Recently I started seeing some weird behavior from NaviServer that I've never seen before. From time to time, it looks like the scheduler thread is getting stuck and not running anything, often for hours at a time. Then, on rare occasions, it will inexplicably come unstuck and go back to normal.
I never see any "scheduled proc took too long" sorts of warnings in the log, so I don't THINK I'm running Tcl code that's tying up the scheduler thread. So far it seems more like the scheduler itself silently just does nothing. Interestingly, when the scheduler comes unstuck, any "missed" ns_schedule_daily jobs suddenly run all at once, in one second! In my case that includes the ns_logroll I schedule daily at 23:59. Just yesterday I saw it finally come unstuck and roll the serer log at 16:45 the next day! So far I've ONLY seen this strange behavior on Windows 7, where I recently upgraded to newer NaviServer code and a newer Microsoft 2019 Visual Studio Community Edition compiler. I suspect the problem doesn't happen on Linux at all, but I haven't checked for that thoroughly. Previously on Windows I was running NaviServer code from c. 2019-07 and an ancient Microsoft compiler from 2010; the problem did NOT happen then. Do you have any advice on how I should go about trying to debug this? E.g., are there commands I should run to check exactly what the scheduler thread is doing at any given time? I've started reading "nsd/tclsched.c" and "nsd/sched.c", but I've never looked at this code before and have no idea what might be wrong. Btw, the sched.c source says it's based on the 1988 paper below, but just how useful is it for understanding things? Those old conference proceedings seem hard to find outside of university libraries. The USENIX website itself says they only have content back to 1993! (Thanks for any help and advice!) https://www.usenix.org/legacy/publications/bibliography/byAuthor.html Author: Ronald E. Barkley Author: T. Paul Lee Title: A Heap-based Callout Implementation to Meet Real-Time Needs Pages: 213-222 Publisher: USENIX Proceedings: USENIX Conference Proceedings Date: Summer 1988 Location: San Francisco Institution: AT&T Information Systems The table of contents (nothing else) is online here: http://www.gbv.de/dms/tib-ub-hannover/303823399.pdf -- Andrew Piskorski <a...@piskorski.com> _______________________________________________ naviserver-devel mailing list naviserver-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/naviserver-devel
