There is an error generated by a new rc script lvm2-monitor. From what I can tell, the script tries to run a vgchange --monitor -y on all of the volume groups. Because LVM2 is set up in cluster mode, instead of a list of volume groups vgs returns the error "WARNING: Falling back to local file-based locking. Volume Groups with the clustered attribute will be inaccessible." which the script feeds one word at a time to vgchange. Each word is fed as a volume group name, generating more error messages, including the first one because the cluster, clvmd, etc are not running.
The man page on vgchange says that --monitor controls where or not dmeventd monitors mirrored logical volumes. I do not have any mirrored volumes so I will probably ignore this message and deploy U5 on my production cluster.
The vgchange man page refers to mirror_image_fault_policy and mirror_log_fault_policy in lvm.conf(5). Unfortunately, lvm.conf(5) does not list mirror_image_fault_policy or mirror_log_fault_policy. There does not appear to be a man page for dmeventd, I am curious what dmeventd does.
Also, lvm2-monitor is invoked very early in the boot up cycle. If you are using iscsi, aoe, or gndb, lvm2-monitor will not see your storage because iscsi, gndb and aoe are started much later in the boot cycle.
It looks like there is an open bugzilla on this issue: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=239779 . If you are using mirrored logical volumes you might want to check into it, also you might also want to find out how to get lvm2-monitor to run after iscsi starts up.
Matt
On Wed, 2007-05-16 at 23:39 +0200, Bob Marcan wrote:
Hi. Are there any issues with the upgrade (up2date) RHEL, CS and GFS from v4.4 to v4.5? TIA, Bob
-- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
