Karen Tung wrote:
> Hi Jean,
>
> Thank you for making the changes.  I still have a few comments:
>
> - line 34: why combine the 3 separate lines into this one line?  I 
> don't think we need everything from dc_utils.py.
Because there is very little in dc_utils.py and I find it annoying to 
have multiple lines as previously done.
I guess just a person "button". I can change if you really want.
> - line 409: I think we don't need to mention the empty snapshot here
Except that it's kind of important. If you leave the empty snapshot in, 
the functionality doesn't work correctly.
> - line 686: why is the "-r" removed from here as well as the 
> finalizer_rollback.py?
-r is added not removed. The reasoning is that if you wish to rollback 
to a step and you've deleted later steps
this should be legal. Without the -r it doesn't work.

Jean

>
> Thanks,
>
> --Karen
>
> Jean McCormack wrote:
>> Karen,
>>
>> I believe I've addressed all of your concerns. The webrev is at : 
>> http://cr.opensolaris.org/~jeanm/slim_5170/
>>
>> Jean
>>
>> Karen Tung wrote:
>>> Hi Jean,
>>>
>>> Here are my comments:
>>>
>>> - Line 402-420.  My understanding of this code segment is
>>> that you are getting a list of all the snap shots, and making
>>> sure that the order of the of checkpoints corresponds to the order 
>>> of the
>>> snapshots list.  Your list of checkpoints is built by reading the 
>>> manifest,
>>> so, I understand that the checkpoints will be in the order that's 
>>> specified.
>>> How do we ensure that the output of "zfs list...." will follow the 
>>> same order?
>>> Right now, it just so happens that they follow the same order.
>>> Is there a rule or something that zfs datasets/snapshots will be listed
>>> in the order they are created?  I don't see an option in "zfs list" 
>>> to sort
>>> the datasets by the time they are created or something.  That will
>>> certainly gurantee the output of "zfs list" will follow the order
>>> of things specified in the manifest.  I am concern
>>> that if "zfs list" ever change the order of the output, then, our 
>>> code won't work.
>>>
>>> - lines 346-348: I think it is inefficient to execute a system shell 
>>> and do a grep on
>>> the temp file content since this loop is executed for each item on 
>>> the step list.  I think
>>> it is more efficient to read the content of the temp file into a 
>>> python list and just
>>> search the internal list.
>>>
>>> - line 348: since you already found the step_num here, can we break 
>>> out the loop?
>>>
>>> - Lines 333-337 and 389-395 are doing the same thing.  I think it is 
>>> better to have the common
>>> code in a function or something.  Even better, you can make this 
>>> function return a
>>> list of snapshot names.  Also, I just saw something on the python 
>>> docs page about how
>>> to do the shell pipeline in Python.  Maybe you want to try something 
>>> like this?
>>> http://docs.python.org/library/subprocess.html#subprocess-replacements
>>>
>>> Thanks,
>>>
>>> --Karen
>>>
>>> Jean McCormack wrote:
>>>> Please review the fix for
>>>> 5170 DC_determine_Resume_step should check snapshots existence not 
>>>> .step files.
>>>>
>>>> Bug:
>>>> http://defect.opensolaris.org/bz/show_bug.cgi?id=5170
>>>>
>>>> Webrev:
>>>> http://cr.opensolaris.org/~jeanm/slim_5170/
>>>>
>>>> Testing included:
>>>>
>>>> 1) distro_const build -l <manifest>
>>>>
>>>> 2) distro_const build -r <legal value> <manifest>
>>>>
>>>> 3) distro_const build -r <illegal value> <manifest>
>>>>
>>>> 4) Modify manifest adding another finalizer script.
>>>>     Repeat 1,2,3
>>>>
>>>> 5) Modify manifest removing a finalizer script
>>>>     Repeat 1,2,3
>>>>
>>>> 6) Modify manifest shuffling the finalizer scripts
>>>>     Repeat 1,2,3
>>>>
>>>> Jean
>>>>
>>>> _______________________________________________
>>>> caiman-discuss mailing list
>>>> caiman-discuss at opensolaris.org
>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>>>   
>>>
>>
>


Reply via email to