[Swift-devel] Re: Kickstart

[Brian Tierney]

Its important to us to be compatible with syslog, so newlines are not 
allowed.

You can talk to Stu or Martin to see how they handled this for WS-GRAM.


Mihael Hategan wrote:
> That doesn't sound quite right. I somewhat doubt a logging format will
> get wide adoption if it does not adapt to the realities of the software
> it tries to service. Perhaps there is a way to reconcile things?
> 
> Mihael
> 
> On Wed, 2007-09-12 at 21:43 -0700, Dan Gunter wrote:
>> Newlines would be replaced with some other allowed character(s).
>>
>> -Dan
>>
>> Mihael Hategan wrote:
>>> So how would a java exception (being a multi-line thing) be logged in
>>> this format?
>>>
>>> Mihael
>>>
>>> On Wed, 2007-09-12 at 16:35 -0500, Michael Wilde wrote:
>>>   
>>>> Dan, sounds good.  As a GRAM and GridFTP client (using CoG interfaces) 
>>>> how would Swift (or Karajan, which Swift calls) insert these identifiers?
>>>>
>>>> (I'll step aside and let Mihael pick the discussion up with you when 
>>>> time permits - he's the developer of all layers of code involved from 
>>>> CoG up, and would know all the details).
>>>>
>>>> - Mike
>>>>
>>>>
>>>> Dan Gunter wrote:
>>>>     
>>>>> Hi,
>>>>>
>>>>> We (CEDPS) have prototype code to load logs, using a simple mapping of
>>>>> one event type to one table, into a relational DB (MySQL, Postgres,
>>>>> SQLite). We could use this as a starting point for some of the analysis.
>>>>>
>>>>> One of the issues that comes up immediately is how to know which events
>>>>> belong to which workflow. Any help kickstart could provide w.r.t.
>>>>> inserting common identifiers would ease the correlation task down the line..
>>>>>
>>>>> -Dan
>>>>>
>>>>> Michael Wilde wrote:
>>>>>       
>>>>>> Hi Brian,
>>>>>>
>>>>>> Lets continue this informally in email. Perhaps Ben (Swift Tech
>>>>>> Coordinator) can join one of the calls.
>>>>>>
>>>>>> It will be pretty simple to get kickstart data in the right form.
>>>>>>
>>>>>> But an intersting and perhaps valuable (dare I say exciting?) demo of
>>>>>> logging would be to do a multi-site workflow, and correlate all he
>>>>>> logs from GRAM, GridFTP, and Swift, into a meaningful ans useful
>>>>>> stream; then do queries/reports on that stream.
>>>>>>
>>>>>> I'm cc'ing the Swift devel list to encourage discussion, but I think
>>>>>> we can make small continuos steps in this direction and make something
>>>>>> interesting happen pretty readily.
>>>>>>
>>>>>> What would we need to do to get CEDPS-conformant logging from GRAM &
>>>>>> GridFTP at the sites we frequent? (the CI's TeraPort, uc-teragrid, and
>>>>>> tg sites at ncsa, sdsc, purdue, tacc, perhaps a few more?)  Plus OSG
>>>>>> sites as we start using them more?
>>>>>>
>>>>>> - Mike
>>>>>>
>>>>>>
>>>>>> Brian Tierney wrote:
>>>>>>         
>>>>>>> Mike:
>>>>>>>
>>>>>>> After reading the Kickstart paper, I think there are several possible
>>>>>>> areas of collaboration with CEDPS.
>>>>>>>
>>>>>>> Do you want to talk about this on one of the CEDPS calls, or perhaps
>>>>>>> better would be to discuss this in person when we are at ANL for the
>>>>>>> CEDPS meeting at the end of October.
>>>>>>>
>>>>>>> Thoughts?
>>>>>>>
>>>>>>>           
>>>>>       
>>>> _______________________________________________
>>>> Swift-devel mailing list
>>>> Swift-devel at ci.uchicago.edu
>>>> http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel
>>>>
>>>>     
>>

-- 
------------------------------------------------------------------------
   Brian L. Tierney,   Lawrence Berkeley National Laboratory (LBNL)
   1 Cyclotron Rd.  MS: 50B-2239,  Berkeley, CA  94720
   tel: 510-486-7381    fax: 510-495-2998   efax: 425-642-4558
   bltierney at lbl.gov   http://www-didc.lbl.gov/~tierney
------------------------------------------------------------------------