A PEGA0055 alert occurs when the clock drift in any of the participating nodes in a multi-node cluster exceeds the configured time threshold.
In most of the hardware platforms, clocks run at different rates, depending on CPU loading and other internal factors. Usually, a network time protocol service or daemon runs to ensure that the clocks are corrected by means of a standard external internet time service, for example, nist.gov.
In Pega 7, the clock drift is measured against the average time of the clocks in the cluster. Once the alert is triggered, an error is reported to the log file in the node that detected the issue.
The alert log message indicates that the clock drift in one of the cluster's nodes has exceeded the configured time threshold in seconds, and then displays the number of seconds by which the clock has drifted from the average time of the clocks in the cluster, as in the following example:
This node's [<node-identifier>] clock has drifted beyond configured threshold [10 sec]; the actual delta is 10.25 seconds.
The threshold is established in the following prconfig.xml setting:
<env name="alerts/cluster/clockdeltathreshold" value="10" />
A large number of PEGA0055 messages in your alert log indicates a problem with the configuration of the system clock on the server(s), resulting in a significant clock drift. This could be due to the failure of the network time protocol to synchronize the clocks.
You can check the following:
Server configuration: A heavy CPU load on the server(s) can cause the clock to drift excessively.
Excessive load on the server(s): This can be caused by transient loads that drift the clock slowly. A correctly configured NTP daemon corrects this problem. Check the server(s) being reported for the root cause of the clock drift.
|
|
alert log, prconfig.xml file |
|
|
Understanding alerts |
SysAdmin category