On Mon, 2003-02-03 at 21:04, Bill Davidsen wrote:
>
> That is a problem with processes left running. I do not forward
> connections, I do not forward X, I do not (in normal practice) leave
> anything running. A typical thing to do is to go to each machine in a
> cluster and look for a user activity:
> grep "user" log/stats.readers
> exit
> nothing more. And every once in a while that hangs after executing the
> logout sequence. With the patch it hasn't to date.
>
> That doesn't mean it's a fix, I don't see it every day, I just haven't
> seen it in a few days since I put in the patch.
The ssh hang on exit "problem" is a policy of the ssh coders. It'll
happen when you have a background job still running when you exit, which
is still connected to the terminal.
As I said, it's an ssh policy issue (which many people disagree with)
and not a bug.
-tony
I've run into this often with a background job however .. it also hangs
when there isn't any background job. I suspect there definitely is a bug.
David
Tony Gale wrote:
>On Mon, 2003-02-03 at 21:04, Bill Davidsen wrote:
>
>
>>That is a problem with processes left running. I do not forward
>>connections, I do not forward X, I do not (in normal practice) leave
>>anything running. A typical thing to do is to go to each machine in a
>>cluster and look for a user activity:
>> grep "user" log/stats.readers
>> exit
>>nothing more. And every once in a while that hangs after executing the
>>logout sequence. With the patch it hasn't to date.
>>
>>That doesn't mean it's a fix, I don't see it every day, I just haven't
>>seen it in a few days since I put in the patch.
>>
>>
>
>The ssh hang on exit "problem" is a policy of the ssh coders. It'll
>happen when you have a background job still running when you exit, which
>is still connected to the terminal.
>
>As I said, it's an ssh policy issue (which many people disagree with)
>and not a bug.
>
>-tony
>
>
On 4 Feb 2003, Tony Gale wrote:
> On Mon, 2003-02-03 at 21:04, Bill Davidsen wrote:
> >
> > That is a problem with processes left running. I do not forward
> > connections, I do not forward X, I do not (in normal practice) leave
> > anything running. A typical thing to do is to go to each machine in a
> > cluster and look for a user activity:
> > grep "user" log/stats.readers
> > exit
> > nothing more. And every once in a while that hangs after executing the
> > logout sequence. With the patch it hasn't to date.
> >
> > That doesn't mean it's a fix, I don't see it every day, I just haven't
> > seen it in a few days since I put in the patch.
>
> The ssh hang on exit "problem" is a policy of the ssh coders. It'll
> happen when you have a background job still running when you exit, which
> is still connected to the terminal.
Please go back and reread either of my comments on the topic, I think I've
made it clear that I have no background jobs, no forwarded ports, and no
forwarded X. The existance of a problem in one area doesn't mean that
nothing else is allow to cause bad behaviour.
> As I said, it's an ssh policy issue (which many people disagree with)
> and not a bug.
--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.