Troubleshoot

Learn to fix Oracle Trace File Analyzer and Compliance Framework (Oracle ORAchk and Oracle EXAchk) issues.

Cluster Nodes are Not Showing As One Cluster When Viewed by Running the tfactl status Command

The cluster appears fragmented because certificates are not synchronized between nodes. As root on any cluster node, run $GIHOME/tfa/nodename/tfa_home/bin/synctfanodes.sh to synchronize the keys; the script uses SSH and SCP and may prompt for root passwords unless passwordless SSH or Expect is available.

Oracle Trace File Analyzer is Not Starting and the init.tfa script is Missing After Reboot

After reboot, TFA may fail to start and its init.tfa script may be missing when the file system containing TFA_HOME was unavailable, slow to mount, inaccessible through NFS, or affected by maintenance or file-system errors. Refer to My Oracle Support note 2224163.1 for corrective steps.

Error Message Similar to "Can't locate **** in @inc (@inc contains:....)"

Errors such as Can't locate ... in @inc can result from using Perl older than version 5.10, which TFA requires. Upgrade Perl, update the Perl path in tfa_home/tfa_setup.txt, or specify the Perl installation directory with the -perlhome option during installation; the selected directory must contain an executable /bin/perl owned by root when installed by root.

Non-Release Update Revisions (RURs) Oracle Trace File Analyzer Patching Fails on Remote Nodes

Remote TFA nodes can fail to upgrade because of socket issues, leaving inconsistent versions, build IDs, or upgrade statuses. Install the TFA package locally on failed nodes, then copy the root SSL files server.jks, client.jks, and internal/ssl.properties from the configuration node with root ownership and mode 700 to restore synchronized communication.

Non-Root Access is Not Enabled After Installation

Non-root access for the Oracle Grid Infrastructure software owner may not be enabled after installation. As root, run tfactl access add -user username, such as tfactl access add -user xyx, to grant the specified user group access to TFA.

TFA_HOME and Repository Locations are Moved After Patching or Upgrade

Patching can relocate TFA_HOME into GRID_HOME and move the repository to the Grid Infrastructure owner’s ORACLE_BASE, potentially losing a customized repository location. Reset the repository path with tfactl set repositorydir and adjust its maximum size with tfactl set reposizeMB; newer releases migrate or upgrade TFA depending on its current location and patch version.

Oracle Trace File Analyzer Fails with TFA-00103 After Applying the July 2015 Release Update Revision (RUR) or Later

TFA upgrades use a two-phase process: software is installed locally on every node, then new SSL keys and the secured communication model are deployed after all nodes upgrade, normally within 24 hours. If TFA-00103 persists, verify matching versions and build IDs with tfactl print status, confirm registered directories and synchronized root-owned SSL files, and run synctfanodes.sh from a node completing phase 2 to force certificate generation and synchronization for nodes upgraded later or excluded during the waiting period.

OSWatcher Parameters are Different After a Reboot or Otherwise Unexpectedly Different

When TFA manages OSWatcher, startup parameters are persisted per user in newer releases, with normal operation using the oracle or grid settings after reboot. Avoid running OSWatcher as root; if settings are incorrect, stop it and restart it as the non-privileged owner with the intended or default parameters, such as tfactl start oswbb, unless OSWatcher is independently installed and not managed by TFA.

Oracle Trace File Analyzer Installation or Oracle Trace File Analyzer Discovery (tfactl rediscover) Fails on Linux 7

Missing Perl modules such as Data::Dumper during TFA installation or tfactl rediscover on Linux 7 are associated with known bugs fixed in TFA 12.1.2.6.4. Link the operating system Perl command to the Perl version supplied in GRID_HOME to work around the issue.

OSWatcher Analyzer Fails When OSWatcher is Not Running from the TFA_HOME

OSWatcher Analyzer commands work only when OSWatcher runs from TFA’s managed installation under TFA_HOME. Stop and disable any OSWatcher instance running elsewhere, then start OSWatcher through TFA.

Oracle Trace File Analyzer Fails to Start with com.sleepycat.je.EnvironmentLockedException Java Exception

A com.sleepycat.je.EnvironmentLockedException indicates that TFA’s Berkeley DB lock file is already held or stale. Check the lock with fuser, terminate any process using it, remove je.lck from the Berkeley DB directory, and restart TFA as root.

Oracle Trace File Analyzer Startup Fails When Solution-Soft Time Machine Software is Installed, but Not Running on the System

TFA startup can fail when Solution-Soft Time Machine is installed even if it is not running. Uninstall the Time Machine software to resolve the startup conflict.

Non-privileged User is Not Able to Run tfactl Commands

Use tfactl access lsusers as root to verify that the non-privileged user is listed and has an Allowed status. Add missing users with tfactl access add -user username; if access is disabled or the issue remains, run tfactl diagnosetfa -local and provide the resulting diagnostic file to Oracle Support.

Oracle Trace File Analyzer Daemon is Not Starting or Not Running?

Errors TFA-00001 and TFA-00002 indicate that the Java-based TFA daemon did not start or is stopped. Check for running processes, confirm the state with tfactl print status, restart it as root with tfactl start, and run tfactl diagnosetfa -local for support analysis if startup still fails.

Oracle Trace File Analyzer Is Not Collecting Diagnostic Traces of Components Such As CRS, DB, ASM, and So On

If TFA cannot collect CRS, database, ASM, or other component traces, verify their Grid Infrastructure base and diagnostic directories with tfactl print directories. Add required subdirectories using tfactl directory add; use -force only when necessary because AHF may reject standard directories it considers invalid.

Oracle Trace File Analyzer Fails to Start

TFA cannot start when its reserved ports 5000 through 5005 are already in use, producing a Java Address already in use or BindException error. Set an available port in data_dir/internal/port.txt and restart TFA with tfactl start.

TFACTL Collection Errors: "METADATA JSON FILE ALREADY EXISTS" in tfa_main.trc

Simultaneous manual collections can fail with METADATA JSON FILE ALREADY EXISTS because processes generate identically tagged metadata files within the same second. Run concurrent collections with unique -tag values so each collection directory and metadata file has a distinct name.

How to Troubleshoot Oracle Orachk and Oracle Exachk Issues

Troubleshoot Orachk and Exachk by selecting the correct tool, confirming it is current with the -v option, reviewing applicable FAQ and support information, and examining error and execution logs. If needed, enable debug output, collect the resulting ZIP and logs, open a support request, and attach the diagnostic materials.

How to Capture Debug Output

Reproduce the issue with the smallest practical check scope, then run Orachk or Exachk with -debug and optionally specify a module such as setup, discovery, execution, or output. For daemon-based runs, restart the daemon with -d start -debug; collect the generated debug logs, temporary output, and resulting ZIP file for analysis.

Data Entry Terminal Considerations

Run Orachk and Exachk from supported UNIX or Linux terminals, recognizing that dropped connections can interrupt interactive prompts and leave processes hung. VNC generally minimizes network disruption; if accessibility software or terminal inactivity controls suspend sessions, adjust the relevant operating-system timeouts before running the tools.

Tool Runs without Producing Files

Orachk and Exachk normally create temporary and output files, but cleanup occurs after cancellation or error. If a completed run produces no output, rerun the tool in debug mode and inspect the results; contact Oracle Support if the problem continues.

Messages similar to “line : Killed $perl_cmd 2>> $ERRFIL?”

Messages stating that a Perl command was killed generally indicate that the built-in watchdog terminated a command exceeding its timeout, typically to prevent a hung process. Investigate the slow or blocked operation and use timeout troubleshooting procedures where appropriate.

Messages similar to “RC-001- Unable to read driver files”

Driver-file read errors can result from an unsupported platform, insufficient temporary or working-directory space, missing installation support files, or inadequate read/write permissions. Confirm platform support, provide sufficient disk space, ensure the hidden .cgrep directory exists, and verify access to the temporary and working directories.

Messages similar to “There are prompts in user profile on [hostname] which will cause issues in [tool] successful execution”

Orachk and Exachk stop when user profiles contain interactive prompts or commands such as read -p, because remote executions cannot answer them and eventually time out. Temporarily comment out profile prompts on all relevant nodes and rerun the tools.

Messages indicating missing SSH or SCP files point to remote-login configuration problems. Verify the executables and configured paths, ensure remote access works manually, correct profile prompts and root authentication issues, and use local runs followed by report merging as a temporary alternative while remote connectivity is repaired.

Other Error Messages in orachk_error.log or exachk_error.log

Many messages in Orachk or Exachk error logs are expected in role-separated environments, where permission-denied OPatch operations occur against homes owned by other users and are handled through alternate inventory methods. These messages, along with repeated null integer-expression warnings, generally do not require reporting; running the tools as root can reduce such output.

Space available on {node_name} at {path} is {x} MB and required space is 500 MB

This message means the tool lacks sufficient space for temporary files, which default to the executing user’s home directory. Free at least the required space or set RAT_TMPDIR to a location with adequate capacity before rerunning Orachk.

Running Oracle Orachk on Microsoft Windows Throws '{oratab}' is empty Error

The Windows {oratab} is empty error occurs when no database is installed, registry-key patterns are nonstandard, or the database initialization file is missing. Set RAT_KEY_DB for nonstandard registry keys when necessary, restore the expected initSID.ora, or explicitly provide the database home and name with orachk -dbconfig db_home_path%db_name.

Operating System Is Not Discovered Correctly

When Orachk or Exachk cannot identify the operating system or derives an unsupported platform, set the RAT_OS environment variable to the correct platform identifier before running the tool.

Oracle Clusterware Software is Installed, but Cannot be Found

Clusterware discovery depends on valid oraInst.loc and Oracle inventory files and paths. Correct those files or set RAT_INV_LOC to the inventory location and RAT_CRS_HOME to the Oracle Clusterware home.

Oracle Database Software Is Installed, but Cannot Be Found

If the database software is installed but not discovered, set RAT_ORACLE_HOME to the applicable Oracle home. Orachk and Exachk will then perform checks for databases running from that home.

Oracle Database Software Is Installed, but Version cannot Be Found

If the database software version cannot be detected, set the RAT_DB environment variable to the installed version, such as 11.2.0.3.0, before running the checks.

Oracle ASM Software is Installed, but Cannot be Found

When ASM is installed but not detected, set RAT_ASM_HOME to the applicable ASM or Grid home so Orachk and Exachk can locate the software.

Oracle Database Discovery Issues on Oracle Real Application Clusters (Oracle RAC) Systems

On RAC systems, Orachk derives database homes from Cluster Registry resource profiles; invalid profiles can prevent discovery or connection. Temporarily specify databases with -dbnames or RAT_DBNAMES, and use RAT_DBHOMES to include additional database homes in recommended patch analysis.

Database login failures usually result from insufficient privileges or incorrect Oracle environment settings. Test the connection independently with the appropriate operating-system user, ORACLE_HOME, ORACLE_SID, PATH, network alias, and SQL_Plus SYSDBA credentials; if SQL_Plus cannot connect, Orachk and Exachk cannot connect either.

Remote checks require working SSH and SCP commands and suitable authentication, including passwordless root access or functional Expect handling where privileged operations are needed. Verify executable paths and manual remote commands, remove interactive profile prompts, enable debugging only under Oracle guidance, and use local collection with later report merging as a workaround.

Permission Problems

Ensure the Orachk and Exachk executables have mode 755 and that users can read and write the installation, temporary, working, and output directories. In clustered environments, install the tools in the recommended locations under the appropriate Grid Infrastructure owner or root so role separation does not block inventory and collection operations.

Slow Performance, Skipped Checks, and Timeouts

Slow or hung child processes are terminated by the watchdog and recorded as skipped checks. Investigate workload, remote profile prompts, and system performance; increase general, check-specific, or password-check timeouts through the relevant RAT_*_TIMEOUT variables, exclude problematic checks for separate runs, or tune database parallelism with -dbparallel, -dbparallelmax, or -dbserial.

Running Compliance Checks on a Subset of Oracle Home and Oracle Databases

Use -dbconfig to restrict Orachk or Exachk checks to selected Oracle homes and databases when discovery is incomplete or multiple homes exist. Specify each home and database set as dbhome%dbname, separate multiple databases with colons, and separate home groups with commas.

SSH Connection Timeout

If an Exachk SSH session times out while searching database homes, set ServerAliveInterval 30 in /etc/ssh/ssh_config on the machine where the run starts, then retry the check.

Oracle Exachk Prompts to Enter Names of RoCE Fabric Switches

Exachk repeatedly prompts for RoCE switch names when they are not defined for supported Exadata environments. Add comma- or line-separated leaf and spine switch names to switches.out, pass them with -switches, or exclude switch checks with the appropriate -excludeprofile option.

Unable to Implement CA Certificates in Oracle Trace File Analyzer

SSL socket failures after CA certificate configuration can occur when intermediate and server certificates are incorrectly combined in caroot.cert.txt, producing an empty certificate chain. Separate the certificates, recreate the server and client JKS files with the prescribed keytool imports, and verify that the resulting keystores contain matching private-key and trusted-certificate entries.

AHF Insights Report Reloads Infinitely in Chrome and Edge

On affected Windows systems, the CrowdStrike browser extension can cause AHF Insights index.html reports to reload continuously in Chrome and Edge. Open the report in another browser without the extension or in an incognito/private window where extensions are normally disabled; enterprise policy may prevent disabling the extension in standard browsing sessions.
Read More: AHF Insights Report Reloads Infinitely in Chrome and Edge