HP-UX HB v13.00 Ch-15 - Serviceguard

HP-UX Handbook – Rev 13.00 Page 72 (of 108)
Chapter 15 Serviceguard
October 29, 2013
contains a PVID ≠ 0.
Example 6
Error: Unable to recv initialize VG message to %s: %s
Symptom: cmcheckconf/cmapplyconf hangs or needs very long to complete.
When performing cmcheckconf/cmapplyconf a cmclconfd helper process is launched on each
node for gathering configuration information. Most likely reasons for hangs or long runtimes are
cmclconfd processes being blocked while trying to access disk devices. By default cmclconfd
expects every claimed disk that is visible in ioscan to be readable without problems. This
assumption is not true in some cases, especially for devices associated to physical data
replication (ContinousAccess/XP, BusinessCopy/XP, EMC SRDF, EMC Timefinder). Making
the devices accessible (e.g. by splitting pairs) should solve related problems.
Use e.g. ps -el | grep cmclconfd on every cluster node. The cmclconfd's Open Files screen
in GlancePlus or the public domain tool lsof should show you what device we are waiting for.
Another workaround is to temporarily remove all VOLUME_GROUP statements (except for
cluster lock volume groups) form the cluster ASCII file (comment them out by prepending a '#'
sign) and use the -K option with cmcheckconf or cmapplyconf.
Example 7
Error: Found two volume groups with the same name vgA but different ids
Warning: The volume group vgA is activated on more than one node
This problem typically occurs when Serviceguard considers "privat" volumes groups to be
"shared", which happens often to the root volume group /dev/vg00. Serviceguard uses LVM's
VGID as unique differentiator to find shared volume groups. In general there are two scenarios
why Serviceguard may get misleading information about volume groups, complaining about
their configuration:
 The systems were cloned, meaning that their private disks were copied using tools like
dd(1). This causes the VGIDs to be identical in their LVM headers, although they should
be unique.
 The VGID in the /etc/lvmtab file does not match the VGID of attached disks. Copying
the lvmtab file to another system might cause this problem; vgscan(1M) can be used to
fix the file.
Example 8
Symptom: Some non-shared LVM volume groups are not automatically activated during bootup.
If the admin customizes the /sbin/lvmrc script to avoid activating cluster-aware VGs at boot
time (to avoid sending error messages like those in example 2 to the console), the admin may
forget to list non-shared volume groups in the file, particularly if adding some later.