[one-users] OCA API / deleting user

[Piotr Kandziora]

Hi,


I have been performing tests with OpenNebula and noticed that using command
line I am not able to delete user while it has some VMs running (it is ok).
But, using OCA API this operation is possible and causes some strange
behaviour. Details below:

a) using command line (deleting user test11 is not possible, because it has
running one-4014 vm instance):

oneadmin at main:~$ oneuser list
  ID USER            PASSWORD
   0 oneadmin        4ff88aaddbd209d8026924c2cc2836b408698823
   3 test11          4ff88aaddbd209d8026924c2cc2836b408698823

oneadmin at main:~$ onevm list
   ID     USER     NAME STAT CPU     MEM        HOSTNAME        TIME
 4014   test11 one-4014 runn   0      0K          root24 00 00:00:32

oneadmin at main:~$ oneuser delete 3
The user 3 still has VMs defined, aborting user delete.

b) using OCA API (in my case Python OCA API):

So, I am deleting user 'test11' using following code:

#!/usr/bin/python
import oca

c = oca.Client('oneadmin:free','http://localhost:2633/RPC2')

usr = oca.user.UserPool(c)
usr.info(oca.ALL)

for i in usr:
    if i.name == 'test11':
        i.delete()

After it command 'onevm list' returns output:

oneadmin at main:~$ onevm list
   ID     USER     NAME STAT CPU     MEM        HOSTNAME        TIME
/usr/lib/one/ruby/client_utilities.rb:143:in `Float': invalid value for
Float(): "" (ArgumentError)
        from /usr/lib/one/ruby/client_utilities.rb:143:in `data_str'
        from /usr/lib/ruby/1.8/rubygems/custom_require.rb:31:in `collect'
        from /usr/lib/one/ruby/client_utilities.rb:139:in `each'
        from /usr/lib/one/ruby/client_utilities.rb:139:in `collect'
        from /usr/lib/one/ruby/client_utilities.rb:139:in `data_str'
        from /usr/lib/one/ruby/client_utilities.rb:138:in `collect'
        from /usr/lib/one/ruby/client_utilities.rb:138:in `data_str'
        from /usr/bin/onevm:208:in `list_short'
        from /usr/bin/onevm:682

At this moment I am not able to perform any action using onevm command.

In this situation I had to fix it by manual modifcation in vm_pool table in
opennebula database (I removed records that belonged to deleted user).

Question is following: Do I have to ensure before deleting user that it does
not have any running VM instances? Or this is unplanned behaviour and should
be fixed?

I am using OpenNebula version 2.0.1.

Best regards
Piotr Kandziora
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20110204/ef7cb3cc/attachment-0002.htm>