cman 无法启动

gl00ad

openais[13904]: [CMAN ] cman killed by node 2 for reason 2

I try to find out what is reason 2 you got, not lucky for now.

My configuration

[root@s12 ~]# lsof -i UDP:5405
COMMAND  PID USER   FD   TYPE DEVICE SIZE NODE NAME
aisexec 2367 root    6u  IPv4   8200       UDP 239.192.2.98:netsupport
aisexec 2367 root    8u  IPv4   8202       UDP s12.example.com:netsupport
[root@s12 ~]# cat /etc/hosts
# Do not remove the following line, or various programs
# that require network functionality will fail.
127.0.0.1       localhost.localdomain   localhost
::1     localhost6.localdomain6 localhost6
192.168.0.12    s12.example.com s12

from this

[root@cms2 ~]# cat /etc/hosts
# Do not remove the following line, or various programs
# that require network functionality will fail.
127.0.0.1       localhost.localdomain   localhost       cms2.Guangdong
::1     localhost6.localdomain6 localhost       cms2.Guangdong6
192.168.201.1 cms1
192.168.201.2 cms2

to

[root@cms2 ~]# cat /etc/hosts
# Do not remove the following line, or various programs
# that require network functionality will fail.
127.0.0.1       localhost.localdomain   localhost
::1     localhost6.localdomain6 localhost6
192.168.201.1 cms1.Guangdong     cms1
192.168.201.2 cms2.Guangdong     cms2

do

hostname
service cman restart
lsof -i UDP:5405

what about your ifcfg-ethx/bondx and ifcfg-ethx:y (floating ip)?

[ 本帖最后由 gl00ad 于 2008-10-26 23:40 编辑 ]

oioilu

我更改了hosts文件后，cman重启还是fence设备报错。log显示依然是"[CMAN ] cman killed by node 2 for reason 2 "

使用lsof -i UDP:5405，已经没有显示，表明端口已经空闲。

以下是ifcfg的配置。

1. 
   
2. [root@cms2 network-scripts]# more ifcfg-bond0
   
3. DEVICE=bond0
   
4. BOOTPROTO=none
   
5. IPADDR=61.8.165.36
   
6. NETMASK=255.255.255.240
   
7. ONBOOT=yes
   
8. TYPE=Ethernet
   
9. GATEWAY=61.8.165.33
   
10. USERCTL=no
    
11. [root@cms2 network-scripts]# more ifcfg-eth0
    
12. BOOTPROTO=none
    
13. DEVICE=eth0
    
14. ONBOOT=yes
    
15. IPADDR=192.168.201.2
    
16. NETMASK=255.255.255.0
    
17. HWADDR=00:1e:4f:39:91:92
    
18. TYPE=Ethernet
    
19. USERCTL=no
    
20. IPV6INIT=no
    
21. PEERDNS=yes
    
22. [root@cms2 network-scripts]# more ifcfg-eth1
    
23. BOOTPROTO=none
    
24. DEVICE=eth1
    
25. ONBOOT=yes
    
26. MASTER=bond0
    
27. SLAVE=yes
    
28. TYPE=Ethernet
    
29. [root@cms2 network-scripts]# more ifcfg-eth2
    
30. BOOTPROTO=none
    
31. DEVICE=eth2
    
32. ONBOOT=yes
    
33. MASTER=bond0
    
34. SLAVE=yes
    
35. TYPE=Ethernet
    

复制代码

[ 本帖最后由 oioilu 于 2008-10-27 11:15 编辑 ]

oioilu

I get an official answer from a Redhat Guy:

It means you have a very old version of cman that needs updating
That message as from 5.0 and lots of things have been fixed (including
that error) since then .

Chrissie

Reason "2" is that someone issued a cman_tool kill command on another
node. So it's nothing wrong with the cluster that has caused that message.

Chrissie

[ 本帖最后由 oioilu 于 2008-10-28 17:51 编辑 ]

gl00ad

congradulation and thanks for sharing!

jerrywjl

原帖由 oioilu 于 2008-10-27 17:58 发表
I get an official answer from a Redhat Guy:

谁说的？！我怎么不知道RHEL5上有这样的问题？

首先，如果你的系统和所使用的集群软件来自同一个光盘，那么版本匹配方面不会有问题，即便RHEL5的GA版本有这样那样的毛病，但是还不至于有这样严重的问题。
所以你能否尽量将情况描述清楚一些？
例如：
你的物理拓扑结构是怎样的？然后现在的具体问题是什么？你按照什么样的步骤进行了排查？！

我建议你将上述情况说清楚，然后将两台设备的时间调整一致，然后同时重启，之后将你的故障现象重现一次之后在两台设备上各运行一次sosreport命令。
最后将两台设备的sosreport产生的bz2文件提供出来，我有时间会帮你看看。

gl00ad

this is mine, how to find cman version?

[root@s12 ~]# cman_tool version
6.1.0 config 17
CMAN 2.0.84 (built May 25 2008 02:03:53) started

His

CMAN 2.0.60 (built Jan 23 2007 12:42:29) started

jerrywjl

原帖由 gl00ad 于 2008-10-27 22:34 发表
this is mine, how to find cman version?


His
  

看所安装的rpm包也行啊！rpm -q cman

gl00ad

his

CMAN 2.0.60 (built Jan 23 2007 12:42:29) started

in the log file in the first page

oioilu

原帖由 jerrywjl 于 2008-10-27 22:17 发表



谁说的？！我怎么不知道RHEL5上有这样的问题？

首先，如果你的系统和所使用的集群软件来自同一个光盘，那么版本匹配方面不会有问题，即便RHEL5的GA版本有这样那样的毛病，但是还不至于有这样严重的问题 ...

我的现在已经是生产环境了，不可能为了这个而重起的，毕竟不是研究性质。

另外redhat的人员联系方式为:Christine Caulfield <ccaulfie@redhat.com>  (Red Hat)

[ 本帖最后由 oioilu 于 2008-10-27 23:47 编辑 ]

oioilu

Update from Redhat ppl:

Reason "2" is that someone issued a cman_tool kill command on another
node. So it's nothing wrong with the cluster that has caused that message.