[erlang-questions] timer server timeout
Mon Apr 26 12:36:20 CEST 2010
At least I did not know about this potential probem with gen_sever timeouts
and I think it should have been reported as a bug.
I think this bad beaviour should be corrected and only if that is not possible
there is reason to document this little nasty side-effect.
On Mon, Apr 26, 2010 at 11:46 AM, Ulf Wiger
> I think we could put this down as one of the things that
> people at Ericsson have known for years but never formally
> raised a complaint about. Mea Culpa, I've avoided the use
> of the internal gen_server timeouts for at least a decade
> for this reason, and have certainly advised others to do it
> as well. I guess I assumed that it was common knowledge by
> now and didn't reflect on whether or not it was in the docs.
> I am almost sure that it has been mentioned a number of times
> on the mailing list, but that's neither here nor there. It
> should be documented.
> It's not just debugging that can cause this. Code change and
> stray messages can do it too. For the same reason that a
> process should be prepared to throw away unrecognized
> messages in some top-level state, it should not allow such
> messages to interfere with timeout handling.
> Ulf W
> Kenneth Lundin wrote:
>> This is not a known issue.
>> After a quick glance I agree with you that a sys:get_status call to
>> the timer_server or any other
>> "gen_server" using the gen_server timeout feature can disturb the
>> timing either by delaying
>> the timeout message or by not letting any timeout message to be sent.
>> I will put this on the list of things to look at.
>> It is whoever very unlikely that this happens in normal operation
>> since there is no reason to call
>> sys:get_status other than for debugging purposes.
>> I have never heard of any real problems caused by this.
>> /Kenneth Erlang/OTP Ericsson
>> On Wed, Apr 21, 2010 at 10:05 PM, johan munk <> wrote:
>>> Following the control flow in the gen_server, sys, and timer modules I
>>> Calling sys:get_status(timer_server) will "reset" the "after time" in the
>>> loop function of gen_server. As the timer server uses that time to get a
>>> timeout if no other messages are received, one can potentially cause
>>> to never trigger.
>>> Is this a known issue/feature?
>> erlang-questions (at) erlang.org mailing list.
>> See http://www.erlang.org/faq.html
>> To unsubscribe; mailto:
> Ulf Wiger
> CTO, Erlang Solutions Ltd, formerly Erlang Training & Consulting Ltd
> WE'VE CHANGED NAMES!
> Since January 1st 2010 Erlang Training and Consulting Ltd. has become ERLANG
> SOLUTIONS LTD.
More information about the erlang-questions