On 21 Sep 2018, at 12:48, Bogdan-Andrei Iancu wrote: > Dan, > > For the delay warnings, that's what the code does - it checks time difference > (internal ticker) between writing and reading from a pipe. If you there is no > load (to prevent the workers to read from the pipe), the only delay may come > from the actual pipe communication. Why the delay, not sure for me too. > > On the second issue (already scheduled), the warning reports an almost that > the timer job was triggered almost 30 seconds ago without completing (note > that the timestamp between the brackets is the absolute time from start, not > a diff -> the 21 minutes are the uptime, not the delay). > > When you get such warnings, try to run an "opensipsctl trap" , maybe we can > see what the timer job is doing - I guess it is very slow (in whatever doing).
It might be triggered by something else. Yesterday when I had the issue I had another issue where radius was not able to insert records in the database and consequently opensips did timeout on those radius accounting requests. It also seems that the radius timeout was the same 30 seconds. After I fixed the radius issue I've never seen the timer warnings again. So it's possible that they only trigger under such circumstances and are somehow related to the radius accounting timing out. -- Dan _______________________________________________ Users mailing list Users@lists.opensips.org http://lists.opensips.org/cgi-bin/mailman/listinfo/users
