Early Release Collector – 22.050

Early Release Collector 22.050 is available starting June 29, 2016.

Note that this Collector version has a known issue with Internal Service Checks for private IPv4 address spaces (i.e. 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16).  As a workaround, you can add the following line to the Collector’s agent.conf configuration file:

enable.ip.blacklist=false

We plan to release a new EA Collector version with a fix for this issue.

Early Release Collector 22.050 includes the following features and fixes.

New Features:

  • This collector version includes support for LogicMonitor’s new Internal Service Checks.  See our release notes and detailed documentation for more information on this new feature.
  • This Collector version includes the ability to cache Collector Data to disk.  Our Collectors have always been able to cache small amounts of data in memory, but not always enough to avoid gaps in data if you were working on your network. Your Collector can now store data in your environment during periods where your collector is unable to deliver data to your LogicMonitor account (e.g. as a result of network issues). Once your collector can reach your account again, the buffered data will be communicated to our servers, eliminating any gaps in data you would have otherwise seen. Currently, a maximum of 30 minutes of data can be cached. For an average monitoring load, we estimate this to consume a couple hundred MB of both disk space and memory. Note that this feature will be enabled by default, but you can disable it by setting reporter.persistent.enable=false in the Collector’s agent.conf configuration file.

Improvements:

  • Upgraded vijava to yavijava 6.0, which includes support for VSphere 6.
  • Previously, the Script Eventsource collection method didn’t accept JSON formatted events that contained more than one level of nested objects. We now accept nested JSON objects and convert the output to string format.
  • Previously, log files being monitored via Eventsource couldn’t be moved or renamed. This is no longer the case, and you can now rename and/or move monitored log files.
  • The Collector previously sent one ping packet to each device during a Netscan, where the device was discovered if 100% of that packet was returned. If that one packet was lost, this sometimes resulted in Netscan missing devices that should be discovered. We now send three ping packets to each device during Netscan, and discover the device if one of those three packets is returned.
  • We’ve changed how we store WMI discovery task history. The previous implementation was using a large amount of memory in some cases, causing the collector to restart periodically.
  • Upgraded the following external jars: httpclient-4.5.2.jar, commons-logging-1.2.jar, httpcore-4.4.4.jar, commons-codec-1.9.jar, httpcore-nio-4.4.4.jar and httpasyncclient-4.1.1.jar

Bug Fixes:

  • The !dir debug command wasn’t working well – we’ve improved this command.
  • The !tlist debug command didn’t return all Eventsource data collection tasks in some cases.
  • We were not always correctly identifying when a LogicMonitor Collector was installed on IPv6 Windows devices.
  • The LogicMonitor Agent service sometimes failed to start if there were spaces in the path.  We’ve improved how the service handles spaces to correct this. 
  • The !snmpget and !snmpwalk debug commands weren’t using credentials included in the commands over credentials set as device properties. Credentials included in these commands are now used, even if they differ from the credentials set for your device. This enables you to test SNMP credentials without changing device properties.
  • Data values exceeding 32bits were reported as No Data. This has been corrected.
  • HTTP Active Discovery configured to use SSL wasn’t working correctly if nothing was specified in the Port # List field. If use SSL is selected and no ports are specified, we will now default to port 443.
  • Linux Collector services were not using the correct temp directory, which caused issues during startup in certain cases. This has been fixed.
  • Uploading collector logs that contain spaces previously failed.
  • For datasources using the BatchScript collection method, restarting the associated collector resulted in NaN data for the first data collection task.
  • The !account collector debug command is only supported for Windows collectors. We previously returned non-sensical results if it was run on a Linux collector. We now display no results when !account is run on a Linux collector.

In this Article: