Troubleshooting

The following list is frequently asked questions about running a collector as a service.

Can all collectors be run as a Windows service? If not, which ones cannot?

The OPC Collector, File collector, Simulation Collector, Calculation collector, and Server-to- Server Collector can be run as services. The iFIX collector run as a background task and cannot be run as a service.

Can all collectors be run as an application? If not, which ones cannot?

All collectors can run as applications (console programs). This includes the Simulation Collector. To make a collector run as a console program, pass a RUNASDOS command line parameter.

What does "running as a service" mean?

It means that the collector appears in the Control Panel list of services. It can run at system boot or be run with a different username and password from the currently logged-in user.

How can the iFIX collector be set up to run when no one is logged in?

It can be set up to run without a user login by adding it to the iFIX SCU task list as a background task and by configuring iFIX to continue running after logoff in local startup.

How do you shut down a collector running as console application?

Collectors started as console applications should be shut down by typing S at the command prompt in the DOS window and pressing Enter.

Can a collector be run as a Windows service and then stopped and restarted?

Yes. Collectors that can run as a service can be stopped and started in Control Panel Services. They can be paused/resumed through Historian Administrator.

What is the difference between running a collector to start as a service on boot up using the Services applet in Control Panel versus running iFIX as a service, which starts the collector through the startup task configuration in the SCU?

The collectors that can run as a service would not be started from iFIX. They can be started from the Control Panel start at system boot. Although you cannot run an iFIX collector as a service, you can log off and on while it is running.

Configure buffer flush speed with the BufferFlushMultiplier key

Store-and-forward buffering is a key feature of Historian collectors. It prevents data loss during planned or unplanned network outages between a collector and Historian server.

If the collector is disconnected from the archiver for several hours or days, many megabytes of data can be buffered and must be delivered by the collector to the Data Archiver upon reconnect. Since all data is sent from the collector to the archiver in time order, the design goal has been to catch up to real time as quickly as possible by sending data as fast as possible.

If this is not the desired behavior because you want to limit the network load on a slow, shared Wide Area Network (WAN) or you want to limit the CPU load on the Data Archiver caused by the incoming data, you can configure the collector to throttle the data it is sending. Throttling Store and Forward does not affect the Alarms and Events collector.

Configuring the throttle is easy, but it requires modifying a registry key, so it should be done with caution.

A DWORD registry key called BufferFlushMultiplier is present under each collector. For Windows 32-bit, it is located under HKey_Local_Machine\Software\Intellution, Inc.\iHistorian\Services\YOUR_COLLECTOR_TYPE. For Windows 64-bit, it is located under HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Intellution, Inc.\iHistorian\Services\YOUR_COLLECTOR_TYPE.

• To slow the store and forward throttling, set the value of BufferFlushMultiplier to 2. The 2 means that the collector should never send data at more than 2 times its normal rate to limit network and CPU load.
• To disable throttling, set the value of BufferFlushMultiplier to 0 or delete the registry key.

Control archiver speed with the NumIntervalsFlush registry key

The NumIntervalsFlush registry key controls how quickly the collector sends data to the archiver. The collector collects from the data source at the user configured rate, but for efficiency it bundles data samples in a single write to archive. By default, the collector will send data to archiver every 2 seconds or 10,000 samples, whichever happens first. Most often, it sends every 2 seconds because the collector is not collecting that many samples that fast.

If you need collected data sent to archiver right away, so that it is available for retrieval or for calculations, use the NumIntervalsFlush registry key.

You will have to create the registry key, as it does not exist by default. Create a DWORD value called NumIntervalsFlush under the collector, in the same place as HISTORIANNODENAME and INTERFACENAME. On a 64-bit Windows Operating System, all 32-bit component-related registry keys (such as collectors, Client Tools, and APIs) will be located under HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Intellution, Inc.\iHistorian\.

You can configure inactive server connections to reset automatically with the SocketRecvTimeOut registry key. SocketRecvTimeOut configures a timeout that forces the connection to drop and re-establish if no data is received during the specified time. Consider this configuration when your collector goes to status Unknown for long periods of time even when the connection between collector and archiver is good.

Create a DWORD registry key SocketRecvTimeOut under the collector where the problem is occurring and set to a value greater than 90 seconds. A typical value would be 300 seconds. If no bytes are received by the collector for 300 seconds, then the network connection will be closed and re-established.

The ABR and the ABC drivers are OPC v2.0 compliant. All other Intellution 7.x drivers, including the MB1, support OPC v1.0 compliance.

Version 7.x drivers also comply with the OLE for Process Control (OPC) v1.0a standard. Any 1.0-compliant OPC client application can access process hardware data through the I/O Server.

The ABC I/O Server also supports the v2.0 standard and with the OPC Alarms and Events v1.0 specification. Refer to Using OLE for Process Control (OPC) Functionality to learn more about the advantages of OPC.

The following is a table of error messages, possible causes, and recommended corrective actions for failed logins sometimes experienced with Historian Administrator.

Troubleshooting Data Collector configuration and/or performance requires a thorough understanding of how Historian works and how the various parameters affect system operation. Armed with this knowledge, you can usually localize a problem to one or more functions or parameter settings and take effective corrective action.

Historian offers several tools to help find the cause of an operating problem.

LOG files

The system creates a new log file each time an archiver or collector is started. You can open these files in Notepad or another text editor. The -nn suffix in the file name indicates the place of each log file in the time sequence.

• The data collector log files, located in the LogFiles folder within the Historian program folder, are a historical journal of every event affecting operation of the collector.
• The DataArchiver-nn.LOG files are sequential files for the archiver only.
• The iFIX collector-nn.LOG files contain performance information on iFIX collector functioning.

SHW file

The Data Collector .SHW file shows configuration data for collectors and is also located in the LogFiles directory under Historian. Verify that the parameter and configuration settings match what you configured. You can open this file in Notepad or another text editor.

Collector Maintenance Performance Indicators

These performance and status indicators can be a major aid in identifying, localizing, and diagnosing a problem with a collector.

Collector Maintenance pages

The Collector Maintenance pages can provide useful information about settings and selections of various options and parameter values. Examine each field and verify that it is appropriate for the current application.

iHistorianSDKerrors.log

When performing any troubleshooting of the File collector Administrator or SDK program you should examine the iHistorianSDKerrors.log file, which is located in the LogFiles folder in your Historian program folder.

Tag Configuration

To diagnose a problem in tag configuration, examine the .LOG and .SHW files in the Historian/Logfiles directory. Since these files are a journal record of all system events and parameter modifications important to a system administrator, they can be helpful in identifying and localizing the source of a system malfunction or data error.

You may also find the Windows Event Viewer helpful in troubleshooting. For more information on the Windows Event Viewer, refer to Review System Alerts and Messages.

When troubleshooting SDK applications, such as Historian Administrator, the Excel Add-In, or File collector, examine the iHistorianSDKerrors.log file, located in the LogFiles folder in the Historian program folder.

Stale Tags

If you are not seeing all stale tags in the Historian Web Admin, the ClientManager service may be down. If so, the thread that processes stale tags is suspended until connection is restored.

If the connection to the ClientManager service is lost while processing a batch of 100 tags, that set of tags is skipped and a Tag Property Update error is logged. When the ClientManager service is back up again, the next set of 100 tags is considered for scanning. The skipped tags must wait until the Diagnostic Manager???s next cycle. Until then, they are not marked as stale or not stale.

Alternately, the DataArchiver or ConfigManager services may be down. If so, all the remaining tags are skipped, and a DataOpenRecordset/Tag Update failure is logged. Even though the thread that processes stale tags is not suspended, all the remaining tags have to wait until the Diagnostic Manager???s next cycle. Until then, they are not marked as stale or not stale.

There are many parameters that affect Historian Server performance and scalability. This topic emphasizes the parameters that help improve performance and scalability.

Buffer Memory Max

Buffer Memory Max allows you to specify the maximum memory buffer size that an archiver queue can take. Use this parameter to control memory allocation to the write queues. If you plan to collect large amounts of data with high update rates, increase the Buffer Memory Max size. Increasing the Buffer Memory Max size provides enough memory for write queues, thus increasing the Historian performance. Refer to The Global Options Section for more information.

Performance Counters

System File Cache Tuning

Archiver cannot create a new archive: If an archive is full and you have disabled Overwrite Old Archives in the Global Options section, Historian can only create a new archive if there is enough disk space available. If you do not have enough disk space to create a new archive, Historian does not generate an alert on the Main page, but instead adds the following messages to the Data Archiver log file: Could not locate empty archive file, nor create one, nor use old one. Stopped processing data write requests. Free up disk space or enable Overwrite Old Archives to continue archiving your data.

Archive service not starting automatically: On some systems the Archiver does not start automatically after install. Start the Archiver manually through the Services window or restart the computer. The Archiver should start on all future restarts. The Archiver service may also fail to start if it is started by an improperly configured domain account. In order to add a user from the domain as the Archiver service start up account, the machine Historian is installed on must have joined the domain. The user changing the service log on account must also have administrative rights to the Historian machine. This user the service is configured with must be a member of the Domain Administrators group and have administrative rights to the Historian machine. The user also needs to be allowed to log on as a service and act as part of the operating system. This is configured in the local policy editor of the Historian machine.

Investigate failed backups: Refer to the ArchiveBackup-XX.log file to investigate failed backups. The backup can fail if the user does not have valid permissions or if the archive name supplied is not valid. For more information, refer to Knowledge Base article: i014491 at http://iglobalcare.gefanucautomation.com.