[ClusterLabs] error external/libvirt device not accessible

Oscar Salvador osalvador.vilardaga at gmail.com
Mon Jun 15 09:36:50 UTC 2015


Hi,

Sorry, you can ignore the whole e-mail. I fixed that, was a
missconfiguration with libvirt primitive parameters.

Thanks
Oscar

2015-06-15 10:53 GMT+02:00 Oscar Salvador <osalvador.vilardaga at gmail.com>:

> Hi,
>
> I've seen I made a mistake with -INFINITY, I confuse something, so you
> can ignore the part I'm talking about -INFINITY.
> I saw I specified this with "-inf"
>
> Thanks
> Oscar
>
> 2015-06-15 10:43 GMT+02:00 Oscar Salvador <osalvador.vilardaga at gmail.com>:
>
>> Hi!
>>
>> I have a problem with fencing.
>> I'm using fence-agents, and a configured a one that uses libvirt device.
>>
>> Versions:
>>
>> fence-agents                   4.0.18
>> corosync                        2.3.4
>> pacemaker                      1.1.12
>>
>> Here is the configuration:
>>
>> node 1053402612: server01
>> node 1053402613: server02
>> primitive IP-rsc_nginx IPaddr2 \
>>         params ip=xx.xx.xx.xx nic=eth0 cidr_netmask=xx.xx.xx.xx \
>>         meta migration-threshold=2 \
>>         op monitor interval=20 timeout=60 on-fail=restart
>> primitive Nginx-rsc nginx \
>>         meta migration-threshold=2 \
>>         op monitor interval=20 timeout=60 on-fail=restart
>> primitive p_fence_server01 stonith:external/libvirt \
>>         params hostlist=server01
>> hypervisor_uri="qemu+tcp://virtnode01/system" \
>>         op monitor interval=60
>> primitive p_fence_server02 stonith:external/libvirt \
>>         params hostlist=server02
>> hypervisor_uri="qemu+tcp://virtnode02/system" \
>>         op monitor interval=60
>> location l_fence_server01 p_fence_server01 -inf: server01
>> location l_fence_server02 p_fence_server02 -inf: server02
>> colocation lb-loc inf: IP-rsc_nginx Nginx-rsc
>> order lb-ord inf: IP-rsc_nginx Nginx-rsc
>> property cib-bootstrap-options: \
>>         stonith-enabled=true \
>>         no-quorum-policy=ignore \
>>         default-resource-stickiness=100 \
>>         dc-version=1.1.12-561c4cf \
>>         cluster-infrastructure=corosync
>>
>>
>> Whe I commit the conf, I can see the followning in the logs:
>>
>> First:
>>
>> *Jun 15 10:16:05 [27835] server02   stonithd:     info:
>> cib_device_update:       Device p_fence_server02 has been disabled on
>> server02: score=-INFINITY*
>>
>> Since I specifed a location with:
>>
>> "*location l_fence_server02 p_fence_server02 -inf: server02*"
>>
>> Why it put an -INFINITY score? It is the default behaviour? As far as I
>> know, -INFINITY means "not run this resource on this node".
>>
>> Then I receive:
>>
>>
>> *Jun 15 10:16:05 [27835] server02   stonithd:     info:
>> xml_patch_version_check:         Current num_updates is too high (8 > 7)*
>>
>> *Jun 15 10:16:07 server02 stonith: [30410]: WARN: external_status:
>> 'libvirt status' failed with rc 1*
>> *Jun 15 10:16:07 server02 stonith: [30410]: ERROR: external/libvirt
>> device not accessible.*
>> *Jun 15 10:16:07 [27835] server02   stonithd:     info:
>> internal_stonith_action_execute:         Attempt 2 to execute fence_legacy
>> (monitor). remaining timeout is 19*
>> *Jun 15 10:16:09 server02 stonith: [30435]: WARN: external_status:
>> 'libvirt status' failed with rc 1*
>> *Jun 15 10:16:09 server02 stonith: [30435]: ERROR: external/libvirt
>> device not accessible.*
>>
>>
>> *Jun 15 10:16:09 [27835] server02   stonithd:     info:
>> update_remaining_timeout:        Attempted to execute agent fence_legacy
>> (monitor) the maximum number of times (2) allowed *
>> *Jun 15 10:16:09 [27835] server02   stonithd:   notice: log_operation:
>> Operation 'monitor' [30434] for device 'p_fence_server01' returned: -201
>> (Generic Pacemaker error)*
>> *Jun 15 10:16:09 [27835] server02   stonithd:  warning: log_operation:
>> p_fence_server01:30434 [ Performing: stonith -t external/libvirt -S ]*
>> *Jun 15 10:16:09 [27835] server02   stonithd:  warning: log_operation:
>> p_fence_server01:30434 [ failed:  1 ]*
>>
>>
>> One think it called my attention was:
>>
>> "Performing: stonith -t external/libvirt -S"
>>
>> If I execute this command, I receive:
>>
>> *# stonith -t external/libvirt -S*
>> *Must specify either -p option, -F option, -E option, or name=value style
>> arguments*
>> *usage:*
>> * stonith [-svh] -L*
>> * stonith [-svh] -t stonith-device-type -n*
>> * stonith [-svh] -t stonith-device-type -m*
>> * stonith [-svh] -t stonith-device-type {-p stonith-device-parameters |
>> -F stonith-device-parameters-file | -E | name=value...} [-c count] -lS*
>> * stonith [-svh] -t stonith-device-type {-p stonith-device-parameters |
>> -F stonith-device-parameters-file | -E | name=value...} [-c count] -T
>> {reset|on|off} nodename*
>>
>> *where:*
>> * -L list supported stonith device types*
>> * -l list hosts controlled by this stonith device*
>> * -S report stonith device status*
>> * -s silent*
>> * -v verbose*
>> * -n output the config names of stonith-device-parameters*
>> * -m display meta-data of the stonith device type*
>> * -h display detailed help message with stonith device description(s)*
>>
>>
>> Si, it seems that it is executed with some missing parameters.
>>
>> Anyone has an idea?
>>
>> Thank you very much in advance
>> Oscar
>>
>>
>>
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20150615/082aa20c/attachment.htm>


More information about the Users mailing list