[Ovirt-devel] ovirt install : several issues

Fri Oct 9 16:29:04 UTC 2009

I installed ovirt-agent on a fresh fedora 11 install (not in a virtual machine). The only thing I modified prior to installation was /etc/resolv.conf in order to be able to yum update and install the ovirt packages via yum  (from the ovirt repo), and /etc/ntp.conf to get the correct time from the NTP server of my network. I also disabled NetworkManager via chkconfig.

I used all the options for ovirt-install exactly as recommended in the install guide in the homepage (yes for DNS, DHCP with PXE, cobbler)

I am using one admin network and one guest network.

Now the web interface is working, PXE booting is fine, however my nodes never appear on the web interface.

I have this error message every 5 sec in /var/log/ovirt-server/ovirt-agent.log  :
ERROR Fri Oct 09 18:02:46 +0200 2009 (2434) ********** get_query
DEBUG Fri Oct 09 18:02:46 +0200 2009 (2434) Query: context=338 class=agent object_id=
DEBUG Fri Oct 09 18:02:46 +0200 2009 (2434) User ID: 
ERROR Fri Oct 09 18:02:46 +0200 2009 (2434) Error in ovirt-agent: Unknown class agent
ERROR Fri Oct 09 18:02:46 +0200 2009 (2434)     /usr/share/ovirt-server/ovirt-agent/ovirt-agent.rb:201:in `get_query'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:282:in `do_agent_events'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:320:in `do_events'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:342:in `sess_event_recv'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:121:in `run'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:83:in `initialize'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:82:in `new'
    /usr/lib/ruby/site_ruby/1.8/qmf.rb:82:in `initialize'
    /usr/share/ovirt-server/ovirt-agent/ovirt-agent.rb:146:in `new'
    /usr/share/ovirt-server/ovirt-agent/ovirt-agent.rb:146:in `initialize'
    /usr/share/ovirt-server/ovirt-agent/ovirt-agent.rb:283:in `new'
    /usr/share/ovirt-server/ovirt-agent/ovirt-agent.rb:283

And in tail /var/log/krb5kdc.log 
Oct 09 17:50:34 management.ovirt.priv krb5kdc[1902](info): AS_REQ (7 etypes {18 17 16 23 1 3 2}) 192.168.50.1: ISSUE: authtime 1255103434, etypes {rep=18 tkt=18 ses=18}, qpidd/management.ovirt.priv at OVIRT.PRIV for krbtgt/OVIRT.PRIV at OVIRT.PRIV
Oct 09 17:50:48 management.ovirt.priv krb5kdc[1902](info): AS_REQ (7 etypes {18 17 16 23 1 3 2}) 192.168.50.1: NEEDED_PREAUTH: libvirt/management.ovirt.priv at OVIRT.PRIV for krbtgt/OVIRT.PRIV at OVIRT.PRIV, Additional pre-authentication required
Oct 09 17:50:48 management.ovirt.priv krb5kdc[1902](info): AS_REQ (7 etypes {18 17 16 23 1 3 2}) 192.168.50.1: ISSUE: authtime 1255103448, etypes {rep=18 tkt=18 ses=18}, libvirt/management.ovirt.priv at OVIRT.PRIV for krbtgt/OVIRT.PRIV at OVIRT.PRIV
Oct 09 17:50:53 management.ovirt.priv krb5kdc[1902](info): TGS_REQ (7 etypes {18 17 16 23 1 3 2}) 192.168.50.1: ISSUE: authtime 1255103448, etypes {rep=18 tkt=18 ses=18}, libvirt/management.ovirt.priv at OVIRT.PRIV for qpidd/management.ovirt.priv at OVIRT.PRIV

I can monitor the booting of my node via serial console and here are the relevant error messages :

First :
Bringing up interface eth0:  RTNETLINK answers: No such file or directory
[  OK  ]
Bringing up interface breth0:  
Determining IP information for breth0... failed.
[FAILED]

This one is I assume because there is no DHCP server on the guest network. It shouldn't be a problem since the hardware node should not be reachable by the guest network ? I plan to configure the IP of my guests statically.

Second :
ntpdate: Synchronizing with time server: [FAILED]
Starting ntpd: [  OK  ]

So the time does not synchronize with the management node. But I did put PEERNTP=yes in ifcfg-eth1 in my management node.

Third :
WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/.
WARNING: Could not open '/lib/modules/2.6.30.5-43.fc11.x86_64/kernel/fs/exportfs/exportfs.ko': No such file or directory
WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/.
WARNING: Could not open '/lib/modules/2.6.30.5-43.fc11.x86_64/kernel/fs/exportfs/exportfs.ko': No such file or directory
FATAL: Error inserting nfsd (/lib/modules/2.6.30.5-43.fc11.x86_64/kernel/fs/nfsd/nfsd.ko): Unknown symbol in module, or unknown parameter (see dmesg)
FATAL: Error running install command for nfsd
Starting NFS services:  exportfs: can't open /var/lib/nfs/rmtab for reading
exportfs: can't open /var/lib/nfs/xtab for reading

I assume this is because I do not have a NFS server in the management node running yet.

Last :
oVirt Node release 1.0.2 Beta (1.fc11)
Kernel 2.6.30.5-43.fc11.x86_64 on an x86_64 (/dev/ttyS0)

localhost.localdomain login: 

So the node did not get DNS information since it should be named node#.ovirt.priv, I assume.

Then I ran a wireshark on the management node eth1 because I was wondering which IP address it got.

I noticed that the DHCP exchange prior to TFTP transfer was giving address 192.168.50.21 to the node. However, I cannot ping this address once the node is booted. Since there is no eth1 configuration message in the boot sequence of the node, I assume that the linux kernel does not know about this IP. I think the DHCP exchange should happen a second time : once for the BIOS, once for the linux kernel.

This is all I can think about. I hope you can sort me out. Thanks in advance.