[OpenSIPS-Users] Issues with internal timer ticks

[Bogdan-Andrei Iancu]

Hi Collin, Hi Vitalii,

On long run, the OpenSIPS timer is not design to guarantee alignment 
with the external time flow. The internal timer is design for measuring 
timer delta intervals (rather than against the original time).

As the timer is looping on ~1 second, after 100000 seconds, that ~ 
accumulates and become significant. Not to mention that the timer 
process is subject to the "realtime-ness" of the Linux kernel.

May I ask way is this a problem ? Once again , the timer is design for 
measuring internal delta-time intervals (usually seconds or minutes), it 
does not want to be a replacement for a clock!

Best regards,

Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
http://www.opensips-solutions.com

On 07.08.2015 01:42, Colin Mason wrote:
>
> I just implemented a timer route that was executing a bit late every 
> time on 1.11.1. Looks like the same issue. This is not a virtual 
> machine. I found this thread when googling. You can see that there is 
> a constant time skew.
>
> CentOS 6.5.
>
> Linux 2.6.32-431.5.1.el6.x86_64 #1 SMP Wed Feb 12 00:41:43 UTC 2014 
> x86_64 x86_64 x86_64 GNU/Linux
>
> opensipsctl fifo uptime;opensipsctl fifo get_statistics timestamp
>
> Now:: Thu Aug  6 18:33:23 2015
>
> Up since:: Thu Aug  6 18:18:57 2015
>
> Up time:: 866 [sec]
>
> core:timestamp = 848
>
> Now:: Thu Aug  6 18:42:17 2015
>
> Up since:: Thu Aug  6 18:18:57 2015
>
> Up time:: 1400 [sec]
>
> core:timestamp = 1374
>
> Colin
>
> *From:*users-bounces at lists.opensips.org 
> [mailto:users-bounces at lists.opensips.org] *On Behalf Of *Vitalii Hurin
> *Sent:* Monday, July 06, 2015 8:35 AM
> *To:* users at lists.opensips.org
> *Subject:* [OpenSIPS-Users] Issues with internal timer ticks
>
> Greetings!
>
> I'm looking for someone who can help with understanding the possible 
> reasons of issue with internal timer ticks in OpenSIPS 1.11.3 . So 
> here is the problem: we are developing custom call center and once 
> I've noticed all time-based values that are influenced by internal 
> timer (as it turned out, by function get_ticks() from timer.c) were 
> constantly getting behind actual system time. For example, when I call 
> such command sequence "opensipsctl fifo uptime; opensipsctl fifo 
> get_statistics timestamp", I get this kind of output:
>
> Now:: Wed Jul  1 13:01:52 2015
> Up since:: Tue Jun 30 15:54:50 2015
> Up time:: 76020 [sec]
> core:timestamp:: 75098
>
> As you can see, for nearly one day we have 922 second and I've got 
> even about 3 hour difference for a couple of days! I can't find 
> certain rule of difference increase but it's happening smoothly every 
> time. After checking out some production servers with an old version 
> of our call center based on OpenSIPS 1.8, I've found no similar issues 
> there and core:timestamp value is always equal to up time despite 
> server setups are almost identical. So please help me with searching 
> the reason in our system. Or can it be a bug or something related to 
> timer counting in OpenSIPS 1.11 core? Can't find anything about 
> similar issues on Google...
>
> Thanks in advance!
>
> Vitalii.
>
>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20150807/d98117d7/attachment.htm>