Skip to content
Documentation

Diagnose Frontend Server Log

You can view the Frontend server log files through the PerformanceGuard web interface: select ADMINISTRATION > Status > System Status, select the Log Files tab, and then select the pgfrontend.log file in the collector/logs folder. Alternatively, you can look directly in the [PerformanceGuard installation folder]\collector\logs folder. The current log is pgfrontend.log. When it reaches 10 MB, it’s renamed to pgfrontend.1.log, and so on, up to 10 older files.

When you analyze the Frontend server log, you may find it helpful to know some log messages that relate to typical problems, and what you can do to solve those problems.

Repeated communication errors

Section titled “Repeated communication errors”

To keep the log readable, the Frontend server doesn’t log every communication error with agents. It logs the first occurrence of an error and then every 100th occurrence of the same error, in this form:

WARN  ... - Experienced the following error <count> times: <error>

The count tells you how often the error has occurred since the Frontend server started. Typical errors are described in the following sections.

Read timed out

Section titled “Read timed out”

An agent didn’t send data to the Frontend server within the Frontend server’s socket timeout. If you get many of these warnings, the Frontend server may be overloaded, or you may have agents at very remote locations. In that case, you may want to increase the socket timeout: in the PerformanceGuard web interface, select ADMINISTRATION > Setup > Parameters, select the Frontend tab, and change Socket Timeout. The value is in milliseconds.

Connection reset

Section titled “Connection reset”

An agent closed the connection while the Frontend server tried to read data from it. It may indicate that the Frontend server is overloaded, or that the agent’s socket timeout value is too small. You’ll typically need to increase the agent’s socket timeout: select ADMINISTRATION > Agent Configuration > Configurations, click the Edit link for the required configuration group, scroll down to Agent Service Parameters, and increase Socket timeout (ms). Then click Save Configuration.

Unexpected EOF on input

Section titled “Unexpected EOF on input”

A communication error occurred while an agent sent data to the Frontend server. EOF means end-of-file: the connection ended before all the expected data was received. Occasional messages are normal, for example when a computer shuts down or loses its network connection during a delivery.

Could not read agent data

Section titled “Could not read agent data”

The Frontend server received data on the agent port that it couldn’t interpret as agent data. The message includes the IP address of the sender. If the sender isn’t a PerformanceGuard agent, another program, for example a network scanner, is connecting to the agent port.

Discarding line: Timestamp too Late

Section titled “Discarding line: Timestamp too Late”

The Frontend server discards measurements that arrive much later than they were made. A measurement is discarded when it’s older than 40 times the agent’s delivery interval. This prevents an agent from reporting old data, for example after a computer was hibernated and then woken up at another location.

Discarding line: Timestamp is in the future

Section titled “Discarding line: Timestamp is in the future”

The agent reported measurements with a time that’s later than the server’s time. The Frontend server discards the measurements and makes the agent synchronize its clock with the server. If the message occurs often for the same agent, check the time settings of that computer.