Ultrapower
解决办法：
a) in the %PATROL_HOME%lib\\knowledge directory, rename StdEvents.ctg to StdEvents.old and StdEvents.ctg_NoEVS to StdEvents.ctg.
b) via a text editor, modify the following line
in %PATROL_HOME%lib\\knowledge\\StdEvents.ctg which is associated with event class 28:
NOTIFY_PSL_ACTION LOAD \ to
NOTIFY_ACTION 784393923 \
The PATROL Agent can now be successfully started and the default account can be changed via wpconfig, xpconfig or pconfig utilities.
2.2 patrol for winows agent 安装失败
patrol 版本：
OS版本： WINDOWS SERVER 问题描述：
[17:13:10] CreateFolder:Error: User Profile does not exist. Path: C:\\Documents and Settings\\patrol\\??????\\?? [17:13:10] Unable to create Start Menu item because the current user's profile does not exist. [17:13:10] Unable to create folder [BMC PATROL] for product [nt86a]. Please refer to install log file for more details. [17:13:10] _InstallAProduct:Installation of &PATROL Agent for Microsoft Windows Servers aborted.
原因分析：
用户权限原因不能建立相关目录 解决办法：
3 4 5
5.1
数据库安装问题 中间件及其他安装问题 Unix 配置问题
重起agent后取不到MEMSize数据
patrol 版本：9211
28357950.doc
21/32
Ultrapower
OS版本： Linux 问题描述：
去年做过内存KM的升级,解决了Linux系统不能取内存总大小指标的问题,并将内存总大小名称由MEMTotalMemory更改为MEMSize,然后将新KM下发给各主机,MEMSize可以正常取值(# snmpwalk -v 1 -c public -p 8161 10.33.25.4 .1.3.6.1.4.1.1031.1.1.1.5.1.4.28.47.77.69.77.79.82.89.47.77.69.77.79.82.89.47.77.69.77.83.105.122.101.47.118.97.108.117.101 enterprises.1031.1.1.1.5.1.4.28.47.77.69.77.79.82.89.47.77.69.77.79.82.89.47.77.69.77.83.105.122.101.47.118.97.108.117.101 = \ 但重起agent之后发现,MEMSize又取不到了,其他数据都有,重新下发KM才能取到.
原因分析：
Commit后出现，表示KM本身无问题，请确认Agent端
$PATROL_HOME/lib/knowledge的MEMORY.KM是否与Console端匹配。如果匹配，重启Agent，不做commit再看下。如果不匹配，手动Copy过去再重启观察下。 解决办法：
Agent端$PATROL_HOME/lib/knowledge的MEMORY.KM仍是老KM,将新KM替换后,问题解决
5.2
在solaris5.8平台安装了patrolgent后，在进行psl模式向dcm切换失败
patrol 版本：9211 OS版本： solaris5.8 问题描述：
在solaris5.8平台安装了patrolgent后，在进行psl模式向dcm切换失败
原因分析：
Sun Solaris 8 Kernel Version 108528-28 and 108528-29 patches 的解决办法：
升级SUN的patch 117000-03，
bug，
参考：
http://documents.bmc.com/supportu/documents/59/68/45968/Output/090f44b1802bd51d.htm
5.3
patrol for unix 下CPUUserTime超过100
patrol 版本：9211 OS版本： AIX5.3 问题描述：
28357950.doc
22/32
Ultrapower
AIX5.3平台安装patrol for unix出现CPUUserTime超过100
原因分析：
AIX系统bug 解决办法：
升级AIX5.3的patch APAR IY77690，
参考：
http://documents.bmc.com/supportu/documents/98/32/59832/Output/090f44b1803c08e6.htm
IBM的官方说法：
http://www-1.ibm.com/support/docview.wss?uid=isg1IY77690 5.4
patrol for linux下启动agent出错
patrol 版本：9300
OS版本： Redhat Linux AS4.0 问题描述：
在Redhat Linux AS4.0下完成安装patrol for linux and unix 9300后，启动patrolagent报如下错误：-> /home/patrol/bmcsoft/Patrol3/Linux-2-6-x86-nptl/bin/PatrolAgent: error while loading shared libraries: libstdc++.so.5: cannot open shared object file: No such file or directory
原因分析：
解决办法： 没有结论
5.5 patrol console for windows unix主机中没有network 选项
patrol 版本：9211 OS版本： AIX 问题描述：
在console端对unix主机进行监控配置，发现没有NETWORK选项
原因分析：
Unix内存KM需要手工激活.
28357950.doc
23/32
Ultrapower
解决办法：
The NETWORK KM is activated on the 9.1.xx series of the PATROL for UNIX and Linux by a menu option from the UNIX_OS icon. From the UNIX_OS icon, select the 'Knowledge Module Admin' menu option, then select 'NETWORK Application', and then select 'Activate Monitoring'.--
5.6 patrol for unix 下patrol agent停止
patrol 版本：9300 OS版本： SunOS 5.8 问题描述：
在console端对unix主机进行监控配置，发现没有NETWORK选项
原因分析：
$ ./PatrolAgent -version
$ PatrolAgent V3.6.60i SunOS 5.8 sun4u sparc Build_200608180248 Aug 18 2006
PATROL AGNET FOR UNIX 9300KM $ uname -a
SunOS bjreuter1 5.8 Generic_117350-38 sun4u sparc SUNW,Sun-Fire-V240 解决办法：
Patrol Agent 3.6.60的 bug 安装Agent 3.6.60.01 patch
参考：
http://documents.bmc.com/supportu/documents/56/29/65629/65629.pdf
注意：BMC反馈是unix和windows都应该升级
5.7
关于patrol for unix agent中文件系统的解释
patrol 版本： OS版本： 问题描述：
patrol for unix agent中文件系统的解释
原因分析：
If the FILESYSTEM exclusion list has not been modified to exclude type NFS or to exclude the
specific instance and the mount command shows the file system, an icon will be created and
monitoring will occur just like any other file system. The mount commands used for each of the
different supported OS types can be seen in the FILESYSTEM discovery psl file,
28357950.doc
24/32
Ultrapower
which is
./lib/psl/FILESYSV2.psl and also are listed below.
============================================================================== OSF1 / RS6000 mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ HPUX mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ DG / SUN4 / BULL mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ ULTRIX mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------
SOLARIS / SINIX / SVR4-SPARC / SVR4-i386 / SVR4-m88K mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ SGI mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ i386-SEQUENT mnt_cmd = \
------------------------------------------------------------------------------------------------------------------------------------------------------------ DC-OSX / Linux mnt_cmd = \
==============================================================================
Also note that due to the inherent dynamic nature of NFS mounted file systems, the mount point will
go away if the file system has not been accessed recently. The file system will automatically get
unmounted if it sets idle for some specific amount of time which can be different on each OS and/or
environment. In most cases it is common to either accept this realizing the icons will gray out going
OFFLINE for periods of time, or put type NFS in the file system exclusion list, or setup an OFFLINE
state change action on the specific file system instance so someone can be alerted when it goes
OFFLINE. Although as dynamic as NFS mounts are, the OFFLINE state change action is not a
very viable alternative unless the NFS mount is known to be and should be active all the time.
This OFFLINE state change action is more useful on local or static mounts 解决办法：
28357950.doc 25/32