client authentication with revoked cert results in server no response.

huican ping pinghuican at gmail.com
Thu Apr 9 12:44:01 EDT 2009


Hello Trustin,

I am really Sorry, my bad.
Just checked, and Actually the CPU usage is very high after I sent in
4 input messages. So it must stuck at some loop.
After several failed messages (actually, 1st failed one cause CPU jump
to 40%, 2nd failed CPU to 80%,  3rd CPU 90%), it always keeps at 90%
level even there is no more messages in, FYI, the system has no other
heavy CPU consuming process running.

I tried your revision 1160, and it is same as before. So please
re-considering (probably revert) the workaround at 1157.

Sorry again.
Huican

On Wed, Apr 8, 2009 at 5:13 PM, Trustin Lee <tlee at redhat.com> wrote:
> FYI, the revision # is 1157
>
> — Trustin Lee, http://gleamynode.net/
>
>
>
> On Thu, Apr 9, 2009 at 7:10 AM, Trustin Lee <tlee at redhat.com> wrote:
>> I've just checked in the experimental workaround for this particular
>> issue.  Honestly I don't understand why SSLEngine.unwrap() is stuck
>> there.  It is not supposed to behave like that.  Please build from the
>> trunk and let me know if my fix works.
>>
>> Thanks,
>>
>> — Trustin Lee, http://gleamynode.net/
>>
>>
>>
>> On Thu, Apr 9, 2009 at 6:55 AM, Trustin Lee <tlee at redhat.com> wrote:
>>> This is interesting.  Let me investigate on this issue further.
>>>
>>> It is correct to close the channel when you get SSLHandshakeException.
>>>
>>> — Trustin Lee, http://gleamynode.net/
>>>
>>>
>>> On Thu, Apr 9, 2009 at 6:40 AM, huican ping <pinghuican at gmail.com> wrote:
>>>> The threadDump file is attached.
>>>>
>>>> The CPU usage is low, it seems stuck at SSLEngine.unwrap().
>>>>
>>>> The exception in the exceptionCaught() I can see is
>>>> SSLHandshakeException, and I close() the channel after.
>>>>
>>>> Do you have any suggestion what I should do when I got SSLHandshakeException?
>>>> It could be the clean up issue after I got ssl exception.
>>>>
>>>>
>>>> On Wed, Apr 8, 2009 at 1:21 AM, Trustin Lee <tlee at redhat.com> wrote:
>>>>> Hi Huican,
>>>>>
>>>>> Thanks for reporting the problem first of all.
>>>>>
>>>>> Could you provide the full thread dump which can be obtained by kill
>>>>> -3 or jstack command?  The attached screenshot doesn't reveal enough
>>>>> information to track down the problem.
>>>>>
>>>>> — Trustin Lee, http://gleamynode.net/
>>>>>
>>>>>
>>>>>
>>>>> On Tue, Apr 7, 2009 at 3:43 AM, huican ping <pinghuican at gmail.com> wrote:
>>>>>> Hello Trustin,
>>>>>>
>>>>>> I am attaching a yourkit picture showing the blocked threads after I
>>>>>> sent in 4 messages.
>>>>>> You can see our code exceptionCaught(), and netty lib channel close(),
>>>>>> and it stuck at the SSLEngineImpl code.
>>>>>>
>>>>>> I will see whether I can generate something else.
>>>>>>
>>>>>> On Fri, Apr 3, 2009 at 1:23 AM, Trustin Lee <tlee at redhat.com> wrote:
>>>>>>
>>>>>>> Did you find something useful in the full thread dump in client and
>>>>>>> server?  Your application might have been stuck somewhere.  Please
>>>>>>> post the full thread dump of the both side.
>>>>>>>
>>>>>>> If you don't think this is the case where the full thread dump reveals
>>>>>>> the cause of the problem, please attach small test application that
>>>>>>> reproduces the problem.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> — Trustin Lee, http://gleamynode.net/




More information about the netty-users mailing list