硬件:2*dell6850(4*cpu,双核,rem:8G),存储:EMC(cx300),交换机,SAN网络
软件:Redhat Linux Server4,OCM,Oracle9204
在上面列的情况下,采用哪种方案,系统的可用性最高?
我在尝试采用RAC,cluster file system:ocfs2,感觉ocfs容易死机,
如:
1、不明原因一台机器死机,另外一台ocfs超时也会死机;
2、reboot一台机器,有时另外一台机器也会死机,如:
May 17 23:33:13 rac2 kernel: ocfs2_dlm: Node 0 leaves domain 128A178940A64CA18AF439EF76C03168
May 17 23:33:13 rac2 kernel: ocfs2_dlm: Nodes in domain ("128A178940A64CA18AF439EF76C03168"
: 1
May 17 23:33:21 rac2 kernel: ocfs2_dlm: Node 0 leaves domain 1DEFB76B539E4BF5AD2BA52FEF5B5BE6
May 17 23:33:21 rac2 kernel: ocfs2_dlm: Nodes in domain ("1DEFB76B539E4BF5AD2BA52FEF5B5BE6"
: 1
May 17 23:33:22 rac2 kernel: o2net: no longer connected to node rac1 (num 0) at 192.168.0.1:7777
May 17 23:33:24 rac2 sshd(pam_unix)[15742]: session opened for user root by root(uid=0)
May 17 23:37:22 rac2 kernel: (3822,4)
2net_connect_expired:1444 ERROR: no connection established with node 0 after 10 second[/COLOR]
May 18 12:46:08 rac2 syslogd 1.4.1: restart.
May 18 12:46:08 rac2 syslog: syslogd startup succeeded
May 18 12:46:08 rac2 kernel: klogd 1.4.1, log source = /proc/kmsg started.
就象上面看到的,系统什么信息也没有记录,就死在哪里,只有断电重启。
综上所述,对高可用性,就上面我们具有的条件,请各位指点,采用什么方式最好,如果采用RAC,要如何改进?