On node A and B are in cluster.
AUTO_VG_ACTIVATE=1 --Non cluster
AUTO_VG_ACTIVATE=0 -- Cluster Environment.
Both nodes have
AUTO_VG_ACTIVATE=1 - this implies that ALL volume groups defined in the
/etc/lvmtab are not part of a high availability cluster, “Cluster Aware”, and
should be activated at boot (vgchange –a y /dev/vg_name).
This is not the
correct configuration for a volume group, that is “Cluster Aware”. The
Serviceguard package named “acms” uses /dev/vg04, /dev/vg04 has 27 logical
volumes. The volume group activation, file system checks (fsck) and subsequent
mounting/un-mounting of the file systems that reside on the logical volumes are
all controlled by serviceguard package control script.
Your current
configuration, at boot, tries to activate /dev/vg00, /dev/vg01 and
/dev/vg04.
/etc/lvmrc should
have AUTO_VG_ACTIVATE=0 and the following custom_vg_activation stanza modified
to include for non-Cluster Aware volume group activation, such as /dev/vg00 and
/dev/vg01 –
Original
custom_vg_activation stanza
________________________________________
#
custom_vg_activation()
{
# e.g.
/sbin/vgchange -a y -s
#
parallel_vg_sync "/dev/vg00 /dev/vg01"
#
parallel_vg_sync "/dev/vg02 /dev/vg03"
return
0
}
#
______________________________________
Modified
custom_vg_activation stanza
________________________________
#
custom_vg_activation()
{
# e.g.
/sbin/vgchange -a y -s
#
parallel_vg_sync "/dev/vg00 /dev/vg01"
#
parallel_vg_sync "/dev/vg02 /dev/vg03"
/sbin/vgchange -a y
/dev/vg00
/sbin/vgchange -a y /dev/vg01
return
0
}
#
____________________________________
The result of
making this change is that /dev/vg00 and /dev/vg01 will be available at system
boot, /dev/vg04 will not be available at system boot. The serviceguard package
control script will make /dev/vg04 available when the acms package is started
through serviceguard.
No comments:
Post a Comment