Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #34385 > unrolled thread

Stallled Gets

Back to article view | Back to comp.lang.java.programmer

Contents

  Stallled Gets Roedy Green <see_website@mindprod.com.invalid> - 2016-05-25 08:26 -0700
    Re: Stallled Gets Andreas Leitgeb <avl@logic.at> - 2016-05-25 16:52 +0000
    Re: Stallled Gets Arne Vajhøj <arne@vajhoej.dk> - 2016-05-25 20:03 -0400
      Re: Stallled Gets Roedy Green <inquiry@mindprod.com> - 2016-05-30 21:25 -0700
        Re: Stallled Gets Brixomatic <brixomatic@yahoo.com> - 2016-06-04 11:05 +0200

#34385 — Stallled Gets

Most of the time a failed Get with connect or read times out. But once
every 6 hours or so of requests it will just stall. I have a number of
threads running at once, so it is possible the stall has to do with
threads.  I wonder if anyone else has seen this and has a suggestion
how to track down or fix the trouble.  It is as though an infinite
timeout had been specified.
-- 
Roedy Green Canadian Mind Products http://mindprod.com
Incorrect documentation is often worse than no documentation.
~ Bertrand Meyer (born:1950 age 65)

[toc] | [next] | [standalone]

#34387

Roedy Green <see_website@mindprod.com.invalid> wrote:
> Most of the time a failed Get with connect or read times out. But once
> every 6 hours or so of requests it will just stall. I have a number of
> threads running at once, so it is possible the stall has to do with
> threads.  I wonder if anyone else has seen this and has a suggestion
> how to track down or fix the trouble.  It is as though an infinite
> timeout had been specified.

I assume that a "Get" in your case is the retrieval of contents
specified by an URL.

In principle, there exist a few different possible timeouts:
-: upon trying to resolve hostnames
-: upon creating the connection

Once the connection is established:
-: a tcp/ip-timeout of the connection
-: an application timeout, if, despite the tcp/ip connection
    being alive, the peer just doesn't send data

It would be useful if you could find out at which of these states
of progress exactly it fails to timeout.

[toc] | [prev] | [next] | [standalone]

#34389

On 5/25/2016 11:26 AM, Roedy Green wrote:
> Most of the time a failed Get with connect or read times out. But once
> every 6 hours or so of requests it will just stall. I have a number of
> threads running at once, so it is possible the stall has to do with
> threads.  I wonder if anyone else has seen this and has a suggestion
> how to track down or fix the trouble.  It is as though an infinite
> timeout had been specified.

HTTP GET?

Are you displaying/logging all exceptions?

Arne

[toc] | [prev] | [next] | [standalone]

#34415

On Wed, 25 May 2016 20:03:31 -0400, Arne Vajhøj <arne@vajhoej.dk>
wrote:

>Are you displaying/logging all exceptions?

Yes.  Nothing happens. The program just stalls as if I had specified
an infinite timeout. No exception.

Part of the problem is this happens rarely and typically after running
6  hours or so.  Is there some way to probe a frozen program to find
out what every thread is waiting for?|

[toc] | [prev] | [next] | [standalone]

#34457

In article <cs3qkbpbhlr1m8j3tdhlpiek8jp30dui71@4ax.com>, Roedy Green 
(inquiry@mindprod.com) says...
> 
> On Wed, 25 May 2016 20:03:31 -0400, Arne Vajhøj <arne@vajhoej.dk>
> wrote:
> 
> >Are you displaying/logging all exceptions?
> 
> Yes.  Nothing happens. The program just stalls as if I had specified
> an infinite timeout. No exception.
> 
> Part of the problem is this happens rarely and typically after running
> 6  hours or so.  Is there some way to probe a frozen program to find
> out what every thread is waiting for?|

Have you tried JVisualVM, which has a thread view and the possibility to 
dump threads?
Have you tried 
jstack <pid> >> threaddump.log
(jstack is included in the jdk\bin folder)?
Have you tried running your programm in the debugger and stopping the 
program using the pause button?

Kind regards,
Wanja



-- 
..Alesi's problem was that the back of the car was jumping up and down 
dangerously - and I can assure you from having been teammate to 
Jean Alesi and knowing what kind of cars that he can pull up with, 
when Jean Alesi says that a car is dangerous - it is. [Jonathan Palmer]

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.java.programmer

csiph-web