[one-users] Monitoring issue right after host creation [No Monitoring Information | OpenNebula 3.2.0]
Hector Sanjuan
hsanjuan at opennebula.org
Fri Mar 9 09:58:34 PST 2012
Hello, your monitoring looks fine.
To get the monitoring information plots in sunstone you need to start the
oneacctd daemon:
http://opennebula.org/documentation:rel3.2:sunstone#monitoring_and_plots
About your libvirtd warning, it think it's not critical. In my box I have
the same warning but i was able to launch VMs with KVM.
Hope it helps,
Hector
En Fri, 09 Mar 2012 17:55:38 +0100, biro lehel <lehel.biro at yahoo.com>
escribió:
> Hello,
>
> I managed to resolve the issues regarding the monitoring, by following
> the instructions at [1] (sections 7.1.1.1. and 7.1.1.2), and at [2] (the
> section regarding PolicyKit). Now, 'onehost show 6' results in the
> following (see below).
>
> As you can see, although the monitoring info returned the correct
> result, it still contains an error message, which is stuck there from an
> earlier time (the timestamp of the error message is not the actual time
> of the monitoring, it shows the timestamp of an earlier unsuccessful
> monitoring try). The oned.log contains no errors. Also, I suppose
> because of this error message (which also appears at the bottom of the
> "Host Template" tab, with my current Host selected in Sunstone), the
> "Monitoring Information" tab shows "There is no monitoring information
> for
> host 6" both at CPU and Memory info. Any idea how to work this around?
>
> Also, my initial socket problem (I copy the following from messages.log):
> Mar 9 10:10:57 pcls libvirtd[5603]: Starting libvirtd WARNING: no
> socket to connect to
> Mar 9 10:10:57 pcls libvirtd[5603]: ..done
>
> , still persist. This message also appears when trying to run anything
> related to 'virsh' on my host,, for ex:
>
>> virsh -c qemu:///system list
> WARNING: no socket to connect to
> Id Name State
> ----------------------------------
>
> , and also, it appears when running run_probes on my host (although
> after the warning, it returns the correct information).
>
> It is kind of bothering me, and I could find the reason of it so far. It
> is also making it hard to trace the fact that besides this warning
> message, no errors messages are issued nowhere, and I couldn't trace it
> in the logs neither, so far. Any ideas what causes this socket warning?
> Please help.
>
> All the best,
> Lehel.
>
>
>
>> onehost show 6
> HOST 6
> INFORMATION
> ID : 6 NAME :
> pcls STATE : MONITORED
> IM_MAD : im_kvm VM_MAD :
> vmm_kvm VN_MAD : dummy
> TM_MAD :
> tm_shared LAST MONITORING TIME : 1331310160
> HOST
> SHARES
> MAX MEM : 16446748 USED MEM (REAL) :
> 773580 USED MEM (ALLOCATED) :
> 0 MAX CPU : 400 USED CPU
> (REAL) : 42 USED CPU (ALLOCATED) :
> 0 MAX DISK : 0 USED DISK
> (REAL) :
> 0 USED DISK (ALLOCATED) : 0 RUNNING
> VMS : 0
> MONITORING
> INFORMATION
> ARCH=x86_64
> CPUSPEED=1600
> ERROR=[
> MESSAGE="Error monitoring host 6 : MONITOR FAILURE 6 Error executing
> kvm.rb
> ",
> TIMESTAMP="Fri Mar 9 12:12:32 2012" ]
> FREECPU=357.2
> FREEMEMORY=15673168
> HOSTNAME=pcls
> HYPERVISOR=kvm
> MODELNAME="Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz"
> NETRX=0
> NETTX=0
> TOTALCPU=400
> TOTALMEMORY=16446748
> USEDCPU=42.8
> USEDMEMORY=773580
>
>
> [1]
> http://doc.opensuse.org/documentation/html/openSUSE/opensuse-kvm/cha.libvirt.connect.html#sec.libvirt.connect.auth
> [2] http://en.opensuse.org/SDB:Cloud_OpenNebula#Setting_up_a_cloud_node
>
>
> --- On Fri, 3/9/12, biro lehel <lehel.biro at yahoo.com> wrote:
>
> From: biro lehel <lehel.biro at yahoo.com>
> Subject: Re: [one-users]
> Monitoring issue right after host creation [No Monitoring Information |
> OpenNebula 3.2.0]
> To: users at lists.opennebula.org
> Date: Friday, March 9, 2012, 1:00 PM
>
> Hi Ruben,
>
> I am using the OpenSUSE package (installed through
> YaST). Just to make sure that I did the configurations right (I'm sure I
> didn't):
>
> At (I quote from the documentation): "To be able to use
> the images copied by OpenNebula, change also the user and group under
> which the libvirtd is run to “oneadmin”", what exactly needs to be done?
> I am a Linux-noob (or at least, not at all an expert), so forget me
> this question, but how do I set the user and group under which a daemon
> is run? I have checked under which user is the libvirtd daemon is
> currently running, and indeed it is root (since I probably set it and
> put it in init.d as the root user).
>
> Also: "(typically <oneadmin>) has to pertain to the <libvirtd> and
> <kvm> groups in order to use the daemon and be able to launch VMs."
> What are the
> <libvirtd> and <kvm> groups? And how how do I achieve the necessary
> configuration?
>
> I did some configuration, but I just want to make sure that I'm not
> missing anything (which I probably do).
>
> Ruben? Anyone?
>
> Thank you,
> Lehel.
>
> --- On Fri, 3/9/12, Ruben S. Montero <rsmontero at opennebula.org> wrote:
>
> From: Ruben S. Montero <rsmontero at opennebula.org>
> Subject: Re: [one-users] Monitoring issue right after host creation [No
> Monitoring Information | OpenNebula 3.2.0]
> To: "biro lehel" <lehel.biro at yahoo.com>
> Cc: users at lists.opennebula.org
> Date: Friday, March 9, 2012, 11:42 AM
>
> Hi
> It seems that a libvirt problem, either is not running at pcls or it is
> misconfigured. You have a couple of hints about configuring libvirt for
> KVM at
> [1].
> My other concern is your permissions problem. Are you using the OpenSUSE
> package or installing from source?
>
>
> [1] http://www.opennebula.org/documentation:rel3.2:kvmg#configuration
> Thanks
>
>
> Ruben
>
> On Fri, Mar 9, 2012 at 7:58 AM, biro lehel <lehel.biro at yahoo.com> wrote:
>
>
>
>
> Hello,
>
> I checked your suggestions, and although the scripts existed inside
> kvm.d, I didn't have execution permissions on them. I also run 'onehost
> sync'.
>
> However, now I get a monitoring error while running 'onehost show pcls',
> oned.log containing the following:
>
>
>
> Fri Mar 9 08:45:45 2012 [InM][I]: Monitoring host pcls (6)
> Fri Mar 9 08:45:50 2012 [InM][I]: Command execution fail: 'if [ -x
> "/var/tmp/one/im/run_probes" ]; then /var/tmp/one/im/run_probes kvm 6
> pcls;
> else exit 42; fi'
> Fri Mar 9 08:45:50 2012 [InM][I]: WARNING: no socket to connect to
> Fri Mar 9 08:45:50 2012
> [InM][I]: error: authentication failed: authentication failed
> Fri Mar 9 08:45:50 2012 [InM][I]: error: failed to connect to the
> hypervisor
> Fri Mar 9 08:45:50 2012 [InM][E]: Error executing kvm.rb
> Fri Mar 9 08:45:50 2012 [InM][I]: ExitCode: 255
>
>
> Fri Mar 9 08:45:50 2012 [InM][E]: Error monitoring host 6 : MONITOR
> FAILURE 6 Error executing kvm.rb
>
> On the other hand, running run_probes on pcls result in the same
> warning, but also outputs the expected data:
>
>
>
> # /var/tmp/one/im/run_probes kvm 6 pcls
> WARNING: no socket to connect to
> ARCH=x86_64 MODELNAME="Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz"
> HYPERVISOR=kvm TOTALCPU=400 CPUSPEED=3301 TOTALMEMORY=16446744
> USEDMEMORY=724860 FREEMEMORY=15721884 FREECPU=381.6 USEDCPU=18.4 NETRX=0
> NETTX=0 HOSTNAME=pcls
>
>
>
> I did check the documentation at [1], the only thing missing is that I
> didn't perform the following line:
>
> /etc/default/libvirt-bin : add -l option to libvirtd_opts
> , since I didn't have the mentioned file on my machine (I'm running
> OpenNebula 3.2.0 on an openSUSE 12.1 machine).
>
>
>
> I was curious, and I run 'virsh -c qemu' command, resulting in the
> following:
>
> # virsh -c qemu
> WARNING: no socket to connect to
> error: internal error unexpected Xen URI path 'qemu', try
> ///var/lib/xen/xend-socket
>
>
> error: failed to connect to the hypervisor
>
> Since I want to use KVM, and not Xen, as my hypervisor for OpenNebula, I
> smell a problem here. What am I missing? Any help?
>
> [1]: http://opennebula.org/documentation:rel3.2:kvmg
>
>
>
>
>
> --- On Thu, 3/8/12, Ruben S. Montero <rsmontero at opennebula.org> wrote:
>
>
>
> From: Ruben S. Montero <rsmontero at opennebula.org>
> Subject: Re: [one-users] Monitoring issue right after host creation [No
> Monitoring Information | OpenNebula 3.2.0]
>
>
> To: "biro lehel" <lehel.biro at yahoo.com>
> Cc:
> users at lists.opennebula.org
> Date: Thursday, March 8, 2012, 7:04 PM
>
> Yes,
> I mean the contents of the directory. run_probes, just executes the
> scripts in /var/tmp/one/im/kvm.d (in your case). The reason for no
> output of run_probes is an empty kvm.d. Could you check the contents of
> that directory (and execute any script there....)?
>
>
>
> BTW run_probes is a very simple shell script (similar to run_parts) you
> may want to execute it with bash -xv /var/tmp/one/im/run_probes kvm 4
> pcls
> On Thu, Mar 8, 2012 at 5:57 PM, biro lehel <lehel.biro at yahoo.com> wrote:
>
>
>
>
> Hello Ruben,
>
>
>
> as I mentioned at the bottom of my previous email, I already run
> /var/tmp/one/im/run_probes with the necessary parameters on pcls (my
> only host), which resulted in the same (empty) output. The file exists,
> it just doesn't produce the expected (or any) output, and I couldn't
> find the reason so far.
>
>
>
>
>
> I didn't try 'onehost sync', but I could try tomorrow morning (since I
> got home already, and I'm not at the respective computers), if you
> consider that it might help.
>
> Any other suggestions?
>
>
>
>
>
> Cheers,
> Lehel.
>
> --- On Thu, 3/8/12, Ruben S. Montero <rsmontero at opennebula.org> wrote:
>
>
>
>
>
> From: Ruben S. Montero <rsmontero at opennebula.org>
> Subject: Re: [one-users] Monitoring issue right after host
> creation [No Monitoring Information | OpenNebula 3.2.0]
> To: "biro lehel" <lehel.biro at yahoo.com>
> Cc: users at lists.opennebula.org
>
>
> Date: Thursday, March 8, 2012, 6:17 PM
>
> Hi,
> Could you please check the contents of /var/tmp/one/im/ at pcls? It
> seems that there are no probes there. If that's the case you can copy
> them by hand or use onehost sync and wait for a monitoring action on the
> host.
>
>
>
>
>
>
> Cheers
> Ruben
>
> On Thu, Mar 8, 2012 at 4:25 PM, biro lehel <lehel.biro at yahoo.com> wrote:
>
>
>
>
>
>
>
>
>
>
>
>
> Hello,
>
> I tried to add my one and only host, by executing on the front-end:
>
>
>
>
>
>
>
>> onehost create pcls im_kvm vmm_kvm tm_shared dummy
>
> Issuing the 'onehost list' command, the result seems to be the one
> expected. However, issuing 'onehost show pcls' command results in zero
> values and NO monitoring information:
>
>
>
>
>
>
>
> oneadmin at linux-4dsu:~/remotes/im> onehost show pcls
> HOST 6
> INFORMATION
> ID : 6 NAME :
> pcls STATE :
> MONITORED IM_MAD : im_kvm
> VM_MAD : vmm_kvm VN_MAD :
> dummy TM_MAD : tm_shared LAST
> MONITORING TIME : 1331219173
>
>
>
>
>
>
> HOST
> SHARES
> MAX MEM : 0 USED MEM (REAL) :
> 0 USED MEM (ALLOCATED) : 0
>
>
>
>
>
> MAX
> CPU : 0 USED CPU (REAL) :
> 0 USED CPU (ALLOCATED) : 0 MAX
> DISK : 0 USED DISK (REAL) :
> 0
>
>
>
>
>
> USED DISK (ALLOCATED) :
> 0 RUNNING VMS : 0
> MONITORING INFORMATION
> (the report ends here, without monitoring information).
>
> I had some errors before, oned.log having:
>
>
>
>
>
>
>
> Thu Mar 8 15:49:03 2012 [InM][I]: Command execution fail: 'if [ -x
> "/var/tmp/one/im/run_probes" ]; then /var/tmp/one/im/run_probes kvm 4
> pcls; else exit 42; fi'
> Thu Mar 8 15:49:03 2012 [InM][I]: ExitCode: 42
>
>
>
>
>
>
> Thu Mar 8 15:49:03 2012 [InM][E]: Error monitoring host 4 : MONITOR
> FAILURE 4 -
>
> However, I
> fixed this by simply adding execution permission to the run_probes file
> mentioned above (since, it didn't have one). At this moment, oned.log
> contains no errors (to be precise, it says,
>
> Thu Mar 8 17:16:12 2012 [ReM][D]: HostPoolInfo method invoked
>
>
>
>
>
>
> Thu Mar 8 17:16:12 2012 [ReM][D]: VirtualMachinePoolInfo method invoked
> Thu Mar 8 17:16:12 2012 [ReM][D]: AclInfo method invoked
> Thu Mar 8 17:16:13 2012 [InM][I]: Monitoring host pcls (6)
> Thu Mar 8 17:16:18 2012 [InM][I]: ExitCode: 0
>
>
>
>
>
>
> Thu Mar 8 17:16:18 2012 [InM][D]: Host 6 successfully monitored.
> Thu Mar 8 17:16:19 2012 [ReM][D]: UserPoolInfo method invoked
> Thu Mar 8 17:16:19 2012 [AuM][D]: Message received: LOG I 874 ExitCode:
> 0
>
> ).
>
>
>
>
>
>
>
> Firing the respective command on the host ( /var/tmp/one/im/run_probes
> kvm 4 pcls ) also results in no output and no errors (it simply puts a
> space on a new line in the CLI).
>
> I don't
> think that this is the expected behavior and I think that there is a
> problem somewhere. What am I missing? Any suggestions?
>
> Thanks,
> Lehel.
>
>
>
>
>
>
>
> _______________________________________________
>
> Users mailing list
>
> Users at lists.opennebula.org
>
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>
>
>
--
Hector Sanjuan
OpenNebula Developer
More information about the Users
mailing list