redhat 群集配置问题求助.

Question

系统: redhat as 4 u3
        gfs 6.1.5
        2 nodes

已成功在本机编译所有相关的 SRPM 包并安装上.

但执行:
/sbin/modprobe lock_dlm
/sbin/service ccsd start     
/sbin/service cman start

出现下面的log ,估计是基本的 redhat 群集服务没有正常连接成功.

在网上苦找了N个不同的 cluster.conf 文件,乱拼大半天也是同样的信息出现.

求教 nntp ....

syslog:
Aug 22 03:56:10 server5 ccsd[7567]: Starting ccsd 1.0.3:
Aug 22 03:56:10 server5 ccsd[7567]:  Built: Aug 21 2006 06:35:32
Aug 22 03:56:10 server5 ccsd[7567]:  Copyright (C) Red Hat, Inc.  2004  All rights reserved.
Aug 22 03:56:11 server5 ccsd: startup succeeded
Aug 22 03:56:20 server5 ccsd[7567]: cluster.conf (cluster name = gfs_cluster, version = 5) found.
Aug 22 03:56:21 server5 kernel: CMAN: Waiting to join or form a Linux-cluster
Aug 22 03:56:21 server5 ccsd[7567]: Connected to cluster infrastruture via: CMAN/SM Plugin v1.1.5
Aug 22 03:56:21 server5 ccsd[7567]: Initial status:: Inquorate
Aug 22 03:56:37 server5 kernel: CMAN: sendmsg failed: -13
Aug 22 03:56:50 server5 last message repeated 4 times
Aug 22 03:56:53 server5 kernel: CMAN: forming a new cluster
Aug 22 03:56:53 server5 kernel: CMAN: quorum regained, resuming activity
Aug 22 03:56:53 server5 kernel: CMAN: sendmsg failed: -13
Aug 22 03:56:53 server5 kernel: CMAN: No functional network interfaces, leaving cluster
Aug 22 03:56:53 server5 ccsd[7567]: Cluster manager shutdown.  Attemping to reconnect...
Aug 22 03:56:53 server5 kernel: CMAN: sendmsg failed: -13
Aug 22 03:56:53 server5 kernel: CMAN: we are leaving the cluster.
Aug 22 03:56:53 server5 cman: cman_tool: can't open cluster socket, cman kernel module probably not loaded failed
Aug 22 03:57:11 server5 ccsd[7567]: Unable to connect to cluster infrastructure after 30 seconds.
Aug 22 03:57:41 server5 ccsd[7567]: Unable to connect to cluster infrastructure after 60 seconds.

==================

Answer 1

哦，恭喜,  不过有空还是看看到底是什么版本依赖性的问题.

不会我就是直接用iSCSI,没有问题. gfs stat看看.

Answer 2

应该是版本的问题,换在 u2 下,使用旧的gfs 版本 (6.1.2) ,成功安装... (回贴庆祝一下)

现在装了三个节点,今晚灌点数据再真正测试一下.

========
还有些新问题,使用

gfs_jadd /gfs目录

Requested size (32768 blocks) greater than available space (1 blocks)

不能动态增加节点上限,有点莫明其妙.

会不会是我没有用 gnbd ,直接用 iscsi 的原因?

Answer 3

"kernel: CMAN: No functional network interfaces, leaving cluster"

请按照redhat 官方的顺序安装准备环境 ->安装cluster package->配置 GFS ->启动集群

贴出来的message应该是configre problem