Or ... is the OP wondering about "attaching to a process?"

Original Poster: have you "attached to process" on your local machine,
or have you always "started with debug" to test locally?

∞ Andy Badera
∞ +1 518-641-1280
∞ This email is: [ ] bloggable [x] ask first [ ] private
∞ Google me: http://www.google.com/search?q=(andrew+badera)+OR+(andy+badera)



On Fri, Sep 4, 2009 at 7:04 AM, Andrew Badera<[email protected]> wrote:
> When someone says "live" I equate that to "production." Staging isn't
> live. It's not mission critical. "Live" is user-facing and activated
> and in-use and possibly enterprise-driving.
>
> Remote debugging can be a pain. Usually is. But if you've got the
> remote debug components in place and you can attach to the process, as
> the OP indicated, it's just like debugging locally. If the "live"
> server doesn't have .pdb symbol files, which a production server
> shouldn't, I believe you should be able to reference the .pdbs in the
> build directory that matches the set of assemblies on the server.
>
> ∞ Andy Badera
> ∞ +1 518-641-1280
> ∞ This email is: [ ] bloggable [x] ask first [ ] private
> ∞ Google me: http://www.google.com/search?q=(andrew+badera)+OR+(andy+badera)
>
>
>
> On Fri, Sep 4, 2009 at 6:59 AM, Cerebrus<[email protected]> wrote:
>>
>> That may be so, Andrew, but there are various types of remote servers
>> - Staging servers, Test servers, Development servers etc. Remote
>> Debugging on such a server is a useful trick. I happen to have tried
>> it once earlier and was unsuccesful, just like the OP. So, I'm very
>> interested to know the answer to this one.
>>
>>
>> On Sep 3, 9:49 pm, Andrew Badera <[email protected]> wrote:
>>>
>>> Short answer: you don't. Most people in their right minds wouldn't
>>> attach to a live process for debug.
>>>
>>> Long answer: you should be building flexible logging and tracing into
>>> your code for these scenarios. log4net and the Enterprise Library
>>> Exception Handling Application Block (EHAB) and Logging Application
>>> Block (LAB) work well for this. You can turn tracing off or on from
>>> the config file, so you wrap all operations in a trace. When you turn
>>> it on, you get a trace log of all activities, so you know where
>>> processing was at when an error occurred.
>>>
>>> Use exception handling and tunable-logging to send more or less,
>>> richer or poor, details to the log. Again, this stuff is set in the
>>> config file. You see problems, you turn it on. Problems solved? Turn
>>> it off.
>>>
>>> ∞ Andy Badera
>>> ∞ +1 518-641-1280
>>> ∞ This email is: [ ] bloggable [x] ask first [ ] private
>>> ∞ Google 
>>> me:http://www.google.com/search?q=(andrew+badera)+OR+(andy+badera)- Hide 
>>> quoted text -
>>>
>>> - Show quoted text -
>>
>

Reply via email to