[arch-dev-public] [testing] Cleanup?

[Dusty Phillips]

2008/12/2 Dan McGee <dpmcgee at gmail.com>:
> On Tue, Dec 2, 2008 at 8:48 PM, Allan McRae <allan at archlinux.org> wrote:
>> Dan McGee wrote:
>>>
>>> On Tue, Dec 2, 2008 at 2:06 PM, Aaron Griffin <aaronmgriffin at gmail.com>
>>> wrote:
>>>
>>>>
>>>> On Tue, Dec 2, 2008 at 1:31 AM, Allan McRae <allan at archlinux.org> wrote:
>>>>
>>>>>>
>>>>>>
>>>>>> <http://www.archlinux.org/packages/?arch=i686&repo=Testing&q=&last_update=&limit=all>
>>>>>>
>>>>>>
>>>>>
>>>>> OK, I was scanning the repo contents.
>>>>> Looks like the server/db-scripts updates broke something on the web
>>>>> site.
>>>>>  Is this a known issue?
>>>>>
>>>>
>>>> Dusty and/or Dan: ideas?
>>>>
>>>
>>> >From the log (/tmp/archweb_update.log):
>>>
>>> Updating testing-i686
>>> /usr/lib/python2.6/site-packages/MySQLdb/__init__.py:34:
>>> DeprecationWarning: the sets module is deprecated
>>>  from sets import ImmutableSet
>>> 2008-12-02 21:02:45 -> INFO: Starting repo parsing
>>> 2008-12-02 21:02:45 -> INFO: Reading repo tarfile
>>> /srv/ftp/testing/os/i686/testing.db.tar.gz
>>> 2008-12-02 21:02:45 -> INFO: Finished repo parsing
>>> 2008-12-02 21:02:45 -> INFO: Starting database updates.
>>> 2008-12-02 21:02:45 -> INFO: Updating Arch: i686
>>> 2008-12-02 21:02:45 -> INFO: 93 packages in current web DB
>>> 2008-12-02 21:02:45 -> INFO: 45 packages in new updating db
>>> 2008-12-02 21:02:45 -> INFO: 12 packages in sync not db
>>> 2008-12-02 21:02:45 -> ERROR: .db.tar.gz has less than 50% the number
>>> of packages in the web database
>>> Traceback (most recent call last):
>>>  File "/srv/http/sites/archlinux/archweb_dev/scripts/reporead.py",
>>> line 385, in <module>
>>>    sys.exit(main())
>>>  File "/srv/http/sites/archlinux/archweb_dev/scripts/reporead.py",
>>> line 374, in main
>>>    db_update(arch,pkgs)
>>>  File "/srv/http/sites/archlinux/archweb_dev/scripts/reporead.py",
>>> line 186, in db_update
>>>    'it looks like the syncdb is twice as big as the new'
>>> __main__.SomethingFishyException: it looks like the syncdb is twice as
>>> big as the newpackages. WTF?
>>>
>>> So it was trying to save our ass- a good thing.
>>>
>>> I'll go ahead and run it, overriding the check, since we know what we are
>>> doing?
>>>
>>> -Dan
>>>
>>
>> Should an exception be added to this check for the [testing] repo?
>>  Otherwise this will occur anytime a large number of packages is moved to
>> [core]/[extra].
>
> Yeah, probably. But it is quite easy from my POV to override that
> check, so unless this happens a lot, we shouldn't have much to worry
> about.
>
> It might make more sense to do something like
>
> if removecount > (packagecount / 2) and removecount > 50:
>
> or something with an absolute number in it.

On the one hand, that check can probably be removed now, as the
original Great Orphaning bug has been fixed. On the other, keeping it
there could prevent Great Orphaning NG.

Dusty