Troubleshooting

Start with the Connector’s log, the newest CiscoUCCEConnectorService - yyyy-MM-dd HH.mm.ss.txt in C:\ProgramData\CallRex\Logs (see Logging). Its warnings and errors usually name the problem and its remedy; the table under Log messages explains them. Restart the QMS Cisco UCCE Connector Service after changing its configuration file.

Common problems

The CTI session never opens

The log repeats connection attempts, retry waits, or "CTI server REFUSED the session".

  • Check CTIServer.IP-Address, and that the port is the CTI server’s: 12028 for UCCX, 42027 for UCCE unless the site uses another.

  • Check that the product, CtiProtocolVersion and the client credentials suit the CTI server; choosing the product in the Configuration Tool sets them.

  • On a UCCX cluster, a "standby node" warning means the address is the standby’s; give both nodes' addresses.

The Connector Trigger Rules section offers no events

It shows The connector has not yet reported the events and parameters it can offer. The Connector reports them when its CTI session first opens, so it has not yet connected. Fix the session first, then reopen General Settings.

Calls are recorded but get no flags

  • Check that a trigger rule exists and is enabled for the event and parameter that carry the value, and for the tenant of the agent’s QMS user.

  • Check that the agent’s phone extension matches a QMS user’s extension. If the log says that a value "does NOT map to a userId", the user mapping found no user: see Mapping events to users. On UCCX, an EventToUserMap written for UCCE is a common cause; the Connector warns about it at startup.

  • Check that the contact center sets the value on the call: on UCCX, in the script’s Set Enterprise Call Info step.

  • After a QMS upgrade on several servers, see Upgrade: a Call Recording Service not yet upgraded can return no system flags until it restarts.

Agents are not associated with their devices

Check that Enable Hot-Desking Support is selected under Contact Center Integration, that the ConnectorMode includes AgentState (AgentState or Normal), and that each agent’s QMS agent ID matches the contact center login ID. See Agent State Synchronization.

More detail is needed

Set Log.FileLogLevel to Debug (the installed level is Info) and, to see the raw CTI messages, CiscoUCCEConnectorService.LogMsgHexOnProcessing to true; restart the service. See Logging settings.

Log messages

Messages this release adds, as they appear in the Connector’s log (abbreviated). The log is described in Logging.

Message Meaning and remedy

The flag connector product in QMS is X, but the appSettings ConnectorKind is Y. Adopting X

QMS and the configuration name different products; the Connector uses the one chosen in QMS. Set ConnectorKind to match. "Y (not set; the default)" means the configuration has no ConnectorKind.

n attempt(s) to open the CTI session have failed since the last OPEN_CONF …​ Waiting s second(s)

The CTI server does not accept the session. Check the product, CtiProtocolVersion and the client credentials; choosing the product in the Configuration Tool sets them. The preceding "OPEN_REQ message(s) have been sent with no OPEN_CONF" or "CTI server REFUSED the session" message shows the values sent.

The CTI server is a standby node (…​ E_CTI_SERVER_NOT_MASTER)

A UCCX standby node refused the session. With two addresses, the Connector moves to the other ("Trying CTI server B instead"). With one, add the other node’s address as the second address in CTIServer.IP-Address.

No reply to OPEN_REQ (InvokeID n) within 15 seconds; resetting the connection

The CTI server did not answer the session request. See the retry-wait message above.

No HEARTBEAT_CONF from <address> within n seconds: n of 3 HEARTBEAT_REQs unanswered

The CTI server did not answer a heartbeat in time. One may be a network delay; if a third goes unanswered, the next message reports the session failed.

No HEARTBEAT_CONF from <address> for the last 3 HEARTBEAT_REQs (n seconds): the CTI session has failed. Resetting the connection

The connection was lost without TCP noticing, or the CTI server has stopped responding. The Connector reconnects, or moves to the other CTI server when two are configured. If it recurs with a healthy CTI server, check the network path; a longer CtiHeartbeatIntervalSeconds tolerates longer delays.

The CTI server answered HEARTBEAT_REQ (InvokeID n) with FAILURE_CONF, status X: the CTI session has failed

The CTI server has ended the session. The Connector resets the connection and reconnects.

CtiHeartbeatIntervalSeconds: n is outside the supported range (0, or 5-29); using m

Set a value from 5 to 29, or 0 for no heartbeats.

Could not connect to <address>; retrying in 10 seconds

Only one CTI server address is configured, and the server is not accepting connections. The Connector keeps retrying.

Could not start a connection to CTI server A (<address>); retrying in 10 seconds

The connection could not even be attempted, typically because a host name in CTIServer.IP-Address does not resolve: the preceding error says why. Check the address, or DNS. The Connector keeps retrying, with the other CTI server when two are configured.

EventToUserMap: configured with the Y default rules, but ConnectorKind is X

The user mapping was written for the other product, so calls may get no flags. Remove EventToUserMap and restart.

EventToUserMap is configured in appSettings (n rule(s)), so it is kept

The Connector adopted a product from QMS and kept the configured user mapping. If it was written for the other product, remove it and restart.

Not configured in appSettings, so set from the Y defaults at startup, and not changed: …​

A CTI session setting came from the other product’s defaults. Set it, or choose the product in the Configuration Tool, and restart.

The CTI session was opened as Y; restart the connector so that it is opened as X

The product changed while the Connector was running.

n of the m appSettings FlagTrigger rule(s) no longer apply

QMS has trigger rules, so these FlagTrigger rules are not applied. Re-enter any that are still wanted in QMS.

Rule …​ skipped: that flag is not among the n system flag(s) queried from QMS

The rule names a flag QMS does not have as a system flag.

n of m rule(s) reference a flag that no longer exists and will be skipped

A rule in QMS has lost its flag (flag missing in General Settings).

Could not publish the parameter catalogue …​ ActionNotSupportedException

The Recording Service is older than the Connector; see Upgrade.

Poll failed (…​). Rules unchanged; will retry

Logged at Debug. The Recording Service is unreachable or older; the Connector keeps its current rules.