Has anyone successfully run openSUSE 12.3 configured with a openais/pacemaker cluster along with using clvm?
Upgraded a two node pacemaker HA configuration from openSUSE 12.2 to 12.3. Previously the 12.2 configuration has been running fine for several months.
Am having trouble getting any cluster enabled logical volumes to start the logival volume or trying to create one. It basically fails the same way for any existing logical volume activation or logical volume creation with invalid argument.
admin01:~ # lvcreate --name test1lv -L 3G vgtest01
Error locking on node c01010a: Invalid argument
Error locking on node d01010a: Invalid argument
Failed to activate new LV.
Decided to see if this was an upgrade issue or not, so stood up a temporary pacemaker cluster and did a clean install of an openSUSE 12.3 operating system. Unfortunately am continuing to see the exact same result from this clean installation.
Some basics;
wilma:~ # rpm -qa | egrep ‘clvm|cluster-glue|corosync|crmsh|dlm|libglue2|openais|pacemaker|resource-agent’ | sort
cluster-glue-1.0.11-2.1.1.x86_64
corosync-1.4.3-4.1.1.x86_64
crmsh-1.2.4-3.1.1.x86_64
libcorosync4-1.4.3-4.1.1.x86_64
libdlm-3.00.01-25.5.1.x86_64
libdlm-devel-3.00.01-25.5.1.x86_64
libdlm3-3.00.01-25.5.1.x86_64
libglue2-1.0.11-2.1.1.x86_64
libopenais3-1.1.4-15.1.1.x86_64
libpacemaker3-1.1.7-3.1.1.x86_64
lvm2-clvm-2.02.98-20.2.1.x86_64
openais-1.1.4-15.1.1.x86_64
pacemaker-1.1.7-3.1.1.x86_64
resource-agents-3.9.5-2.4.1.x86_64
in /etc/lvm/lvm.conf locking_type is set to 3.
I am starting dlm, o2cb and clvm via a cloned resource group, base-clone;
============
Last updated: Wed Jul 24 15:02:52 2013
Last change: Wed Jul 24 11:57:58 2013 by root via cibadmin on wilma
Stack: openais
Current DC: fred - partition with quorum
Version: 1.1.7-61a079313275f3e9d0e85671f62c721d32ce3563
2 Nodes configured, 2 expected votes
8 Resources configured.
Online: wilma fred ]
Full list of resources:
st-fred (stonith:external/esxi_free): Started wilma
st-wilma (stonith:external/esxi_free): Started fred
Clone Set: base-clone [base-group]
Started: fred wilma ]
wilma:~ # crm configure show base-group
group base-group dlm o2cb clvm
wilma:~ # crm configure show clvm
primitive clvm ocf:lvm2:clvmd
wilma:~ # pvs
PV VG Fmt Attr PSize PFree
/dev/sda2 vg00 lvm2 a-- 11.80g 1.30g
/dev/sdb vgtest01 lvm2 a-- 4.00g 4.00g
wilma:~ # vgs
VG #PV #LV #SN Attr VSize VFree
vg00 1 2 0 wz–n- 11.80g 1.30g
vgtest01 1 0 0 wz–nc 4.00g 4.00g
wilma:~ # lvs
LV VG Attr LSize Pool Origin Data% Move Log Copy% Convert
rootlv vg00 -wi-ao— 9.00g
swaplv vg00 -wi-ao— 1.50g
wilma:~ # lvcreate --name apachelv -L 3G vgtest01
Error locking on node 2d01010a: Invalid argument
Error locking on node 2e01010a: Invalid argument
Failed to activate new LV.
From messages, it appears there is something amiss between dlm and clvm based on this error reported from dlm.
2013-07-24T14:51:24.986528-04:00 wilma lvm[2652]: dlm_ls_lock returned 22
2013-07-24T14:51:24.986905-04:00 wilma lvm[2652]: hold_lock. lock at 1 failed: Invalid argument
2013-07-24T14:51:24.987288-04:00 wilma lvm[2652]: Command return is 22, critical_section is 0