[Linux-cluster] LVM, Cluster, GFS, iSCSI problems

Brian spatuality at yahoo.ca
Wed Oct 11 17:38:00 UTC 2006


Hi group,

I am trying to setup GFS on two systems but there is a clvm / LVM problem. Here's what's happening:

RHEL 4.4 system x86_64 fully updated (except for latest .3 kernel as cluster dependencies are missing). Running cluster suite fully updated. Using 2 Dell PowerEdge 1955 Blade servers connected to a Promise m500i iSCSI disk array unit.

iSCSI is connecting okay to both servers. One server had LVM and GFS mounted properly and working but the other did not. What's happening now with both servers is the following:

# lvdisplay
    Logging initialised at Wed Oct 11 12:00:45 2006
    Set umask to 0077
    Loaded external locking library liblvm2clusterlock.so
    Finding all logical volumes
  --- Logical volume ---
  LV Name                /dev/vg_gfstest/lv_gfstest
  VG Name                vg_gfstest
  LV UUID                seA0f2-DBgZ-D0gX-gNtT-cT4n-dKFM-0XAdzL
  LV Write Access        read/write
  LV Status              NOT available
  LV Size                20.00 GB
  Current LE             5120
  Segments               1
  Allocation             inherit
  Read ahead sectors     0


I'm guessing that the "NOT Available" bit is causing the volume group to not show up in /dev/ :). This is bad. I'm not sure how to fix this.


Running vgchange to activate the volume group doesn't work:
# vgchange -a y
    Logging initialised at Wed Oct 11 13:24:21 2006
    Set umask to 0077
    Loaded external locking library liblvm2clusterlock.so
    Finding all volume groups
    Finding volume group "vg_gfstest"
  Error locking on node javaapps2: Internal lvm error, check syslog
  Error locking on node javaapps1: Internal lvm error, check syslog
  0 logical volume(s) in volume group "vg_gfstest" now active


The log in /var/log/messages:
Oct 11 13:24:23 server1 lvm[6407]: Volume group for uuid not found: yiH21FlyZwB03nTxylMnyNQsRGnvTqbMseA0f2DBgZD0gXgNtTcT4ndKFM0XAdzL


Has anyone seen this before, or know what to do with this? I suspect this is cluster related as I have not come across this on a stand alone system, but do correct me if I am wrong.

Brian









More information about the Linux-cluster mailing list