How do I determine why a test has the Unknown status?
In Traverse, tests can enter an Unknown test status for several reasons:
- When a new test is created, it is added to the provisioning database and also placed in the appropriate monitor queue to be executed. During this time the web application shows the test in an UNKNOWN state.
- Some tests do a rate calculation ([result1 - result2] / time_elapsed_between_tests), which requires two polled results. For example, most network interface tests (Traffic In/Out, Util In/Out) are in this category. Until the second result is polled, these tests show an UNKNOWN state. If a test is configured for a five-minute polling interval, it remains in an UNKNOWN state for approximately ten minutes, until two results are received and the rate is calculated.
- If the flap-prevention feature is enabled, any test that is in the process of changing its state will show a TRANSIENT state for the configured cycles. For example, if flap-prevention cycle is configured to be 2, and a ping test is configured for a 3 minute interval, when the ping test switches from OK to WARNING, until the test remains in the new state for 2 additional cycles (6 min), the test will be shown in TRANSIENT state.
- If a Traverse process is not running, newly added tests will not return any results and the tests will show an UNKNOWN state. When you drill down into devices with older tests, they will show values under TEST TIME and DURATION columns in a light blue color, indicating outdated results.
- If the DGE component is unable to communicate with the Traverse BVE, newly added tests may remain in UNKNOWN state, with UNKNOWN test time and UNKNOWN duration.
- If a device is not reachable (e.g., it's been turned off or there are network problems, etc.), tests for that device appear in an UNKNOWN state, indicating that no polled value could be retrieved.
- In the case of SNMP tests, if the OID is no longer valid (for example, if the Index has changed), the test appears in an UNKNOWN state, indicating that no polled value could be retrieved.
To see the associated Error message for a test with an Unknown status, navigate to STATUS >> DEVICES and select the 'Hand' Icon to the left of the Test name.
Here are some troubleshooting steps for the most common errors you would see when a test goes Unknown based on the lists provided.
Server returned error: Unable to connect to server: Connection failed: This message states that the DGE/DGEx can no longer connect to the device to poll data. To resolve this, verify the device is online and can connect to the DGE/DGEx that provisions tests on it. If possible, please check the test type agent is running and is configured properly, i.e. WMI service, SNMP agent, etc., if applicable.
Request timed-out while waiting for result: This message states that the DGE/DGEx is not receiving any response from the polled device. To resolve this, please check the test type agent is running and is configured properly, i.e. WMI service, SNMP agent, etc., if applicable.
Invalid SNMP OID: Some OIDs change over time and this can cause SNMP tests to become Unknown. The easiest way to resolve this is to enable test parameter rediscovery on the device with these errors. This will periodically update the tests that are provisioned on the device, and optionally add any new tests, and delete the tests that are no longer valid.
Calculated percentage is over 100%: This means that the result value is over the maximum value configured in the test configuration. You can simply adjust the 'maximum' value in the test configuration page to the appropriate value. If the maximum value of the test is not correct when it is provisioned, please submit a separate ticket so we may validate the device signature.
Timeout waiting for response from device: Same as "Request timed-out while waiting for result".
Server returned error: Unable to add PDH counter: CANNOT FIND OBJECT: This message denotes that the WMI query in the test configuration page either contains a PDH counter that does not exist or that the query is formatted incorrectly. To resolve this, verify that the PDH counter you are trying to monitor exists on that device and that the format of the query is correct.
Cannot use negative value to calculate: This means that the post-processing directive configured on the device results in a negative value. To resolve this, change the post-processing directive to a more appropriate setting.
Polled value X exceeds configured maximum value X: Same as "Calculated percentage is over 100%". A rediscovery is advised so the maximum value is recalculated.
Encountered possible counter rollover or reset: This means that the test value is a counter that will continuously increment and has reached its maximum value and rolled over and reset the value. Traverse can handle one rollover event between consecutive polled values, however subsequent events will generate an Unknown error. For these, please submit a separate ticket as it requires more in depth investigation.
Server returned error: Unable to connect to server: Access is denied: This message means that the credentials provided for the polled test are no longer valid. To resolve this, please verify that the credentials being used are valid and that they are entered correctly in Traverse and exist on the device being polled.
Server returned error: Server did not return any results for query: This message means that the counter being polled does not contain any data. This means that the counter library on the device may be corrupted and that issue would need to be resolved in Windows. If you verify that the counter can be polled successfully outside of Traverse, please submit a separate ticket as it requires more in depth investigation.
Server returned error: Query failed: The object invoked has disconnected from its clients: This message is a DCOM error message from the target machine that would need to be resolved in Windows. If you verify that the counter can be polled successfully outside of Traverse, please submit a separate ticket as it requires more in depth investigation.
doTest was never called for test: For these, please submit a separate ticket as it requires more in depth investigation.
Unable to get test result due to internal error: composite counter (null): For these, please submit a separate ticket as it requires more in depth investigation.
Server returned error: Unable to connect to server: No such interface supported: This means that the interface that this test was provisioned to poll is no longer available. To resolve this, please verify that the interface is still operational on the target device.
All versions of Traverse