entry 'string' does not exist in directory 'string'
Fix: List the parent directory (asmcmd ls -l +DATA//) to see the real names, and use find to locate the file.
Fix: List the parent directory (asmcmd ls -l +DATA//) to see the real names, and use find to locate the file.
Fix: Install on a certified OS release, or set CV_ASSUME_DISTID to a release the installer knows before starting it, and apply the latest Release Update, which adds support for newer OS...
Fix: Find the constraint and its columns in DBA_CONS_COLUMNS, then either skip the duplicate row, update the existing row instead (MERGE), or take the key from a sequence / identity col...
Fix: Check the spelling and case against DESCRIBE table. To filter or group on an alias, repeat the expression or wrap the query in an inline view and filter the outer query. Aggregate...
Fix: Check which instances run (ps -ef | grep pmon) and that ORACLE_SID matches one of them, including case. If the instance is down, start it with STARTUP as SYSDBA and read the alert...
Fix: Supply a value for that column in the INSERT, or give the column a DEFAULT (ALTER TABLE ... MODIFY col DEFAULT ...). For generated keys, use a sequence or an identity column.
Fix: Handle it where a missing row is a normal case: add an EXCEPTION WHEN NO_DATA_FOUND block, or select COUNT(*) / use a cursor loop instead of SELECT INTO.
Fix: Use IN / ANY / EXISTS instead of =, or make the subquery return one row (an aggregate such as MAX, a tighter WHERE, or FETCH FIRST 1 ROW ONLY) — whichever matches what the query is...
Fix: Insert the parent row first (or in the same transaction, before the child), correct the foreign-key value, or check which parent table the constraint references in DBA_CONSTRAINTS...
Fix: Read the rest of the message and the line number in the error stack, then enlarge the variable (anchor it with %TYPE to the column) or fix the value being converted.
Fix: Create the directory named in AUDIT_FILE_DEST (from the pfile/spfile, or the default $ORACLE_BASE/admin//adump) with the right owner, and check free space.
Fix: Run lsnrctl status and compare the services listed with SERVICE_NAME in tnsnames.ora. Use the full name including the domain, or register the instance: check LOCAL_LISTENER and run...
Fix: Start the listener (lsnrctl start, or srvctl start listener on Grid Infrastructure), check the HOST and PORT in tnsnames.ora against lsnrctl status, and open the port in the firewa...
Fix: Clear the remote LOG_ARCHIVE_DEST_n on both databases (keep the local destination), then add the database to the broker again. Set transport options through broker properties such...
Fix: Register the database with srvctl add database (with its DB_UNIQUE_NAME, Oracle home, spfile and role), and add instances on RAC, then manage it with srvctl.
Fix: Check the registered resources with crsctl stat res -t or srvctl config database, and register the missing database with srvctl add database.
Fix: Remove the domain line (keep search) on all nodes, making sure NetworkManager does not write it back.
Fix: Add memory to the node. In a lab, where less memory is a deliberate choice, the check can be ignored in the installer.
Fix: Add a swap file or partition on each node, or ignore the check in a lab environment.
Fix: If you do not use libvirt on the nodes, remove the default network so virbr0 disappears; otherwise the warning can be ignored as virbr0 is not a cluster interface.
Fix: Add the node names to DNS, or, in a lab that uses /etc/hosts and has no DNS, ignore the check.
Fix: Make the node names resolvable in DNS with the right address type, or ignore it in a lab that resolves names from /etc/hosts only.
Fix: With the Clusterware stack down on that node, remove the leftover files, then run the check again.
Fix: Nothing needs fixing — it is a warning and the files are created in the target disk group with OMF names. To avoid it, convert only the disk group (e.g. '+DATA/PRIM', '+DATA/STBY')...
Nothing here for that yet. Try searching all articles.