GFS2文件系统无法挂载

草中宝

三台服务器，其中两台是oracle 11G R2 rac节点，另外一台是oracle 11G dataguard。有共享存储系统，上面部署了gfs2文件系统供三台电脑访问。
每次自启，dataguard上的gfs2文件系统能正常挂载，其它的两台电脑（rac节点）则无法挂载gfs2文件系统，报错类似如下
Mounting GFS2 filesystem (/opt/app/oracle/archivelog/db_backup): invalid device path "/dev/vg_arch/lv_db_backup"

下面是rac两服务器节点详细的启动日志
[root@rac1 log]# more boot.log
Gmount: mount point /proc/bus/usb does not exist
                Welcome to Oracle Linux Server
Starting udev: G[  OK  ]
Setting hostname rac1.com:  [  OK  ]
Setting up Logical Volume Management:   Skipping clustered volume group vg_arch
  Skipping volume group vg_arch
[FAILED]                                                  <= dataguard服务器 也有该错误，但不影响dataguard服务器上的gfs2文件系统的挂载
Checking filesystems
/dev/sda2: clean, 91241/655360 files, 694002/2621440 blocks
/dev/sda1: clean, 44/76912 files, 73657/307200 blocks
/dev/sda5: clean, 54501/804672 files, 2445104/3214848 blocks
[  OK  ]
Remounting root filesystem in read-write mode:  [  OK  ]
Mounting local filesystems:  [  OK  ]
Enabling local filesystem quotas:  [  OK  ]
Enabling /etc/fstab swaps:  [  OK  ]
Entering non-interactive startup
Calling the system activity data collector (sadc)...
Bringing up loopback interface:  [  OK  ]
Bringing up interface eth0:  Determining if ip address 90.0.12.151 is already in use for device eth0...
[  OK  ]
Bringing up interface eth1:  Determining if ip address 5.0.12.151 is already in use for device eth1...
[  OK  ]
Starting auditd: [  OK  ]
Starting system logger: [  OK  ]
Starting irqbalance: [  OK  ]
Starting cluster:
   Checking if cluster has been disabled at boot... [  OK  ]
   Checking Network Manager... [  OK  ]
   Global setup... [  OK  ]
   Loading kernel modules... [  OK  ]
   Mounting configfs... [  OK  ]
   Starting cman... [  OK  ]
   Starting qdiskd... [  OK  ]
   Waiting for quorum... [  OK  ]
   Starting fenced... [  OK  ]

   Tuning DLM kernel config... [  OK  ]
   Starting gfs_controld... [  OK  ]
   Unfencing self... [  OK  ]
   Joining fence domain... [  OK  ]
Starting system message bus: [  OK  ]
Starting clvmd: clvmd startup timed out
Mounting GFS2 filesystem (/opt/app/oracle/archivelog/db_backup): invalid device path "/dev/vg_arch/lv_db_backup"
[FAILED]
Mounting GFS2 filesystem (/opt/app/oracle/archivelog/local_arch): invalid device path "/dev/vg_arch/lv_local_arch"
[FAILED]
Mounting GFS2 filesystem (/opt/app/oracle/archivelog/standby_arch): invalid device path "/dev/vg_arch/lv_standby_arch"
[FAILED]
Starting HAL daemon: [  OK  ]
Starting sshd: [  OK  ]
Starting ntpd: [  OK  ]
Starting crond: [  OK  ]
Starting ohasd:
CRS-4123: Oracle High Availability Services has been started.
Starting Cluster Module - cluster monitor: Setting verbosity level to LogBasic
[  OK  ]
Starting Cluster Service Manager: [  OK  ]
Starting oddjobd: [  OK  ]
Starting ricci: [  OK  ]
Starting smartd: [  OK  ]

查询文件系统，rac两服务器节点确实是没有以“/dev/vg_arch/lv_”开头的设备的，也没有对应的“/dev/dm-”设备，而dataguar服务器上是有以“/dev/vg_arch/lv_”开头的设备的，也有对应的“/dev/dm-”设备。

草中宝

下面是dataguard服务器详细的启动日志
[root@dataguard log]# more boot.log
Gmount: mount point /proc/bus/usb does not exist
                Welcome to Oracle Linux Server
Starting udev: G[  OK  ]
Setting hostname dataguard.com:  [  OK  ]
Setting up Logical Volume Management:   Skipping clustered volume group vg_arch
  Skipping volume group vg_arch
[FAILED]                                             <= dataguard服务器 也有这个错误，但不影响dataguard服务器上的gfs2文件系统的挂载
Checking filesystems
/dev/sda2: clean, 95895/460560 files, 701644/1841152 blocks
/dev/sda1: clean, 44/76912 files, 73718/307200 blocks
/dev/sda5: clean, 50815/697632 files, 1841818/2786816 blocks
[  OK  ]
Remounting root filesystem in read-write mode:  [  OK  ]
Mounting local filesystems:  [  OK  ]
Enabling local filesystem quotas:  [  OK  ]
Enabling /etc/fstab swaps:  [  OK  ]
Entering non-interactive startup
Calling the system activity data collector (sadc)...
Bringing up loopback interface:  [  OK  ]
Bringing up interface eth0:  Determining if ip address 90.0.12.153 is already in use for device eth0...
[  OK  ]
Starting auditd: [  OK  ]
Starting system logger: [  OK  ]
Starting irqbalance: [  OK  ]
Starting cluster:
   Checking if cluster has been disabled at boot... [  OK  ]
   Checking Network Manager... [  OK  ]
   Global setup... [  OK  ]
   Loading kernel modules... [  OK  ]
   Mounting configfs... [  OK  ]
   Starting cman... [  OK  ]
   Starting qdiskd... [  OK  ]
   Waiting for quorum... [  OK  ]
   Starting fenced... [  OK  ]
   Starting dlm_controld... [  OK  ]
   Tuning DLM kernel config... [  OK  ]

   Unfencing self... [  OK  ]
   Joining fence domain... [  OK  ]
Starting system message bus: [  OK  ]
Starting clvmd:
Activating VG(s):   3 logical volume(s) in volume group "vg_arch" now active
[  OK  ]
Mounting GFS2 filesystem (/opt/app/oracle/oradata/orcl/archivelog/db_backup): [  OK  ]
Mounting GFS2 filesystem (/opt/app/oracle/oradata/orcl/archivelog/local_arch): [  OK  ]
Mounting GFS2 filesystem (/opt/app/oracle/oradata/orcl/archivelog/standby_arch): [  OK  ]
Starting HAL daemon: [  OK  ]
Starting sshd: [  OK  ]
Starting crond: [  OK  ]
Starting saslauthd: [  OK  ]
Start luci...
Point your web browser to https://dataguard.com:8084 (or equivalent) to access luci
Starting Cluster Module - cluster monitor: Setting verbosity level to LogBasic
[  OK  ]
Processing Database instance "orcl": log file /opt/app/oracle/product/11.2.0/db_1/startup.log
OK
Starting Cluster Service Manager: [  OK  ]
Starting oddjobd: [  OK  ]
Starting ricci: [  OK  ]
Starting smartd: [  OK  ]
[root@dataguard log]#

草中宝

比较过三台服务器的服务，要说可疑服务，dataguard服务器上就比rac服务器多了一个luci服务。但luci服务是管理cluster的啊。

此时再重启rac服务器，gfs2文件系统又可以正常挂载了。

实在是弄不明白了。还请各位老鸟多多指教。

草中宝

似乎是哪个服务器先启动，那个服务器上的gfs2文件系统就无法挂载。等所有的服务器都启动之后，再重启一次之前不能挂载gfs2文件系统的服务器，之前挂载失败的，也可以正常挂载了。

hiyachen

redhat吗？gfs2是redhat的。

草中宝

hiyachen 发表于 2015-08-13 18:48
redhat吗？gfs2是redhat的。

是redhat的衍生版本，oracle linux 6.5。
看了些文档，貌似要整个qdisk，并设置qdisk的票数。只是不知道怎么设置呢。