Ah, I understand the logging better then. Not only is it logging it's own
error, but it's recording the one thrown by the remote server when the
logout request fails.

I think I should build that into the spec for client app deployers, i.e. use
a client that accepts single log out, or build it into your custom client.

Kim

~=|=~

Kim Cary
Chief Information Security Officer
Pepperdine University



On Mon, Dec 14, 2009 at 5:48 AM, Scott Battaglia
<[email protected]>wrote:

> On Fri, Dec 11, 2009 at 12:39 PM, Kim Cary <[email protected]>wrote:
>
>> Thanks all, let me just check what's happening here:
>>
>> CAS server sends single signout to all apps with service tickets for the
>> user.
>> Some apps don't give the expected response (and give server errors on
>> receiving the SSO traffic over https).
>> The CAS server doesn't get the expected response from SSO and throws an
>> error (OR the CAS server logs the server error from the external app in its
>> server log).
>
>
> The CAS server logs the errror (and its own generated error).  The error is
> not shown to the user though.
>
> Cheers,
> Scott
>
>
>
>>
>>
>> Kim
>>
>> ~=|=~
>>
>> Kim Cary
>> Chief Information Security Officer
>> Pepperdine University
>>
>>
>>
>> On Fri, Dec 11, 2009 at 5:41 AM, Scott Battaglia <
>> [email protected]> wrote:
>>
>>> Created CAS-828 to track it.
>>>
>>> Cheers,
>>> Scott
>>>
>>> On Fri, Dec 11, 2009 at 8:35 AM, Marvin Addison <
>>> [email protected]> wrote:
>>>
>>>> > The alternative is that if this list agrees we can change the logging
>>>> level
>>>> > of that error message and possibly not include the entire stack trace
>>>> ;-)
>>>>
>>>> +1 for some change here.  It's important to log it, but the stack
>>>> trace is both verbose and virtually useless.  We just need to know
>>>> what URL failed and the HTTP response code.  Suggested message text
>>>> off the top of my head:
>>>>
>>>> WARN Single sign out failed for http://example.com/app/?data.
>>>> Application returned HTTP 500 response.
>>>>
>>>> We also see lots of connection failures in our environment, which
>>>> should be treated similarly.  I'd recommend logging this at WARN since
>>>> it has security implications.
>>>>
>>>> M
>>>>
>>>> --
>>>> You are currently subscribed to [email protected] as:
>>>> [email protected]
>>>>
>>>> To unsubscribe, change settings or access archives, see
>>>> http://www.ja-sig.org/wiki/display/JSG/cas-user
>>>>
>>>
>>> --
>>> You are currently subscribed to [email protected] as: 
>>> [email protected]
>>>
>>>
>>>
>>>
>>>
>>> To unsubscribe, change settings or access archives, see 
>>> http://www.ja-sig.org/wiki/display/JSG/cas-user
>>>
>>>
>> --
>>
>> You are currently subscribed to [email protected] as: 
>> [email protected]
>>
>>
>>
>> To unsubscribe, change settings or access archives, see 
>> http://www.ja-sig.org/wiki/display/JSG/cas-user
>>
>>
> --
> You are currently subscribed to [email protected] as: 
> [email protected]
>
> To unsubscribe, change settings or access archives, see 
> http://www.ja-sig.org/wiki/display/JSG/cas-user
>
>

-- 
You are currently subscribed to [email protected] as: 
[email protected]
To unsubscribe, change settings or access archives, see 
http://www.ja-sig.org/wiki/display/JSG/cas-user

Reply via email to