Alon, the diff between the two hardware.py files you sent is: $ diff hardware.py_first hardware.py_second 108c108 < ret = struct.unpack('L', f.read(8))[0] --- so they seem quite the same.... And in engine.log it complains about 2013-01-31 17:49:57,189 ERROR [org.ovirt.engine.core.bll.InstallerMessages] (VdsDeploy) Installation 192.168.1.103: Internal error: invalid syntax (hardware.py, line 109) while diff between patched one and original file gives, around the line 109: diff hardware.py hardware.py.orig 105a90 107,113c92,94 < f.seek(index) < ret = struct.unpack('L', f.read(8))[0] < except struct.error: <---- seems this line... < pass < < self.logger.debug('prdmsr: %s', ret) < return ret --- 121c102 < ret = ( --- I don't know Python so I can't correct the typo.... let me know.

On Thu, Jan 31, 2013 at 9:50 PM, Alon Bar-Lev wrote: Actually I think it's me who has to sorry about that... let me explain All my tests during these months on different environments were made without DNS because I had no chance to manage it on the lan where I was. For this test insead I had opportunity to make things well and so I configured a CentoS 6.3 vm on another server as DNS server and IPA server for the environment. So I configured my DNS this way and it works ok: ipacentos 192.168.1.103 ---> dns server + IPA server f18ovn03 192.168.1.102 ---> node to install f18engine 192.168.1.111 ---> engine So far so good, DNS ok. But when this morning I tried to install the host, my input in web admin portal was: Name ---> f18ovn03.localdomain.local ip ---> 192.168.1.103 (that is WRONG! It is the one of the CentOS 6.3 vm, running on a R310 that has indeed an Intel cpu...) In fact [root@f18engine ~]# cd /var/log/ovirt-engine/host-deploy/ [root@f18engine host-deploy]# ll total 112 -rw-r--r--. 1 ovirt ovirt 110947 Jan 31 12:17 ovirt-20130131121728-192.168.1.103.log se the .103 in filename Unfortunately I realized that only this evening ;-( I tried to remove and install the node with correct ip value from the gui but I always got ssh connection interrupted messages... So I decided to restart clean with engine-cleanup and reboot of both servers. Probably an uuid for the hostname and/or the ip during the initial wrong deploy was causing inability to proceed.... So I reinstalled the beta versions for host-deploy and host-deploy-java rpm on engine And I was able to correctly install the node.... and after reboot it is now up and I can proceed Very very sorry to have let you waste your time ;-( But I managed to learn one thing and to ask you another one: 1) never put the same password on different servers ;-) 2) the "Name" field when adding a node should match its hostname or is meant only to be a label? In the first case, I think it could be a safe thing to verify hostname lookup and if it doesn' t match the ip provided, send a warning window to the user, because strange things could happen In the second case, probably if I removed the host from webadmin and tried to install with a different label (Name) and correct ip (.102) would have been successful in your opinion, without reconfiguring all the engine? Gianluca