This is a fairly major update. You won't notice anything visually per se, but it's a major reliability release for the Event Server service. We fixed some random once a year type bugs, added crash prevention, and we've increased performance by seconds to half-seconds here and there. All-in-all it should feel a bit snappier to use if you've used it for years. If you have any questions about the previous 2030 releases, please review all of the prior release notes.
---
Event Server Service:
- NEW: Added a "Heartbeat Check" entry to the Windows Event Viewer log (written every few minutes) to prove the service is alive and show what it has been doing (Input queue scans, XML files launched, processes running, and the status of your Mail/File/SQL rules); if a problem stays active (like the Input queue being unreachable), it is re-logged with each Heartbeat Check until it clears -- a lasting problem can never scroll out of view. The old per-scan "Reading directory" repeating lines has been removed, too (it flooded the log without really telling you anything other than it was working)
- NEW: Event Server now watches for hung processes; any process it launched (XML batch, RPT, AutoPrint, or a Mail/File/SQL rule run) still running after 23 hours is assumed hung, ended, and noted in the log. You can change the number of hours with the new KillLongRunningHungExesAfterXHours setting under [Event Server] in jnrr.ini (set it to 0 to turn this off); Event Server adds the setting to jnrr.ini automatically if it's missing, so you'll find it there ready to adjust. See the Event Server manual for more information
- NEW: If the same XML file fails to process 3 times, Event Server now moves it to the Error queue automatically (instead of trying it over and over) and notes it in the log; a hung process that had to be ended counts as a failed try
- NEW: File rules and SQL rules now run independently; they still share the same interval setting, but a long-running SQL rule no longer delays your File rules (and vice versa) -- previously neither could run again until BOTH had finished
- NEW: Process "spots" are now reserved for active rules; each active rule type (Mail, File, SQL) holds one of the process spots for itself so the rules can ALWAYS run on schedule, and the XML/RPT/AutoPrint side is capped to fit -- the total never goes over 20 (this keeps our monitoring tools correct, since they track up to 20 processes). See the Event Server manual for more information
- UPDATED: RPT and AutoPrint processing now count toward max batch processes (they "queue up" for a process spot just like XML files do). NOTE: if max batch processes is set to 0, the Input queue is off entirely -- XML, RPT, and AutoPrint files will not be processed (previously RPT and AutoPrint still ran at 0); if you use AutoPrint, make sure max batch processes is set to at least 1
- UPDATED: A momentary problem reaching the AutoPrint folder (like a network blip) no longer turns AutoPrint off until the service is restarted; the scan is suspended with a log entry and automatically resumes when the folder is reachable again (re-checked at each Heartbeat Check)
- UPDATED: A problem writing to the Windows Event Viewer log itself (a full or broken log) will no longer stop jobs from launching; if Windows won't take a log entry, the entry is skipped and processing continues (your jobs matter more than our log entries)
- UPDATED: The Event Server Monitor feels much smoother during heavy activity; status values are now written to disk only when they actually change (far less file churn for the Monitor to fight with), status entries left behind by a previous run are verified at startup and cleared if their process is gone, and any "phantom" entry that slips through is cleaned up automatically (no more phantom running processes in the Monitor)
- UPDATED: Faster Input queue scanning when the queue is holding a lot of files (rewrote the internal file list and sort)
- UPDATED: Log messages were cleaned up throughout to consistently name the "Report Runner Event Server Service" and several confusing messages were reworded and made clearer
- BUG FIX: A rare startup timing issue could leave the service showing "Running" in Windows Services while actually doing nothing at all (no scans, no rules, no log entries) until it was restarted; the internal startup order was corrected (this is the like once-a-year random bug we referred to earlier)
- BUG FIX: Fixed an issue where a failed process launch (for example, the executable briefly locked by antivirus or an update) could permanently strand a process spot or rule until the service was restarted; a failed launch now simply retries
- BUG FIX: Fixed an internal timing issue where scan cycles could overlap under load (which could double-launch the same XML file), and where an error during a scan could disappear without any log entry at all; scans are now strictly one-at-a-time and errors are always logged
- BUG FIX: Fixed an issue where a persistent Input directory error (like a network share going offline) would log the same error once per second for the life of the outage AND could slowly corrupt the internal free-process count until the service was restarted
- BUG FIX: The Mail/File/SQL rules now keep running on schedule even while the Input queue directory is unreachable or a long batch process is finishing up; previously either situation could hold the rules until it cleared
- BUG FIX: Fixed an issue where failure counts against an XML file could be inherited by a NEW file arriving later with the same name; also fixed a rare case where one process exit could be counted twice
- BUG FIX: If the service can't start properly (for example, a bad value in jnrr.ini, which should never happen), it now stops with a clear log entry explaining why, instead of showing "Running" while doing nothing
- BUG FIX: A negative Mail or File/SQL rule interval in jnrr.ini (which should never happen) is now treated as disabled (with a log warning) instead of running those rules back-to-back nonstop
---
Event Server Controller (the [E] icon in the System Tray):
- NEW: When starting, stopping, pausing, or resuming the service, the controller now waits up to 10 seconds for the service to respond; if the service does not reach the expected state, a message is displayed showing the current status (previously a failed start could go unreported)
- NEW: Starting the controller when it is already running now shows a System Tray balloon message ("The controller is already running in the System Tray") instead of doing nothing
- UPDATED: Service error messages now match the actual cause of the error; the Run as Administrator guidance is only shown when the error really is a permissions error, and other failed starts now point to checking the JNRRB_ES Log file (make sure it is set to overwrite itself as needed)
- UPDATED: While the Event Server Monitor is open, the controller's right-click menu options are temporarily disabled; they are available again when the monitor is closed. The service commands are also disabled while the service is mid-transition (starting, stopping, pausing, or resuming) so a second command cannot be issued at the wrong time
- BUG FIX: The controller no longer crashes when it cannot read the service status (usually caused by running/installing without proper Administrator permissions); the tray icon now shows a status of "Unavailable" and the controller keeps checking in the background, returning to the normal status display on its own as soon as the status can be read again
---
Event Server Monitor (started from the Controller):
- NEW: If a queue directory cannot be read (for example, a network share that goes offline), the monitor no longer stops working; the queue caption shows "(Unavailable)" and the display recovers on its own when the directory can be read again
- UPDATED: The monitor now keeps your place while it refreshes; the queue lists no longer rebuild (losing your scroll position and selection) unless something actually changed
- UPDATED: Large queues refresh much faster; the report names shown next to the XML files are now cached and only re-read when a file changes, instead of every XML file being re-read on every 2-second refresh
- UPDATED: Double-clicking an XML file to find its log now searches for the complete file name including the .xml extension (so Batch_1.xml can no longer match log entries for Batch_10.xml) and opens the most recent log file that mentions it
- UPDATED: When the Processed queue holds more than 100 files, the caption now shows the total count, e.g. "Processed 2500 (Last 100 Shown)"
- UPDATED: Clearing a queue now takes its snapshot of files before the confirmation prompt is shown, so XML files arriving while the prompt is open are not deleted
- UPDATED: The File/SQL and Mail rule labels now show green only when the interval setting is greater than zero, matching how Event Server itself decides whether the rules run
- UPDATED: The Open Services and Open Event Viewer buttons now launch services.msc and eventvwr.msc using their full Windows system paths
- UPDATED: Many internal optimizations and cleanup throughout the controller and monitor (fewer disk reads per refresh, legacy code removed, smaller and faster executable)
---
Downloads page:
https://reportrunner.com/downloads
For those with a recent Event Server installation, you ONLY need to update the executables (there's an option for that)
