Kern Sibbald wrote:
> On Sunday 10 January 2010 14:00:22 Dan Langille wrote:
>> Kern Sibbald wrote:
>>> On Sunday 10 January 2010 03:31:43 Dan Langille wrote:
>>>> I am seeing a number of interesting notices during regression testing.
>>>>
>>>> Should we be concerned?
>>>>
>>>>> Logcheck system account wrote:
>>>>>> Security Events
>>>>>> =-=-=-=-=-=-=-=
>>>>>> Jan  9 16:02:05 dbclone postgres[82038]: [2-1] ERROR:  must be owner
>>>>>> of relation basefiles Jan  9 16:03:15 dbclone postgres[82349]: [2-1]
>>>>>> ERROR: must be owner of relation basefiles Jan  9 16:03:42 dbclone
>>>>>> postgres[82593]: [2-1] ERROR:  must be owner of relation basefiles Jan
>>>>>> 9 16:04:17 dbclone postgres[82804]: [2-1] ERROR:  must be owner of
>>>>>> relation basefiles Jan  9 16:05:45 dbclone postgres[83099]: [2-1]
>>>>>> ERROR: must be owner of relation basefiles Jan  9 16:06:48 dbclone
>>>>>> postgres[83415]: [2-1] ERROR:  must be owner of relation basefiles Jan
>>>>>> 9 16:07:13 dbclone postgres[83647]: [2-1] ERROR:  must be owner of
>>>>>> relation basefiles Jan  9 16:07:58 dbclone postgres[83884]: [2-1]
>>>>>> ERROR: must be owner of relation basefiles Jan  9 16:08:50 dbclone
>>>>>> postgres[84188]: [2-1] ERROR:  must be owner of relation basefiles Jan
>>>>>> 9 16:10:34 dbclone postgres[84469]: [2-1] ERROR:  must be owner of
>>>>>> relation basefiles Jan  9 16:16:44 dbclone postgres[84986]: [2-1]
>>>>>> ERROR: must be owner of relation basefiles Jan  9 16:20:15 dbclone
>>>>>> postgres[85289]: [2-1] ERROR:  must be owner of relation basefiles Jan
>>>>>> 9 16:21:03 dbclone postgres[85571]: [2-1] ERROR:  must be owner of
>>>>>> relation basefiles Jan  9 16:21:19 dbclone bacula-dir: message.c:590
>>> The above messages seem to be some permissions error in setting up the
>>> PostgreSQL tables -- in particular the BaseFiles table.  You might want
>>> to take a closer look at the tables and whether Bacula (probably user
>>> regress) has appropriate permissions on them.   This is a new table added
>>> in 3.1.x and possibly something went wrong during the upgrade.
>> Confusion.
>>
>> This is regression testing.  
> 
> No, no confusion at all.  I recognize very easily the output of this 
> particular regression test (messages-test).
> 
>> The regress database is created and 
>> permissions assigned during that testing.  I would therefore conclude,
>> the upgrade process is not relevant to this situation.  It must be
>> something else.
>>
>> For what it's worth, I've already checked
>> src/cats/grant_postgresql_privileges.in and basefiles is given the same
>> treatment as other files.
>>
>> Later today, I'll run the regression testing and inspect the database
>> table permissions at that time.
> 
> You can probably just run tests/messages-test

That did indeed generate that 'owner' message.  This is interesting 
though.  dan is the owner.  So why an issue? I suspect: Because the 
owner must be regress, the owner as which the tests are running.

  Schema |     Name     | Type  | Owner
--------+--------------+-------+-------
  public | basefiles    | table | dan
  public | cdimages     | table | dan
  public | client       | table | dan
  public | counters     | table | dan
  public | device       | table | dan
  public | file         | table | dan
  public | filename     | table | dan
  public | fileset      | table | dan
  public | job          | table | dan
  public | jobhisto     | table | dan
  public | jobmedia     | table | dan
  public | location     | table | dan
  public | locationlog  | table | dan
  public | log          | table | dan
  public | media        | table | dan
  public | mediatype    | table | dan
  public | path         | table | dan
  public | pool         | table | dan
  public | status       | table | dan
  public | storage      | table | dan
  public | unsavedfiles | table | dan
  public | version      | table | dan
(22 rows)

regress=# \dp basefiles
                               Access privileges
  Schema |   Name    | Type  |  Access privileges  | Column access 
privileges
--------+-----------+-------+---------------------+--------------------------
  public | basefiles | table | dan=arwdDxt/dan     |
                             : regress=arwdDxt/dan
(1 row)


I suspect it is this line:

db_sql_query(db, "CREATE INDEX basefiles_jobid_idx ON BaseFiles ( JobId 
)" , NULL, NULL);

but by hand, it works:

regress=# CREATE INDEX basefiles_jobid_idx ON BaseFiles (jobid);
CREATE INDEX
regress=#

But as the user regress:

regress=> CREATE INDEX basefiles_jobid_idx ON BaseFiles (jobid);
ERROR:  must be owner of relation basefiles
regress=>

I'm not sure of the correct fix for this.

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_______________________________________________
Bacula-devel mailing list
Bacula-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-devel

Reply via email to