[one-users] database malformed

[Shi Jin]

Hi there,
I wanted to add some features to the oned.conf file, so I did a one stop and
then one start. However the start receives some message about the .lock.

Then I realized that something is wrong about the system. For example, onevm
list shows vms without host:
  27 oneadmin xubuntu- runn   0 1048576           node3 00 21:21:00
  28 oneadmin xubuntu- runn   0 1048576                 00 21:20:59
  29 oneadmin xubuntu- runn   0 1048576                 00 21:20:57
  30 oneadmin xubuntu- runn   0 1048576           node2 00 21:20:56
And newly deployed VMs are always in pending state.

Then I went to oned.log file and found error messages:
Fri Oct  2 14:49:00 2009 [ONE][E]: SQL command was: SELECT oid FROM vm_pool
WHERE state == 3 and ( lcm_state == 3 or lcm_state == 16 ), error: database
disk image is malformed

The sched.log file shows similar messages:
Fri Oct  2 14:50:24 2009 [ONE][E]: SQL command was: SELECT * FROM
vm_attributes WHERE id=35, error: database disk image is malformed
Fri Oct  2 14:49:34 2009 [HOST][D]: Discovered Hosts (enabled): 4
Fri Oct  2 14:49:34 2009 [ONE][E]: SQL command was: SELECT oid FROM vm_pool
WHERE state == 1, error: database disk image is malformed
Fri Oct  2 14:49:34 2009 [VM][D]: Pending virtual machines :

So I guess the SQLite Database was corrupted, most likely during my earlier
one stop/start process. I am wondering what is the best practice to restart
the OpenNebula service, without causing database corruption. Also, would it
be possible to simply replace the SQLite database with a MySQL database. I
guess the SQL commands will remain the same, right? Has anyone done this?

Thanks a lot.
-- 
Shi Jin, Ph.D.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20091002/c0da8860/attachment-0001.htm>