Troubleshooting Oracle monitor connection failures

Troubleshooting Oracle monitor connection failures

Applies to: Oracle Database Server and Oracle RAC monitors. Applies both when adding a monitor and when an existing monitor reports a connection error.

When adding an Oracle Database Server monitor in Applications Manager, you may encounter connection errors that prevent the monitor from being added successfully. These errors may occur due to listener connectivity issues, incorrect connection details, database instance startup or shutdown operations, or network connectivity restrictions.

This article explains the common Oracle connection errors, their possible causes, and the steps to troubleshoot them.

1. Which section do I need?

Applications Manager shows a short message, and the Oracle driver error appears underneath it. The driver error is the one that identifies the problem, so read it first and use this table.

Driver errorMeaningSection
IO Error: The Network Adapter could not establish the connection
ORA-12541
No listenerListener Not Reachable
ORA-01033Instance starting or stoppingOracle Instance Starting or Stopping
ORA-12170TCP connect timeoutTCP Connect Timeout
ORA-01017Wrong user name or passwordInvalid Username or Password
ORA-00020, ORA-00018Process or session limit reachedProcess or Session Limit Reached
ORA-00257Archiver stuck, archive space fullArchiver Is Stuck
ORA-01031Monitoring user lacks a privilegeInvalid Username or Password
ORA-28001Password expiredPassword Has Expired
ORA-28000Account lockedThe Account Is Locked
ORA-12505
ORA-12514
Listener is up but the database is not registeredListener Is Up but the Database Is Not Registered
ORA-12518Listener could not hand off the connectionListener Could Not Hand Off the Connection

If the driver message carries no ORA- code at all, collect it in full and contact support. An unmapped driver error is worth reporting.

2. Listener Not Reachable

Error messageOracle Server is down.

Underlying driver error: IO Error: The Network Adapter could not establish the connection, or ORA-12541, or ORA-12505.

Possible causes:

  • The Oracle listener is not running on the database server.
  • The hostname or IP address configured in Applications Manager is incorrect.
  • The listener is not listening on the configured port.
  • The configured Service Name or SID is incorrect.
  • The database service is not registered with the listener.
  • A firewall or network security policy is blocking the connection.
  • An Oracle Connection Manager or SCAN listener configuration is preventing the connection.

Troubleshooting steps:

  1. Verify that the Oracle listener is running on the database server.
    lsnrctl status
    Verify the listener status, listening endpoints, hostname, port and registered database services in the output.
  2. Verify that the database service is registered with the listener.
    lsnrctl services
    Ensure that the required database service is listed and is available for connections.
  3. Verify the Service Name or SID configured in Applications Manager, and ensure that the connection type selected matches the Oracle connection identifier.
    • Service Name: use the database service name registered with the listener.
    • SID: use the Oracle System Identifier when the connection is configured using SID-based addressing.
    A mismatch between the Service Name and SID configuration may result in connection failures even when the listener is running.
  4. Verify that the hostname and listener port are reachable from the Applications Manager server.
    Windows:  Test-NetConnection <oracle-host> -Port <listener-port>
    Linux: nc -vz <oracle-host> <listener-port>
  5. If the connection is configured using an Oracle SCAN address, verify that the SCAN hostname resolves correctly and that the SCAN listeners are running. For Oracle RAC environments, also verify that the required database service is registered and available through the SCAN listener.
  6. If Oracle Connection Manager is configured, verify that it is running and that the configured connection rules allow connections from the Applications Manager server.
  7. Check whether any firewall or network security policy is blocking communication between the Applications Manager server and the Oracle listener.
  8. If the issue persists, verify the listener log for connection-related errors and check whether the database service is correctly registered with the listener.
If the driver error is ORA-12505, the listener answered but has no service registered under the name given. Work through the steps above, starting with lsnrctl services. If the instance was started only a moment ago, wait up to a minute and retry before changing anything — the listener can take that long to register a newly started instance.

3. Oracle Instance Starting or Stopping

Error messageUnable to connect to Oracle Server.

Underlying driver error: ORA-01033: ORACLE initialization or shutdown in progress

Possible causes:

  • The Oracle database instance is currently starting up or shutting down.
  • The database is undergoing maintenance or restart operations.
  • The database has not completed its initialization process.
  • The database is in MOUNT mode and is not yet open for normal connections.
  • In Oracle Data Guard environments, the standby database may be in a mounted state or undergoing a role transition.

Troubleshooting steps:

  1. Verify whether the Oracle database instance is currently starting or shutting down. Connect with an account that has the required administrative privileges and run:
    SELECT INSTANCE_NAME,
    STATUS,
    DATABASE_STATUS
    FROM V$INSTANCE;
    Check whether the instance is in STARTED, MOUNTED or OPEN state.
  2. Verify the database open mode:
    SELECT NAME,
    OPEN_MODE,
    DATABASE_ROLE
    FROM V$DATABASE;
    Verify whether the database is open and accepting the type of connections Applications Manager requires.
  3. If the database is undergoing startup or shutdown operations, wait until the operation completes and retry adding the monitor.
  4. If the database is in MOUNT mode, verify whether it is intentionally maintained in that state or whether the startup process is incomplete.
  5. For Oracle Data Guard environments, verify the role and open mode of the database. A physical standby database may remain in MOUNT mode or operate in READ ONLY mode, depending on its configuration. Verify that the configured monitoring service is available and permits the required connections.
  6. If the database is expected to be open but remains in a non-open state, check the Oracle alert log for startup, recovery or shutdown-related errors.
  7. Once the database is available and the required service is accepting connections, retry adding the Oracle monitor.
Do not change the standby database open mode solely to resolve a monitoring connection issue. Confirm the intended Data Guard configuration with the database administrator.

4. TCP Connect Timeout

Error messageUnable to connect to Oracle Server.

Underlying driver error: ORA-12170: Cannot connect. TCP connect timeout

Possible causes:

  • The Oracle database host is not reachable from the Applications Manager server.
  • The listener port is not reachable or is not responding within the configured timeout.
  • A firewall or network security device is dropping connection requests.
  • The Oracle listener or Oracle Connection Manager is overloaded.
  • There is a network routing issue between the Applications Manager server and the Oracle database server.
  • The configured hostname resolves to an incorrect or unreachable IP address.

Troubleshooting steps:

  1. Verify that the Oracle database hostname resolves to the expected IP address from the Applications Manager server.
    nslookup <oracle-host>
  2. Verify that the Oracle listener port is reachable from the Applications Manager server.
    Windows:  Test-NetConnection <oracle-host> -Port <listener-port>
    Linux: nc -vz <oracle-host> <listener-port>
  3. Verify that the Oracle listener is running and listening on the configured port.
    lsnrctl status
  4. Check whether any firewall, network security device or routing policy is blocking the connection.
  5. If Oracle Connection Manager is configured, verify that its service is running and that the connection rules permit the connection.
  6. For Oracle RAC environments, verify the SCAN listener availability, SCAN hostname resolution, and connectivity to the listener endpoints returned during connection redirection.
  7. Check the Oracle listener logs for incoming connection requests and timeout-related errors.
  8. If the listener is reachable but the connection continues to time out, check whether the database server or network is experiencing high load or connection delays.
  9. If required, review the connection timeout configuration in Applications Manager and verify whether it is sufficient for the database environment.

5. Password Has Expired

Underlying driver errorORA-28001: the password has expired

The credentials are correct, but Oracle will not accept them until the password is changed. This normally follows a password-lifetime policy expiring.

Troubleshooting steps:

  1. Confirm the account state:
    SELECT username, account_status, expiry_date
    FROM dba_users
    WHERE username = '<monitoring-user>';
    An expired account shows EXPIRED or EXPIRED(GRACE).
  2. Set a new password for the monitoring user, then update it in the monitor's credentials and save.
  3. To stop it recurring, ask the DBA to place the monitoring user in a profile whose PASSWORD_LIFE_TIME is UNLIMITED, or to track the expiry date. A monitoring account that expires silently takes the monitor down with it.

6. Listener Could Not Hand Off the Connection

Underlying driver errorORA-12518: TNS:listener could not hand off client connection

The listener accepted the connection and then failed to pass it to a server process. This is nearly always a resource limit on the database host rather than a configuration mistake.

Troubleshooting steps:

  1. Check whether the instance has reached its process limit:
    SELECT resource_name, current_utilization, max_utilization, limit_value
    FROM v$resource_limit
    WHERE resource_name IN ('processes', 'sessions');
  2. If current_utilization is at limit_value, the instance needs a higher processes setting, or fewer connections. This is a DBA change and requires a restart.
  3. Check the listener log and the alert log for the same timestamp. They usually name the reason the hand-off failed.

7. Oracle RAC Specific Errors

Listener and SCAN connectivity for RAC is covered in the sections above. Three further messages apply only to RAC monitors, and none of them is a connection failure in that sense.

The instance added does not belong to a Oracle RAC database type. Kindly check the details provided.

The instance was reachable and the credentials worked, but it is a single-instance database, not a member of a RAC cluster.

  1. Confirm whether the database is clustered:
    SELECT name, value FROM v$parameter WHERE name = 'cluster_database';
    TRUE means RAC; FALSE means single-instance.
  2. If it is single-instance, add it as an Oracle Database monitor instead of an Oracle RAC monitor.
  3. If it is clustered, check that you entered the SCAN name or a cluster node, and not a stand-alone database on the same host.

Kindly check if this RAC node is part of the cluster.

A node that Applications Manager is monitoring no longer appears in the cluster. Usually the node was evicted, or it was removed from the cluster deliberately.

  1. Ask the DBA for the current node list:
    olsnodes -n -s
  2. If the node was evicted, the cluster logs (ocssd.log) explain why — commonly a network heartbeat or voting disk problem. Resolve that first; the monitor recovers once the node rejoins.
  3. If the node was removed on purpose, delete it from the RAC monitor in Applications Manager so it stops being polled.

Data collection failed due to connectivity issues with the Applications Manager backend database.

This one is not about Oracle. Applications Manager could not reach its own database while storing the collected data, so the data was lost even though the Oracle connection may have been fine.

  1. Check whether other monitors reported problems at the same time. If many did, the backend database is the common cause.
  2. Check <AppManager_Home>/logs/AMErrorLog*.txt around that timestamp for backend database errors.
  3. If Applications Manager uses a remote backend database, check the network between the two, and whether the backend reached its connection limit.

8. The Account Is Locked

Error messageThe account of this user is locked. Please Contact Database Administrator.

Driver error behind it: ORA-28000: the account is locked

What it means: Oracle has locked the account the monitor signs in with. The password may be perfectly correct — this is not an authentication failure. Oracle locks an account for one of two reasons: an administrator locked it deliberately, or the account's profile hit its FAILED_LOGIN_ATTEMPTS limit.

That second case matters here, because Applications Manager polls on a schedule. If the password stored on the monitor is wrong, every poll is another failed login, and the account will lock again shortly after you unlock it. Check the stored password before unlocking.

Troubleshooting steps:

  1. Check the account's current state.
    SELECT username, account_status, lock_date, profile
    FROM dba_users
    WHERE username = UPPER('<monitoring-user>');
    The account_status tells you which case you are in:
    • LOCKED — an administrator locked it.
    • LOCKED(TIMED) — too many failed logins. It unlocks by itself after the profile's PASSWORD_LOCK_TIME.
    • EXPIRED & LOCKED — the password also needs resetting. Do section 5 as well.
  2. Unlock the account.
    ALTER USER <monitoring-user> ACCOUNT UNLOCK;
  3. Open the monitor and use Test Credential straight away. A locked account refuses every password, right or wrong, so this is the first point at which you can tell whether the password saved on the monitor is correct. If it now fails with ORA-01017: invalid username/password, the stored password is wrong: correct it before the next poll locks the account again.
  4. If the account keeps locking and the monitor's password is known good, something else is signing in with stale credentials. Check the profile's limit, then look for other clients using the same account.
    SELECT profile, resource_name, limit
    FROM dba_profiles
    WHERE profile = (SELECT profile FROM dba_users WHERE username = UPPER('<monitoring-user>'))
    AND resource_name IN ('FAILED_LOGIN_ATTEMPTS', 'PASSWORD_LOCK_TIME');
    A dedicated account used only by Applications Manager makes this much easier to diagnose, because any failed login then belongs to the monitor.
  5. Save the monitor again, or wait for the next poll, and confirm it recovers.

If the account must stay locked by policy, create a separate account for monitoring instead. It needs only read access to the data dictionary and the dynamic performance views — not to any application data. The privileges required are listed in the Oracle monitor privileges article.

9. Listener Is Up but the Database Is Not Registered

Error messageThe Instance Name (Service Name) mentioned is wrong or currently inactive. Please configure with proper values.

Underlying driver error: ORA-12505 when the monitor connects by SID, or ORA-12514 when it connects by service name. For example:

ORA-12505: Cannot connect to database. SID orcl is not registered with
the listener at host localhost port 1521.

What it means: this is not the same as section 2. The listener is running, reachable, and answering — it replied in order to tell you that it has no instance registered under the name you asked for. The host, the port and the firewall are all fine. Only the SID or service name is unaccounted for.

There are three common reasons:

  • The database is down. An instance only registers with the listener while it is running, so a stopped database looks exactly like a wrong name.
  • The name is wrong, or the wrong kind of name. A SID and a service name are not interchangeable, and the monitor must be configured for whichever one you entered.
  • The database is still starting. See the note below — this one is easy to misread.
If this appeared during an Oracle restart, it may be temporary. The listener starts accepting connections before the instance registers with it, and registration can lag by up to a minute. During that window every connection gets ORA-12505 or ORA-12514. Once the instance registers, the same connection attempt starts returning ORA-01033 instead, which is section 3. A monitor that polls during a restart can therefore show either error, and seeing this one does not by itself mean anything is misconfigured. Check again after a minute before changing any settings.

Troubleshooting steps:

  1. Ask the listener what it actually knows about. Run this on the database server:
    lsnrctl status
    lsnrctl services
    Each registered database appears in the Services Summary, with a status of READY or BLOCKED. If the name the monitor uses is not in that list, the listener genuinely does not have it, and the message is accurate.
  2. Confirm the database is open.
    SELECT instance_name, status FROM v$instance;
    SELECT name, open_mode FROM v$database;
    STATUS should be OPEN and OPEN_MODE should be READ WRITE. Anything else — MOUNTED, STARTED, or no connection at all — explains the error on its own.
  3. Check the exact names the database publishes, and compare them with the monitor.
    SHOW PARAMETER service_names
    SHOW PARAMETER db_name
    SHOW PARAMETER instance_name
    In Applications Manager, open the monitor and make sure the Service Name / SID field holds one of these, and that the selector beside it matches which kind you entered. A service name typed into a SID field produces this exact error even when the database is perfectly healthy.
  4. If the database is open but still not listed by lsnrctl services, registration itself has failed. Check that LOCAL_LISTENER points at the right address, then force a re-registration:
    SHOW PARAMETER local_listener
    ALTER SYSTEM REGISTER;
    This is also worth doing when the listener was restarted after the database, which leaves it with nothing registered until the next automatic attempt.
  5. If the monitor must keep working while the database is shut down and restarted, ask the DBA to add a static SID_LIST_LISTENER entry in listener.ora. With a static entry the listener accepts the connection and passes it to the instance, so a restarting database reports the more precise ORA-01033 rather than looking unregistered.
  6. Save the monitor again, or wait for the next poll, and confirm it recovers.

10. Additional Troubleshooting

Invalid Username or Password

Underlying driver errorORA-01017: invalid username/password; logon denied

Verify that the username and password configured in Applications Manager are correct. Also ensure that the monitoring user exists in the correct database or container and has the required privileges.

For additional troubleshooting steps, refer to the Oracle monitor privileges article.

Oracle SSL / Certificate Errors

If the connection fails due to SSL or certificate-related errors, verify that the required certificates are correctly configured and trusted by Applications Manager. Also ensure that the SSL configuration in Applications Manager matches the Oracle database server's security configuration.

For additional information, refer to How to Import Certificates for Monitoring Oracle Database with SSL Authentication and Troubleshooting Oracle SSL Connection Errors.

Process or Session Limit Reached

Underlying driver errorORA-00020: maximum number of processes (N) exceeded
ORA-00018: maximum number of sessions exceeded

The instance answered and refused the connection because it has no process or session slot left. Nothing is wrong with the host, the listener or the credentials, and the database itself is still serving the sessions it already has. Applications Manager draws on the same pool as the applications using the database, so this usually appears first during a load peak.

  1. Compare what is in use against the limit.
    SELECT resource_name, current_utilization, max_utilization, limit_value
      FROM v$resource_limit
     WHERE resource_name IN ('processes','sessions');
  2. If max_utilization is at or near limit_value, the instance is sized too small for its workload. Ask the DBA to raise the limit; both parameters need a restart.
  3. ALTER SYSTEM SET processes = <new value> SCOPE=SPFILE;
  4. Raising the monitor's polling interval makes Applications Manager ask for a connection less often. Treat that as a stop-gap: it hides the symptom rather than addressing the limit.

Archiver Is Stuck

Underlying driver errorORA-00257: archiver error. Connect internal only, until freed.

The database is in archivelog mode and the archiver cannot write a new log, almost always because the fast recovery area or the archive destination is full. Until space is freed the instance refuses new connections from ordinary users, so monitoring stops along with everything else. This is a database outage rather than a monitoring problem, and it needs the DBA.

  1. Check how full the recovery area is.
    SELECT name, space_limit, space_used, space_reclaimable
      FROM v$recovery_file_dest;
  2. Ask the DBA to free space: back up and delete archive logs, reclaim space in the recovery area, or raise db_recovery_file_dest_size.
  3. Once the archiver can write again the instance accepts connections and the monitor recovers on its next poll, with no change needed in Applications Manager.

11. When to Contact Applications Manager Support?

If the issue persists after following the troubleshooting steps above, contact the Applications Manager support team with the following details:

  • A screenshot of the Oracle Monitor Addition page.
  • The complete error message displayed while adding the monitor, including the underlying driver exception.
  • The Oracle database version.
  • The Applications Manager build number.
  • The hostname or IP address and listener port configured for the Oracle monitor.
  • The configured connection type (Service Name or SID).
  • The output of the lsnrctl status command.
  • The output of the connectivity test performed from the Applications Manager server.
  • The Oracle listener log and database alert log entries corresponding to the time of the connection failure.
  • The Applications Manager Support Information File (SIF).
When should you provide the Support Information File? If the listener is running, the configured host and port are reachable, the database service is available, and the connection still fails from Applications Manager, provide the SIF along with the relevant Oracle listener and database alert logs. It helps the support team identify the underlying JDBC exception and determine whether the issue originates from Applications Manager, the Oracle JDBC driver, or the database and network configuration.

                    New to ADSelfService Plus?

                      • Related Articles

                      • Troubleshooting IBM i (AS400) Monitor Connection Failures

                        Troubleshooting guide for IBM i (AS400) monitoring in ManageEngine Applications Manager. Use this article to troubleshoot common IBM i monitor failures, including SSL certificate trust, invalid credentials, disabled user profiles and host-server ...
                      • Troubleshooting Oracle SSL Connection Errors

                        Troubleshooting guide for Oracle Database monitoring in ManageEngine Applications Manager. Some Oracle connection errors are caused by TLS configuration or listener endpoint settings rather than by an untrusted certificate. Use the sections below to ...
                      • DNS Monitor - Troubleshooting

                        Common DNS Monitor Errors and Troubleshooting Guide 1. Host Not Found Description: The DNS server was unable to locate the requested lookup address. Possible Causes: This may happen if the hostname is incorrect, the domain does not exist, or there is ...
                      • Real User Monitor (RUM) - Troubleshooting guide

                        If your Real User Monitor has not collected data for an extended period, follow the steps below to troubleshoot the issue. Step 1: Verify the RUM Agent configuration Real User Monitoring requires the RUM Agent to be installed and mapped to ...
                      • APM Insight Java Agent – Onboarding & Troubleshooting Guide

                        This document is a step-by-step guide to onboarding and troubleshooting the APM Insight Java Agent, for both the standalone agent and the FSO Auto Profiler. It covers agent attachment, version compatibility, configuration, connectivity, data ...