[Box Backup-dev] RC 3 testing: NetBSD 3.0 fails

Ben Summers boxbackup-dev at fluffy.co.uk
Fri Feb 24 13:39:06 GMT 2006


On 22 Feb 2006, at 07:14, Per Thomsen wrote:

> On 2/21/06 10:56 PM, Ben Summers wrote:
>>
>> On 22 Feb 2006, at 05:33, Per Thomsen wrote:
>>
>>> On 2/21/06 11:27 AM, Per Thomsen wrote:
>>>> On 2/21/06 10:48 AM, Ben Summers wrote:
>>>>
>>>>> On 21 Feb 2006, at 18:41, Per Thomsen wrote:
>>>>>
>>>>>> When run individually, backupstorepatch passed (problem appeared
>>>>>> to be a
>>>>>> non-dying bbstored from a previous test).
>>>>>>
>>>>>>
>>>>>> The bbackupd test, when run exits at this point in the test:
>>>>>>
>>>>>> [ 0 (of 0) differences probably due to file modifications after
>>>>>> the last
>>>>>> upload ]
>>>>>> Differences: 0 (0 dirs excluded, 0 files excluded)
>>>>>> ..
>>>>>> Add files with current time
>>>>>> --------
>>>>>> bbackupd: Add files with current time
>>>>>>
>>>>>>
>>>>>> I can't find a core file anywhere. I am re-running the test now.
>>>>>>
>>>> Here are the results (same as with HAVE_KQUEUE):
>>>>
>>>> : {240} grep -A 3 -B 3 FAIL debug-bbackupd-2.out
>>>> [ 11 (of 12) differences probably due to file modifications  
>>>> after the
>>>> last upload ]
>>>> Differences: 12 (3 dirs excluded, 4 files excluded)
>>>>
>>>> FAILURE: Condition [compareReturnValue == 1*256] failed at
>>>> testbbackupd.cpp(837)
>>>> Rename directory
>>>> waiting: ............Local file 'testfiles/TestDir1/apropos'  
>>>> exists,
>>>> but
>>>> store file '/Test1/apropos' does not exist.
>>>> (the file above was modified after the last sync time -- might be
>>>> reason
>>>> for difference)
>>>> -- 
>>>> [ 5 (of 10) differences probably due to file modifications after  
>>>> the
>>>> last upload ]
>>>> Differences: 10 (3 dirs excluded, 4 files excluded)
>>>>
>>>> FAILURE: Condition [compareReturnValue == 1*256] failed at
>>>> testbbackupd.cpp(845)
>>>> FAILURE: Condition [compareReturnValue == 1*256] failed at
>>>> testbbackupd.cpp(849)
>>>> Rename files
>>>> waiting: ............Local file 'testfiles/TestDir1/ 
>>>> continousupdate'
>>>> does not exist, but store file '/Test1/continousupdate' does.
>>>> Local file 'testfiles/TestDir1/df324' does not exist, but store  
>>>> file
>>>> '/Test1/df324' does.
>>>> -- 
>>>> [ 9 (of 16) differences probably due to file modifications after  
>>>> the
>>>> last upload ]
>>>> Differences: 16 (3 dirs excluded, 4 files excluded)
>>>>
>>>> FAILURE: Condition [compareReturnValue == 1*256] failed at
>>>> testbbackupd.cpp(859)
>>>> Create a file with timestamp to way ahead in the future
>>>> waiting: ............
>>>> [ 0 (of 0) differences probably due to file modifications after the
>>>> last
>>>> upload ]
>>>> -- 
>>>> 00000052 -d---- 1970-01-01T00:00:00 0000000000000000
>>>> Test1/renamed-dir/dsfsfdslkjfkdsjf
>>>> 00000053 f----- 2006-02-21T11:23:27 6c56ee3b9bb9e1e8
>>>> Test1/fileaftermarker
>>>>
>>>> FAILED: 4 tests failed
>>>> --------
>>>> bbackupd: FAILED: 4 tests failed
>>>>
>>>>
>>> The bbackupquery return code is 512, and not 256, in every  
>>> failure. I
>>> will debug this, but is there a key to the return codes?
>>
>>
>> In bbackupquery, type "help compare", which includes:
>>
>>     If -c is set, then the return code (if quit is the next command)
>> will be
>>         1    Comparison was exact
>>         2    Differences were found
>>         3    An error occured
>>     This can be used for automated tests.
>>
>> Not the obvious location for the info!
>>
>>>
>>> I ran the bbackupd tests in release mode, just for laughs, and  
>>> all tests
>>> pass, and the return code from bbackupquery is 256 in the tests that
>>> failed in debug mode.
>>
>> I'm wondering if it's just a timing issue. There isn't enough context
>> in your listing to tell, though, as you miss out the bit which says
>> what it was doing at the time.
> The first failing debug test in bbackupd is "Add files with current
> time",  then "Rename directory" (2 FAILUREs), and finally "Rename
> files". I have attached the output.

Since ones after this are working, this means it just took a bit  
longer than expected to do it's magic.

I hate timing sensitive tests, but you can either have them being a  
bit annoying sometimes or taking hours to run.

Ben







More information about the Boxbackup-dev mailing list