AW: AW: [Linux-cluster] Problems starting a VM Service
Rene Rosenberger
r.rosenberger at netbiscuits.com
Thu Oct 1 12:12:27 UTC 2009
Hi again,
here is the debug output when i try to start Log-Server:
[root at cluster-node01 tmp]# cat DEBUG
+ PATH=/bin:/sbin:/usr/bin:/usr/sbin
+ export PATH
++ dirname /usr/share/cluster/vm.sh
+ . /usr/share/cluster/ocf-shellfuncs
+++ basename /usr/share/cluster/vm.sh
++ __SCRIPT_NAME=vm.sh
++ consoletype
++ '[' 1 -eq 1 ']'
++ __SERIAL=yes
++ __LOG_PID=7780
++++ readlink /proc/7780/exe
+++ basename /usr/sbin/clurgmgrd
++ __LOG_NAME=clurgmgrd
++ __ocf_set_defaults stop
++ __OCF_ACTION=stop
++ unset LANG
++ LC_ALL=C
++ export LC_ALL
++ OCF_SUCCESS=0
++ OCF_ERR_GENERIC=1
++ OCF_ERR_ARGS=2
++ OCF_ERR_UNIMPLEMENTED=3
++ OCF_ERR_PERM=4
++ OCF_ERR_INSTALLED=5
++ OCF_ERR_CONFIGURED=6
++ OCF_NOT_RUNNING=7
++ '[' -z vm ']'
++ '[' -z 1 ']'
++ '[' -z /usr/share/cluster ']'
++ '[' '!' -d /usr/share/cluster ']'
++ '[' x1 '!=' x1 ']'
++ '[' -z 0 ']'
++ '[' xstop = xmeta-data ']'
++ '[' -z vm:Log-Server ']'
+ export OCF_APP_ERR_INDETERMINATE=150
+ OCF_APP_ERR_INDETERMINATE=150
+ case $1 in
+ validate_all
++ id -u
+ '[' 0 = 0 ']'
+ '[' -z auto ']'
+ '[' auto = auto ']'
++ virsh version
++ grep 'Running hypervisor:'
++ tr A-Z a-z
++ awk '{print $3}'
+ export OCF_RESKEY_hypervisor=xen
+ OCF_RESKEY_hypervisor=xen
+ '[' -z xen ']'
+ echo Hypervisor: xen
Hypervisor: xen
+ '[' 1 = 0 ']'
+ '[' -z '' ']'
+ echo 'Management tool: virsh'
Management tool: virsh
+ export OCF_RESKEY_use_virsh=1
+ OCF_RESKEY_use_virsh=1
+ '[' -z auto -o auto = auto ']'
+ '[' 1 = 1 ']'
+ '[' xen = qemu ']'
+ '[' xen = xen ']'
+ OCF_RESKEY_hypervisor_uri=xen:///
+ echo Hypervisor URI: xen:///
Hypervisor URI: xen:///
+ '[' -z auto -o auto = auto ']'
+ '[' 1 = 1 ']'
+ '[' xen = qemu ']'
+ '[' xen = xen ']'
+ export OCF_RESKEY_migration_uri=xenmigr://%s/
+ OCF_RESKEY_migration_uri=xenmigr://%s/
+ '[' -n xenmigr://%s/ ']'
++ printf xenmigr://%s/ target_host
+ echo Migration URI format: xenmigr://target_host/
Migration URI format: xenmigr://target_host/
+ '[' -z Log-Server ']'
+ return 0
+ do_stop shutdown destroy
+ declare domstate rv
++ do_status
++ '[' 1 = 1 ']'
++ virsh_status
++ declare state pid
++ '[' xen = xen ']'
++ service xend status
++ '[' 0 -ne 0 ']'
+++ pidof libvirtd
++ pid=7002
++ '[' -z 7002 ']'
+++ virsh domstate Log-Server
++ state='shut off'
++ echo shut off
++ '[' 'shut off' = running ']'
++ '[' 'shut off' = paused ']'
++ '[' 'shut off' = 'no state' ']'
++ '[' 'shut off' = idle ']'
++ return 1
++ return 1
+ domstate='shut off'
+ rv=1
+ ocf_log debug 'Virtual machine Log-Server is shut off'
+ '[' 2 -lt 2 ']'
+ declare __OCF_PRIO=debug
+ declare -i __OCF_PRIO_N
+ shift
+ declare '__OCF_MSG=Virtual machine Log-Server is shut off'
+ case "${__OCF_PRIO}" in
+ __OCF_PRIO_N=7
+ pretty_echo debug 'Virtual machine Log-Server is shut off'
+ declare pretty
+ declare 'n='
+ declare __OCF_PRIO=debug
+ shift
+ declare '__OCF_MSG=Virtual machine Log-Server is shut off'
+ '[' -n yes ']'
+ echo '<debug> Virtual machine Log-Server is shut off'
<debug> Virtual machine Log-Server is shut off
+ return 0
++ which clulog
+ '[' -z /usr/sbin/clulog ']'
+ clulog -p 7780 -n clurgmgrd -s 7 'Virtual machine Log-Server is shut off'
+ '[' 1 -eq 150 ']'
+ '[' 1 = 1 ']'
+ do_virsh_stop shutdown destroy
+ declare -i timeout=60
+ declare -i ret=1
+ declare state
++ do_status
++ '[' 1 = 1 ']'
++ virsh_status
++ declare state pid
++ '[' xen = xen ']'
++ service xend status
++ '[' 0 -ne 0 ']'
+++ pidof libvirtd
++ pid=7002
++ '[' -z 7002 ']'
+++ virsh domstate Log-Server
++ state='shut off'
++ echo shut off
++ '[' 'shut off' = running ']'
++ '[' 'shut off' = paused ']'
++ '[' 'shut off' = 'no state' ']'
++ '[' 'shut off' = idle ']'
++ return 1
++ return 1
+ state='shut off'
+ '[' 1 -eq 0 ']'
+ return 0
+ return 0
+ exit 0
Please, it is very important to get this running as it should!
Regards,
rene
-----Ursprüngliche Nachricht-----
Von: Lon Hohberger [mailto:lhh at redhat.com]
Gesendet: Mittwoch, 30. September 2009 17:08
An: r.rosenberger at netbiscuits.com
Cc: 'linux clustering'
Betreff: Re: AW: [Linux-cluster] Problems starting a VM Service
On Wed, 2009-09-30 at 15:25 +0200, Rene Rosenberger wrote:
> Hi,
>
> rgmanager-2.0.52-1
>
> [root at cluster-node02 ~]# cat /etc/cluster/cluster.conf
> <?xml version="1.0"?>
> <cluster alias="cluster01" config_version="16" name="cluster01">
> <fence_daemon clean_start="0" post_fail_delay="0"
> post_join_delay="3"/>
> <clusternodes>
> <clusternode name="cluster-node01.netbiscuits.com"
> nodeid="1" votes="1">
> <fence>
> <method name="1">
> <device name="Fence_Device_01"/>
> </method>
> </fence>
> </clusternode>
> <clusternode name="cluster-node02.netbiscuits.com"
> nodeid="2" votes="1">
> <fence>
> <method name="1">
> <device name="Fence_Device_02"/>
> </method>
> </fence>
> </clusternode>
> </clusternodes>
> <cman expected_votes="1" two_node="1"/>
> <fencedevices>
> <fencedevice agent="fence_ipmilan"
ipaddr="192.168.100.141"
> login="root" name="Fence_Device_01" passwd="emoveo11wap"/>
> <fencedevice agent="fence_ipmilan"
ipaddr="192.168.100.142"
> login="root" name="Fence_Device_02" passwd="emoveo11wap"/>
> </fencedevices>
> <rm>
> <failoverdomains>
> <failoverdomain name="Failover_Domain_01"
> nofailback="0" ordered="0" restricted="0">
> <failoverdomainnode
> name="cluster-node01.netbiscuits.com" priority="1"/>
> <failoverdomainnode
> name="cluster-node02.netbiscuits.com" priority="1"/>
> </failoverdomain>
> </failoverdomains>
> <resources/>
> <vm autostart="1" domain="Failover_Domain_01"
exclusive="0"
> migrate="live" name="Nagios" path="/rootfs/vm/" recovery="relocate"/>
> <vm autostart="1" domain="Failover_Domain_01"
exclusive="0"
> migrate="live" name="Log-Server" path="/rootfs/vm/" recovery="relocate"/>
> </rm>
> </cluster>
>
> Regards, rene
Ok, so it's not the one fixed here:
http://git.fedorahosted.org/git/?p=cluster.git;a=commit;h=a9ac1e635c559b4651
2cf4251fe71c015bb6d70f
I don't recall if this will matter much, but get rid of the trailing
slash on /rootfs/vm/.
Also, ensure /rootfs/vm/Nagios and /root/fs/vm/Log-Server file names
match the names contained within the respective config files.
(e.g. name = "Nagios" / name = "Log-Server" )
This is because rgmanager wants a vm "name" but xm wants a "config file"
- so they have to match.
-- Lon
More information about the Linux-cluster
mailing list