chukwa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric Yang <ey...@yahoo-inc.com>
Subject Re: rolling a 0.3 release?
Date Thu, 01 Oct 2009 17:00:26 GMT
This is broken by CHUKWA-308.  There are several patches to HICC after 308.
I also attempted to fix the current HICC with minimized change for Stephen
Nelson, but my conclusion was that was not possible without removing taglib
dependency.  Since then, I have been working actively on this over weekends,
but it will take another 3 weekends of work to complete this.

Regards,
Eric 


On 10/1/09 9:11 AM, "Jerome Boulon" <jboulon@netflix.com> wrote:

> Eric, do you know which update/patch broke HICC?
> /Jerome.
> 
> On 10/1/09 12:27 AM, "Jiaqi Tan" <tanjiaqi@gmail.com> wrote:
> 
>> Is there any way we can mark out only the log collection changes for
>> 0.3 release, and "freeze" HICC at the last point where it was working?
>> 
>> Jiaqi
>> 
>> On Thu, Oct 1, 2009 at 3:19 PM, Ariel Rabkin <asrabkin@gmail.com> wrote:
>>> I dissent.  I think robust log collection, is an important application
>>> in its own right, particularly log collection that makes the data
>>> MapReduceable.  A release with log collection but not HICC would be
>>> much better than no release.
>>> 
>>> Is there harm in doing a release without HICC, or without a working HICC?
>>> 
>>> --Ari
>>> 
>>> On Wed, Sep 30, 2009 at 10:58 PM, Jiaqi Tan <tanjiaqi@gmail.com> wrote:
>>>> I second Eric on that, without HICC there's only the back-end data
>>>> collection which I believe would not be as useful.
>>>> 
>>>> Regards,
>>>> Jiaqi
>>>> 
>>>> On Thu, Oct 1, 2009 at 12:56 AM, Eric Yang <eyang@yahoo-inc.com> wrote:
>>>>> Hicc is currently broken.  The "create dashboard" feature is broken
due to
>>>>> incompatible Jasper.runtime.ELContextImpl bundled in Apache taglib and
>>>>> jetty-jsp api.  I have a working solution to remove apache taglib from
>>>>> Hicc
>>>>> code, and going with pure REST & serve HICC views from HDFS.  I
need at
>>>>> least 3 more weekends to complete this.  I think it's better that we
put
>>>>> off
>>>>> the 0.3 release and go with 0.4.
>>>>> 
>>>>> Regards,
>>>>> Eric
>>>>> 
>>>>> On 9/29/09 10:39 PM, "Ariel Rabkin" <asrabkin@gmail.com> wrote:
>>>>> 
>>>>>> Hrm.  Are there other features on the horizon that are worth waiting
>>>>>> for?  Even without SALSA, Chukwa supplies useful functionality and
I
>>>>>> think it's worth getting something out there.
>>>>>> 
>>>>>> I'm sort of inclined to do a 0.3 release without SALSA, and then
a 0.4
>>>>>> whenever SALSA  is ready to go. A 0.4 release could also include
a
>>>>>> number of other improvements -- there are things I want to do over
the
>>>>>> coming months, and I suspect Jerome is also implementing things.
>>>>>> 
>>>>>> Thoughts?
>>>>>> 
>>>>>> --Ari
>>>>>> 
>>>>>> On Tue, Sep 29, 2009 at 10:34 PM, Jiaqi Tan <tanjiaqi@gmail.com>
wrote:
>>>>>>> We're still missing a subclass of the AbstractDataProcessor for
fully
>>>>>>> automating the end-to-end process of converting post-Demux output
to
>>>>>>> state-machine entries for SALSA processing, but I don't see myself
>>>>>>> finishing that for at least a month. I would commit some of the
>>>>>>> command-line tools with documentation on how to get the end-to-end
>>>>>>> process done manually as a stop-gap, I can change those JIRAs
to
>>>>>>> blockers for 0.3 and I can get those done in a week or two.
>>>>>>> 
>>>>>>> Jiaqi
>>>>>>> 
>>>>>>> On Wed, Sep 30, 2009 at 12:39 PM, Ariel Rabkin <asrabkin@gmail.com>
>>>>>>> wrote:
>>>>>>>> Howdy.
>>>>>>>> 
>>>>>>>> How does everybody feel about making a formal release of
0.3 ? I can't
>>>>>>>> get to it until next week, but I think we could do it then.
The
>>>>>>>> current SVN trunk is stable, reasonably well documented and
fairly
>>>>>>>> well tested. Are there blockers for release that aren't listed
on
>>>>>>>> JIRA?
>>>>>>>> 
>>>>>>>> --Ari
>>>>>>>> 
>>>>>>>> --
>>>>>>>> Ari Rabkin asrabkin@gmail.com
>>>>>>>> UC Berkeley Computer Science Department
>>>>>>>> 
>>>>>>> 
>>>>>> 
>>>>>> 
>>>>> 
>>>>> 
>>>> 
>>> 
>>> 
>>> 
>>> --
>>> Ari Rabkin asrabkin@gmail.com
>>> UC Berkeley Computer Science Department
>>> 
> 


Mime
View raw message