About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
SQL Server raises error 4214 when a job or person asks for a transaction log backup of a database that has no full backup behind it. Take a full database backup of that database, then run the log backup again. The log job will keep failing every cycle until that first full backup exists.
Microsoft's own wording for the user action is short: perform a full database backup before you try to back up the transaction log. The harder part is working out why the full backup is missing, because a log job that fails every fifteen minutes may have a schedule, a recovery-model change or a failing full job hiding behind it.
What the error means
Microsoft documents error 4214 with the message text "BACKUP LOG cannot be performed because there is no current database backup". The symbolic name is DMPX_NODBBACKUP, the event source is MSSQLSERVER and the page lists SQL Server 2017 and later. The explanation on that page reads: "The error is raised when you attempt a transaction log backup before doing a full backup of a database." Microsoft also says you receive a second message in the error log: "Backup Error: 3041, Severity: 16, State: 1" with "BACKUP failed to complete the command BACKUP LOG" and the database name.
A log backup is a slice of the transaction log that is only useful when it can be applied after a data backup. Microsoft's log-backup page puts it plainly: "Minimally, you must have created at least one full backup before you can create any log backups."
The recovery model matters as well. The simple recovery model does not support transaction log backups. SQL Server Enterprise and Standard editions use the full recovery model by default for new databases, while Express uses the simple recovery model by default. A database that moved from simple to full is in a particular state: Microsoft's page on changing the recovery model says that immediately after switching to the full or bulk-logged model you should take a full or differential database backup to start the log chain, and that "The switch to the full or bulk-logged recovery model takes effect only after the first data backup."
Three terms help when you read the checks below:
- Log chain: the unbroken sequence of log backups after a data backup.
- Copy-only backup: a backup taken outside the normal sequence. Microsoft says a copy-only full backup "can't serve as a differential base or differential backup and doesn't affect the differential base". The page does not say whether a copy-only full backup satisfies a log backup, so a normal full backup is the safe base.
- msdb backup history: the
msdb.dbo.backupsettable records each backup set, including a type letter (D,IorL), anis_copy_onlyflag and abegins_log_chainflag. Microsoft definesbegins_log_chainfor log backups only: "1 = This is the first in a continuous chain of log backups. A log chain begins with the first log backup taken after the database is created or when it is switched from the simple to the full or bulk-logged recovery model." It describes log backups, so read it onLrows and never onDrows.
A database in the simple recovery model that receives a log backup statement returns a different error. Microsoft lists error 4208 as "The statement %hs is not allowed while the recovery model is SIMPLE. Use BACKUP DATABASE or change the recovery model using ALTER DATABASE." Error 4214 is the one for a database whose log backup has no full backup behind it.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| New database that has never had a full backup | Log job fails from the first run; database was created recently | msdb.dbo.backupset has no row with type D for the database |
| Recovery model switched from simple to full or bulk-logged, no data backup since | Log job starts failing straight after the change | recovery_model_desc is FULL, and no full or differential backup finished after the change time |
| Full backup job failing, so the log job has nothing to follow | Full job fails or is disabled; log job fails with 4214 | SQL Server Agent job history for the full job; no recent type D row |
| Log job targets the wrong database or a database restored under a new name | Error names a database nobody backs up | Compare the name in the error with the names in the full-backup job |
Prove the cause before changing anything
All of these are read-only. Run them in SSMS or sqlcmd against the instance that logged the error. Replace SalesDB with the database named in your error.
First, find the recovery model and state of the database. Microsoft documents recovery_model_desc in sys.databases.
SELECT name, recovery_model_desc, state_desc
FROM sys.databases
WHERE name = N'SalesDB';
If the result is SIMPLE, SQL Server will not take log backups of that database at all, and the real question is whether you want the full recovery model.
Second, list the last backups SQL Server has recorded. The columns used here are documented on the backupset page.
SELECT TOP (10)
bs.database_name, bs.type, bs.is_copy_only,
bs.backup_finish_date, bs.begins_log_chain
FROM msdb.dbo.backupset AS bs
WHERE bs.database_name = N'SalesDB'
ORDER BY bs.backup_finish_date DESC;
Microsoft lists the type values as D for a database backup, I for a differential, and L for a log. Read the result this way:
- No rows at all: the database has never been backed up on this instance, or the history was cleared. This points to the first cause.
- Rows of type
Lonly, with noDrow: the history holds log backups but no full backup. Look for the full backup in another history or on the instance that took it. - A
Drow exists butis_copy_only= 1 on every one of them: only copy-only full backups are recorded. Microsoft does not say whether these satisfy a log backup, so take a normal full backup. - A
Drow withis_copy_only= 0 exists, but itsbackup_finish_dateis earlier than the time the recovery model was switched to FULL: Microsoft's statement that the switch "takes effect only after the first data backup" applies. Take a new full backup now.
Do not read begins_log_chain on D rows. It describes L rows, and marks only the first log backup after the database was created or switched from simple.
Third, read the SQL Server Agent job history for the full backup job and for the log job. If a maintenance plan or third-party tool runs the backups, find the actual statement it sends. The 4214 line in the error log should be followed by the 3041 line Microsoft describes, naming the database.
Finally, check whether the failing log backup is meant to exist. A database in SIMPLE recovery that a log job still names is a configuration mistake, not a missing backup.
Fix
New database, or failing full job. Take a full backup. Microsoft's Transact-SQL form is BACKUP DATABASE to a device or file.
BACKUP DATABASE [SalesDB]
TO DISK = N'D:\Backup\SalesDB_full.bak';
Risk: a full backup reads the database and writes a backup file, which competes with users for disk and network bandwidth. It changes nothing in the database. To back out, delete the backup file once it is no longer needed. Keep it while it is the only base for later log backups.
Then run the log backup once by hand to prove the chain starts.
BACKUP LOG [SalesDB]
TO DISK = N'D:\Backup\SalesDB_log1.trn';
Risk: a log backup truncates the transaction log. Microsoft recommends frequent log backups "to truncate the transaction log". The .trn file becomes a link in the log chain, so keep it with the full backup for any later point-in-time restore. There is no way to un-truncate the log; deleting the file breaks the chain from that point.
Recovery model is SIMPLE but the application needs point-in-time restore. A SIMPLE database returns error 4208 for a log backup, not 4214, so treat this as a configuration decision. Microsoft's page on changing the recovery model gives the statement ALTER DATABASE [model] SET RECOVERY FULL; as its example. For your database:
ALTER DATABASE [SalesDB] SET RECOVERY FULL;
Disruptive step: this changes how the log is managed. Microsoft says that under the full recovery model the transaction log continues to grow until you perform a transaction log backup, so a missing log job can fill the log drive. Microsoft also states that plan cache entries for the database are flushed. To back out, set the model back to SIMPLE. That discards the log chain: Microsoft says that to remove the log backup chain you switch to the simple recovery model.
Take a full or differential backup straight after the switch, as Microsoft advises, then schedule regular log backups.
Only copy-only full backups are recorded. Take a normal full backup (without the copy-only option) using the BACKUP DATABASE statement above. Risk: the backup reads the database and writes a backup file, which competes with users for disk and network bandwidth; it changes nothing in the database. To back out, delete the backup file once it is no longer needed, but keep it while it is the only base for later log backups. Microsoft says a copy-only backup is "independent of the sequence of conventional SQL Server backups".
Log job names the wrong database. Correct the job step. Do not take a full backup of a database nobody runs.
Confirm the fix
Re-run the second query. You should see a D row with is_copy_only = 0 and a backup_finish_date later than any recovery-model change, then an L row after it. Confirm the next scheduled log backup succeeds with no 4214 in the error log. If the database was just created or just switched from simple, begins_log_chain on that first L row is 1 under Microsoft's definition; read the flag only on L rows.
Example (illustrative): on sql01, a developer creates SalesDB at 01:30. The log backup job runs every 15 minutes from 02:00. The nightly full backup job is set for 22:00, so the first log attempt at 02:00 fails with error 4214, and so does each later one. An administrator sees the failing job at 09:00 and starts BACKUP DATABASE by hand at 09:02; it finishes at 09:10. The failed attempts run from 02:00 to 09:00 inclusive. That is 7 hours, or 420 minutes, divided by 15 equals 28 intervals, plus the first run, for 29 failed log backup attempts. The log job at 09:15 succeeds. Nothing is lost, because no log backup ever existed, but the database has no restore point for the 7 hours 40 minutes between 01:30 and 09:10. Had nobody acted, that gap would have lasted 20.5 hours, until the 22:00 full backup.
What to watch afterwards
- Alert on the first failed log job, not the tenth. A job that fails every 15 minutes generates 96 failures a day (24 hours times 4 runs an hour).
- Make the full backup the first step after any recovery-model change, restore or database creation script.
- After a full backup that follows a switch to FULL, check that an
Lrow appears after theDrow in the backup history. - Keep the SIMPLE-to-FULL switch and the log job schedule in the same change record so one cannot go live without the other.
- Watch log file size in the first week. Microsoft says that under the full recovery model the log is truncated after a log backup, so a log file that keeps growing means log backups are not completing.
Read next
Related articles you may want next.
- Backup & RecoverySQL Point-in-Time Restore: Plan and Verify the Backup ChainCheck native SQL backup chains, select a recovery point and rehearse a separate restore without confusing readable backups with recovered service.Read article
- Backup & RecoveryAd Hoc SQL Backups: Protect the Differential Base with Copy-OnlyChoose the correct native backup type and record the actual differential base and log ownership.Read article
- TroubleshootingSQL Transaction Log Keeps Growing: Find the Real CauseSeparate log reuse, physical file growth and disk exhaustion, then identify the evidence needed for a DBA-approved SQL Server remedy and recovery check.Read article
Risks and limits
These commands have not been run or lab-tested. Microsoft's error page lists SQL Server 2017 and later, and the wording of the Agent job messages can differ by tool. The page does not list every route to the error; the causes in the table are the ones the documented behaviour supports, and the copy-only reading above is an undocumented possibility, not a Microsoft statement. Third-party backup products may issue log backups on their own schedule; read their job text before changing a recovery model they depend on.
Do not run a full backup of a large database during business hours just to clear an alert. Do not switch to SIMPLE to silence log jobs without agreeing the loss of point-in-time restore, because that switch discards the log chain.
NetLeverage includes a daily full disaster-recovery backup of each hosted server at no extra charge. Talk to an engineer.
Sources
- MSSQLSERVER_4214 — Microsoft Learn, page dated 2026-03-05, accessed 18 November 2026
- Recovery models (SQL Server) — Microsoft Learn, page dated 2026-02-20, accessed 18 November 2026
- View or change the recovery model of a database (SQL Server) — Microsoft Learn, page dated 2025-11-25, accessed 18 November 2026
- Transaction log backups (SQL Server) — Microsoft Learn, page dated 2026-01-21, accessed 18 November 2026
- Back up a transaction log (SQL Server) — Microsoft Learn, page dated 2025-10-28, accessed 18 November 2026
- BACKUP (Transact-SQL) — Microsoft Learn, page dated 2026-09-21, accessed 18 November 2026
- backupset (Transact-SQL) — Microsoft Learn, page dated 2026-09-21, accessed 18 November 2026
- Copy-only backups (SQL Server) — Microsoft Learn, page dated 2026-07-20, accessed 18 November 2026
- sys.databases (Transact-SQL) — Microsoft Learn, page dated 2026-09-21, accessed 18 November 2026
- Database Engine events and errors (4000 to 4999) — Microsoft Learn, page dated 2026-10-27, accessed 18 November 2026
This article was originally published on .