PowerHA_5.4连接EMC_Symmtrix启动慢的解决办法.docVIP

PowerHA 5.4连接EMC Symmtrix启动慢的解决办法在使用PowerHA / HACMP 时，如果连接的存储为 EMC 的 CLARiiON 、Symmetrix，并且使用了 PowerPath，那么可能会出现 Cluster 启动慢、甚至无法成功，takeover 慢、甚至无法成功的问题。现象判断：查看 hacmp.out 中是否出现 SCIOLSTART…… Invalid argument……关键词（如下图），可能 会有 EVENT START: config_too_long 360 TE_JOIN_NODE 事件。 本次案例环境为 AIX 5308 + PowerHA 5.4，所遇问题为 Cluster 启动非常慢，并且发生 config_too_long 360 TE_JOIN_NODE 事件。处理方法为：/usr/sbin/cluster/utilities/clcustomdisk -c -tdisk/pseudo/power -Ndisk/pseudo/power -gSCSI3 -hSCSI_TUR -b/usr/lpp/EMC/Symmetrix/bin/emcpowerreset -ptrue -mMKDEV（这里红色标记为参考文档1中错误的地方，以 emc69100为准，emc122557 也有提到。）处理完以后的正确信息样例： 详细信息，请见参考文档。参考文档： emc156011----cl_fscsilunreset routine reports error Invalid argument during the start of HACMP/usr/lpp/EMC/README.* emc69100emc122557CLARiiON 的处理，多了两个步骤，参考emc97234，emc104555。HDS 的存储也有类似问题，不过，HDS安装文档中有清晰的说明。本次案例 hacmp.out 处理：Oct 21 00:26:12 EVENT START: node_down htxx2Oct 21 00:26:31 EVENT COMPLETED: node_down htxx2 0Oct 21 00:26:31 EVENT START: node_down_complete htxx2Oct 21 00:26:31 EVENT COMPLETED: node_down_complete htxx2 0Oct 21 00:27:25 EVENT START: network_down -1 net_ether_010Oct 21 00:27:25 EVENT COMPLETED: network_down -1 net_ether_010 0Oct 21 00:27:25 EVENT START: network_down_complete -1 net_ether_010Oct 21 00:27:26 EVENT COMPLETED: network_down_complete -1 net_ether_0100Oct 21 01:48:30 EVENT START: node_up htxx1Oct 21 01:54:28 EVENT START: config_too_long 360 TE_JOIN_NODEOct 21 02:21:51 EVENT START: node_up htxx1Oct 21 02:26:04 EVENT COMPLETED: node_up htxx1 0Oct 21 02:26:04 EVENT START: node_up_complete htxx1Oct 21 02:26:05 EVENT COMPLETED: node_up_complete htxx1 0Oct 21 02:26:18 EVENT START: node_up htxx2Oct 21 02:26:21 EVENT COMPLETED: node_up htxx2 0Oct 21 02:28:24 EVENT START: node_up_complete htxx2Oct 21 02:28:24 EVENT COMPLETED: node_up_complete htxx2 0Oct 21 02:46:33 EVENT START: node_down htxx2 gracefulOct 21 02:46:53 EVENT COMPLETED: node_down htxx2 graceful 0Oct 21 02:46:53 EVENT START: node_down_complete htxx