4.14 Poll.sh Error on glusterfs images

Hi, I got a monitoring error on 4.14 after an upgrade from 4.12, on hosts that have VM’s with kvm and disks from a glusterfs datastore.

Seems like poll.sh cant retrieve disk usage and marks the host as unknown.

From oned.log

ERROR="Thu Oct 1 19:08:56 2015 : Error monitoring Host nebulaXXX (44): …/…/vmm/kvm/poll:368:in block in get_disk_usage': undefined methodmatch’ for nil:NilClass (NoMethodError)
from /usr/lib/ruby/2.1.0/rexml/element.rb:902:in block in each' from /usr/lib/ruby/2.1.0/rexml/xpath.rb:67:ineach’
from /usr/lib/ruby/2.1.0/rexml/xpath.rb:67:in each' from /usr/lib/ruby/2.1.0/rexml/element.rb:902:ineach’
from …/…/vmm/kvm/poll:329:in get_disk_usage' from ../../vmm/kvm/poll:145:inblock in get_all_vm_info’
from …/…/vmm/kvm/poll:129:in each' from ../../vmm/kvm/poll:129:inget_all_vm_info’
from …/…/vmm/kvm/poll:856:in print_all_vm_template' from ../../vmm/kvm/poll:908:in'
ERROR MESSAGE --8<------
Error executing poll.sh
ERROR MESSAGE ------>8–

On the host, from virsh i got this info on the disks

<disk type='network' device='disk'>
  <driver name='qemu' type='qcow2' cache='none'/>
  <source protocol='gluster' name='vg0/e0a60044c15d3b1725a657de6ed1f5c0'>
    <host name='nebulaXXX.example.com' port='24007'/>
  </source>

This works on other hosts we have, with nfs datastores, with this info from virsh

<disk type='file' device='disk'>
  <driver name='qemu' type='raw'/>
  <source file='/var/lib/one//datastores/0/528/disk.2'/>

There is a bug in the probe. Here is the issue tracking it:

http://dev.opennebula.org/issues/4020

Download the attached file from that ticket and copy it to /var/lib/one/remotes/kvm/poll. After that execute:

$ onehost sync --force

The probe disables disk size monitoring for drivers that do not use RDB (Ceph) or plain files.

1 Like

Thanks, it works.