论坛徽章:: 0

电梯直达

1楼 [收藏(0)] [报告]

发表于 2012-01-18 19:47 |只看该作者 |倒序浏览

本帖最后由 shitoryu 于 2012-01-20 01:06 编辑

参照网上的文档，配置RHCS，并添加了一个http的集群服务，资源包括虚拟IP,filesystem（gfs），httpd的启动脚本。
目前测试服务切换，当一个节点ha1.com的网络失效或者节点ha1.com关机，http服务都能切换到节点ha2.com上。
把ha1上的HBA卡的光纤线拔掉，用clustat查看集群状态，
显示如下：
Member Name                            ID Status
------ ----                            ---- ------
ha1.com                                  1 Online, Local,rgmanager
ha2.com                                  2 Online, rgmanager

Service Name                Owner (Last)                State
------- ----                ----- ------                -----
service:redhat_http          (ha2.com)                      recoverable

服务一直不能切换到ha2,直到ha1完全关机，则服务会切换到ha2.com上。

我的疑问是 http服务的资源中包括了filesystem,hba卡链路断掉，相当于不能访问文件系统，为什么切换没有发生？

ha1中的log如下：Jan 18 19:38:26 ha1 kernel: qla2xxx 0000:06:01.0: LIP reset occured (f7f7).
Jan 18 19:38:31 ha1 kernel: qla2xxx 0000:06:01.0: LOOP DOWN detected (2 e678 0).
Jan 18 19:38:42 ha1 kernel:  rport-0:0-0: blocked FC remote port time out: saving binding
Jan 18 19:38:46 ha1 kernel: sd 0:0:0:1: SCSI error: return code = 0x00010000
Jan 18 19:38:46 ha1 kernel: end_request: I/O error, dev sda, sector 48386825
Jan 18 19:38:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: gfs2_quotad: statfs error -5
Jan 18 19:39:16 ha1 kernel: sd 0:0:0:1: SCSI error: return code = 0x00010000
Jan 18 19:39:16 ha1 kernel: end_request: I/O error, dev sda, sector 48386825
Jan 18 19:39:16 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: gfs2_quotad: statfs error -5
Jan 18 19:39:46 ha1 kernel: sd 0:0:0:1: SCSI error: return code = 0x00010000
Jan 18 19:39:46 ha1 kernel: end_request: I/O error, dev sda, sector 47443065
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14439
Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14440
Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14441
Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14442
Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14443
Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: sd 0:0:0:1: SCSI error: return code = 0x00010000
Jan 18 19:39:46 ha1 kernel: end_request: I/O error, dev sda, sector 47443105
Jan 18 19:39:46 ha1 kernel: Buffer I/O error on device sda9, logical block 14444

Jan 18 19:39:46 ha1 kernel: lost page write due to I/O error on sda9
Jan 18 19:39:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: fatal: I/O error
Jan 18 19:39:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: block = 14444
Jan 18 19:39:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: function = log_write_header, file = fs/gfs2/log.c, line = 622
Jan 18 19:39:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: about to withdraw this file system
Jan 18 19:39:46 ha1 kernel: GFS2: fsid=new_cluster:gfs1.0: telling LM to withdraw
Jan 18 19:40:09 ha1 clurgmgrd: [2984]: <err> /share_fs is not a directory
Jan 18 19:40:09 ha1 clurgmgrd[2984]: <notice> status on clusterfs "share_fs" returned 1 (generic error)
Jan 18 19:40:09 ha1 clurgmgrd[2984]: <notice> Stopping service service:redhat_http
Jan 18 19:40:10 ha1 avahi-daemon[2480]: Withdrawing address record for 192.168.101.15 on eth0.
Jan 18 19:42:41 ha1 kernel: qla2xxx 0000:06:01.0: Loop down - aborting ISP.
Jan 18 19:42:41 ha1 kernel: qla2xxx 0000:06:01.0: Performing ISP error recovery - ha= f7c602e0.
Jan 18 19:43:01 ha1 kernel: qla2xxx 0000:06:01.0: Cable is unplugged...

谢谢！

文库|博客

shitoryu

丰衣足食

论坛徽章:: 0

2楼 [报告]

发表于 2012-01-18 20:45 |只看该作者

本帖最后由 shitoryu 于 2012-01-18 20:46 编辑

机器是 HP dl380 g3,操作系统redhat 企业版5.5,光纤卡是fc2214,用系统自带的驱动，每个主机各一块光纤卡，存储是hp的msa1000
cluster.conf配置如下：
<?xml version="1.0"?>
<cluster alias="new_cluster" config_version="11" name="new_cluster">
      <fence_daemon post_fail_delay="0" post_join_delay="3"/>
      <clusternodes>
            <clusternode name="ha1.com" nodeid="1" votes="1">
                     <fence>
                              <method name="1">
                                    <device name="ILOD306LDN1H634"/>
                              </method>
                     </fence>
            </clusternode>
            <clusternode name="ha2.com" nodeid="2" votes="1">
                     <fence>
                              <method name="1">
                                    <device name="ILOD306LDN1H566"/>
                              </method>
                     </fence>
            </clusternode>
      </clusternodes>
      <cman expected_votes="1" two_node="1"/>
      <fencedevices>
            <fencedevice agent="fence_ilo" hostname="ilo_ha1" login="administrator" name="ILOD306LDN1H634" passwd="xxxxx"/>
            <fencedevice agent="fence_ilo" hostname="ilo_ha2" login="administrator" name="ILOD306LDN1H566" passwd="xxxxx"/>
      </fencedevices>
      <rm>
            <failoverdomains>
                     <failoverdomain name="fd" ordered="0" restricted="0">
                              <failoverdomainnode name="ha1.com" priority="1"/>
                              <failoverdomainnode name="ha2.com" priority="1"/>
                     </failoverdomain>
            </failoverdomains>
<resources>
                     <clusterfs device="/dev/sda9" force_unmount="1" fsid="42707" fstype="gfs2" mountpoint="/share_fs" name="share_fs" options=""/>
                     <ip address="192.168.101.15" monitor_link="1"/>
                     <script file="/etc/init.d/httpd" name="redhat_init_http"/>
            </resources>
            <service autostart="1" domain="fd" name="redhat_http">
                     <ip ref="192.168.101.15"/>
                     <clusterfs ref="share_fs"/>
                     <script ref="redhat_init_http"/>
            </service>
      </rm>
</cluster>