extract the 2 DLLs (download file db-4.6.21.msi from OTN -> Download ->
Database -> Berkeley DB: Berkeley DB 4.6.21.msi Windows installer, with AES encryption (16.1M)) . From the box where you installed
BerkeleyDB copy the msvcp71.dll and msvcr71.dll (c:\program Files (x86)\Oracle\Berkeley DB 4.6.21\bin) to the install node into the
unzipped crfpack-winnt bin directory. Now start the install again

8. Increase the retention time for CHM data

By default the Berkeley DB stores CHM data only for 24 hours. To have a chance to investigate a Saturday morning incident on Monday, it is
recommended to increase the retention time to the maximum of 72 hours.

C:\Program Files\oracrf\bin>oclumon manage -bdb resize 259200

In the current version you cannot query the actual retention time.

Install the GUI

(Oracle recommends you run the gui from outside the cluster you are monitoring)
(You need to install the GUI with -g separately.)

C:\Program Files\oracrf\install>perl crfinst.pl -g d:\gui_ipdos

(d:\gui_ipdos is the loacation where the gui software will be installed)

Run the GUI


If you are running GUI from a node outside of the set of installed nodes, you MUST use the '-m' option:

d:\crfgui -m win1

(where win1 is a node where the tool is installed and is part of the cluster install provided to crfinst.pl script)

The default refresh rate of the GUI is 1 second. To change the refresh rate, use -r with number of seconds (i.e. -r 5 for a 5 second refresh
rate)

d:\crfgui -r 5 -m win1

(where win1 is a node where the tool is installed and is part of the cluster install provided to crfinst.pl script)

Invoking the GUI with '-d' option starts it in historical mode.

d:\crfgui -d "<hh>:<mm>:<ss>" -m <nodename>

where -d is used to specify hours (<hh>), minutes (<mm>) and seconds (<ss>) in the past from the current time to start the GUI from e.g.
crfgui -d "05:10:00"
starts the GUI and displays information from the database which is 5 hours and 10 minutes in the past from the current time.

Invoking the GUI with '-i' option provides the same shell at the command prompt as is seen in the GUI windows with a prompt of 'toprac>'.
You can use '?' at
this prompt to get detailed information about available commands and options.

To obtain help about more options execute:

d:\crfgui -h

Run Oclumon

3 de 4 13/12/2019 12:44
Document Display https://support.oracle.com/epmos/faces/SearchDocDisplay?_adf.ctrl-stat...

The commandline tool oclumon is included in the 'Cluster Health Monitor' package. This can be used to query the Berkeley DB backend to
print out the node specific metrics for a specified time period. The tool also supports a query to print the durations and the states for a
resource on a node during a specified time period.

To obtain information about this tool and its options execute:

d:\oclumon -h

To get detailed information about all nodes in the cluster use verbose (includes
process and device stats) mode:

d:\oclumon dumpnodeview -v -allnodes -last "00:30:00"

which will dump all stats for all nodes for last 30 minutes from the current
time.

To limit the scope of dump, use start and end times:

d:\oclumon dumpnodeview -allnodes -s "2008-11-12 12:30:00" -e "2008-11-12 13:30:00"

which will dump stats for all nodes from 12:30 to 13:30 on Nov 12th, 2008.

Note that you don't specify the timezone in these clocks. To find the timezone on the servers in the cluster, use oclumon dumpnodeview
-allnodes. All your
time specification should be in that timezone, without actually referring the timezone string (like UTC or PST8PDT) in the time.

You can use '-n' to specify a node name:

d:\oclumon dumpnodeview -v -n mynode -last "00:10:00"

will dump all stats for 'mynode' for last 10 minutes.

To use oclumon to query for alerts only, use the '-alert' option:

d:\oclumon dumpnodeview -v -allnodes -alert -last "00:30:00"

which will dump all records for all nodes for last 30 minutes, which contains at least one alert. It should be noted that if '-v' is not used
you may see records which apparently don't have any system wide alerts. (NOTE: system wide alerts are the alerts on the data that is
shown in
dumpnodeview output without the -v option.) The alert in those cases may come from one of processes or the devices.

Be aware that OCLUMON output prints to screen and you will need to spool the output to a file in order to send it to Oracle
Support

oclumon dumpnodeview -allnodes -s "2009-06-19 11:44:00" -e "2009 -06-19 11:46:00">c:\temp\output.txt

REFERENCES

NOTE:847485.1 - How to Collect 'Cluster Health Monitor' (former IPD/OS) Data on Windows Platform for Oracle Support
NOTE:810915.1 - How to Monitor, Detect and Analyze OS and RAC Resource Related Degradation and Failures on Windows
Didn't find what you are looking for?

4 de 4 13/12/2019 12:44

You might also like