<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1"> <META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2658.24"> <TITLE>Problem with initial service scheduling (2.0b3)</TITLE> </HEAD> <BODY> Hi all, I currently have a configuration with 4800 services : 600 active and 4200 passive. And, as the number was growing, I noticed a problem in the way nagios scheduled their initial check time : With the 2.0b3 original code, with max_service_check_spread=30, when I look at the scheduling queue just after start, I see that the last service checks are scheduled to run in 4 hours ! This delay corresponds to : max_service_check_spread * (total_services / total_scheduled_services) And should be equal to max_service_check_spread. I found the reason in event.c/init_timing_loop() and I am including a change which appears to correct the problem but, as I am not sure to fully understand the 'interleave_block' logic, this change should be taken with care : The reason : in the 'schedule service checks' section of init_timing_loop(), next check time is incremented for each service, and not for each SCHEDULED service. So, in my case it is incremented 'total_services' times and the last check time is equal to : Current_time + total_services * service_inter_check_delay Where it should be : Current_time + total_scheduled_services * service_inter_check_delay Which is coherent with the way service_inter_check_delay is computed. My change consists of taking the 'should_be_scheduled' check out of the inner loop, and add a line in order to have the code enter the inner 'interleave_block' loop only for active checks. This way current_interleave_block goes from 0 to total_schedules_services instead of going up to total_services. Once again, the patch I am submitting seems to correct the problem in MY case. But I don't know if it is correct when interleave variables have some different values. Regards François </BODY> </HTML>