HP 集群软件

我是007

问题
# cmcheckconf -v -C /etc/cmcluster/cmclconfig.ascii
Begin cluster verification...  
Checking cluster file: /etc/cmcluster/cmclconfig.ascii  
Defaulting MAX_CONFIGURED_PACKAGES to 300.  
Checking nodes ... Done  
Checking existing configuration ... Done  
Defaulting MAX_CONFIGURED_PACKAGES to 300.  
Gathering storage information  
Unable to receive device query message from mucs3173: Software caused connection abort                                <--- 问题
Found 148 devices on node mucs3088  
Found 148 devices on node mucs3090  
Found 150 devices on node mucs3091  
Found 0 devices on node mucs3173     <----------  这里
Found 148 devices on node mucs3179  
Analysis of 594 devices should take approximately 16 seconds  
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%  
Found 59 volume groups on node mucs3088  
Found 59 volume groups on node mucs3090  
Found 59 volume groups on node mucs3091  
Found 0 volume groups on node mucs3173  
Found 59 volume groups on node mucs3179  
Analysis of 236 volume groups should take approximately 1 seconds  
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%  
Gathering network information  
Beginning network probing (this may take a while)  
Completed network probing  
Gathering polling target information  
cmcheckconf: Unable to reconcile configuration file /etc/cmcluster/cmclconfig.ascii  

不管是在哪个节点运行 cmcheckconf ，结果都一样.
配置 :
HPUX 11.31.
Serviceguard:A.11.19和修补软件PHSS_40152（在受影响的节点上安装了PHSS_41162，但是没有帮助）.
top

解决办法
mucs3173 syslog.log 中包含很多以下信息:
cmclconfd[29685]: Could not get vg (/dev/vg3139_TAQ_A) info: 3

使用 cmscancl 命令取得所有节点的 /etc/lvmtab 内容并且发现只有受影响的节点上有 vg3139_TAQ_A 和 vg3139_TAQ_old :
$ grep -e lvmtab -e vg3139_TAQ scancl.out  
------ Output of strings /etc/lvmtab (mucs3090) ------  
/dev/vg3139_TAQ
------ Output of strings /etc/lvmtab (mucs3088) ------  
/dev/vg3139_TAQ
------ Output of strings /etc/lvmtab (mucs3173) ------  
/dev/vg3139_TAQ_A      <---??
/dev/vg3139_TAQ_old    <---??
/dev/vg3139_TAQ     
------ Output of strings /etc/lvmtab (mucs3179) ------  
/dev/vg3139_TAQ
------ Output of strings /etc/lvmtab (mucs3091) ------  
/dev/vg3139_TAQ

如果加上 -k 选项cmcheckconf 会顺利执行. 这个选项消除了检查LVM磁盘, 所以这就确定了问题和LVM问题有关系.
动作 :

• ·检查 /dev/vg3139_TAQ_A 是否存在于受影响的节点上.
  
• 如果卷组 vg3139_TAQ_A 不需要了, vgexport 它.
  

效果 :
vgexport 解决了问题.

地址：http://h20566.www2.hp.com/portal/site/hpsc/template.PAGE/public/kb/docDisplay?javax.portlet.begCacheTok=com.vignette.cachetoken&javax.portlet.endCacheTok=com.vignette.cachetoken&javax.portlet.prp_ba847bafb2a2d782fcbb0710b053ce01=wsrp-navigationalState%3DdocId%253Demr_na-c02491156-1%257CdocLocale%253D%257CcalledBy%253D&javax.portlet.tpst=ba847bafb2a2d782fcbb0710b053ce01&sp4ts.oid=4162060&ac.admitted=1408344051145.876444892.199480143