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.
Postfix is refusing the message because the queue filesystem does not have enough free space. In smtpd_check_queue the reply is the code 452, the status 4.3.1, and the text Insufficient system storage. The log warning beside it, in the source, is not enough free space in mail queue: N bytes < 1.5*message size limit, where N is the free-byte count.
The client can try again later. Mail already queued is still there. New mail is turned away until the free-space check passes.
What the symptom means
postconf(5) defines queue_minfree as the minimal amount of free space, in bytes, in the queue file system that is needed to receive mail. The default is 0. It also says that, by default, the SMTP server rejects MAIL FROM when free space is less than 1.5 times message_size_limit (Postfix 2.1 and later). To require still more free space, set queue_minfree to at least 1.5 times message_size_limit.
The source matches that factor. smtpd_space_multf is 1.5, and the comment says free space must be at least smtpd_space_multf * message_size_limit. The check rejects when queue_minfree is already greater than or equal to the free space, or when message_size_limit does not fit in the free space divided by 1.5.
message_size_limit defaults to 10240000 bytes. At that default, the filesystem needs about 1.5 times 10240000 bytes free, even when queue_minfree is 0. A much larger message_size_limit raises the free-space bar by the same factor. postconf warns that an excessively small message_size_limit drops non-delivery notifications when a bounce is larger than the limit. Raising it to get past a full disk does the opposite of what you want: the required free space grows.
The source checks the process's current directory with fsspace("."). postconf describes that space as the queue file system. postqueue(1) lists /var/spool/postfix as the mail queue. queue_directory is the parameter, and its default is whatever postconf -d prints. Check the filesystem that holds that directory, not a different disk that still looks empty.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| The queue filesystem is full | 452 4.3.1, and df shows little or no space available | df on the queue directory |
message_size_limit was raised, so 1.5 times that value is larger than the free space |
The warning compares free bytes with the size limit, and the disk is not actually full | postconf message_size_limit and df |
queue_minfree is set above the free space |
postconf -n shows a large queue_minfree |
Compare it with df |
| The queue holds a backlog of deferred mail | postqueue -p is long, and the filesystem is the queue disk |
List the queue, then df |
| Another log or database on the same filesystem grew | df is full, and the queue listing is short | df, then ls -l on the logs and data files you know are on that filesystem |
Prove the cause before changing anything
All of these are read-only.
1. Read the two limits.
postconf -n
postconf -d message_size_limit
postconf -d queue_minfree
postconf queue_directory
-n prints only parameters set in main.cf. -d prints defaults. message_size_limit defaults to 10240000. queue_minfree defaults to 0, which does not disable the 1.5 times size-limit check.
2. See free space on the queue filesystem. df(1) shows the space available on the filesystem containing the file you name. -h prints sizes in powers of 1024.
df -h /var/spool/postfix
If postconf queue_directory prints a different path, give that path to df instead. Read the available-space column.
3. See whether the queue itself is the bulk. postqueue -p lists queued mail and, when the last attempt failed, the reason. A long listing of deferred mail on a full queue disk is a backlog. A short listing on a full disk means something else on that filesystem is using the space.
Fix
There are two separate changes. Make only the one your checks point to. Never delete files under the queue directory: that deletes mail.
Save the limits and the df output first. Each answer goes to its own file, kept on a different filesystem from the queue:
postconf -n > <backup-directory-on-another-filesystem>/postconf-before.txt
df -h /var/spool/postfix > <backup-directory-on-another-filesystem>/df-before.txt
Use the path from postconf queue_directory if it is not /var/spool/postfix.
Change 1: free space by removing old logs
First select the file. These two commands are read-only. Name one log file you no longer need. Run df on it and on the queue directory. The Filesystem and Mounted on columns must be the same for both, because df(1) shows the space on the file system that contains the file you name. A log on a different filesystem frees nothing here:
df /var/spool/postfix <log-file>
ls -l <log-file>
Note the size that ls -l prints for the log file.
Save the file next. Copy the log to a different filesystem that has room, and check that the copy is complete: its size in the second listing must equal the size you noted:
cp -a <log-file> <backup-directory-on-another-filesystem>/
ls -l <backup-directory-on-another-filesystem>/
Risk: the next step deletes that one log file from the queue filesystem. A deletion cannot be undone except from your copy. Do not delete a log file that a program is still writing to, and do not delete anything under the queue directory.
Remove only the one file you selected, copied and size-checked:
rm <log-file>
Check with df -h /var/spool/postfix that the available bytes are above 1.5 times the message_size_limit you intend to keep, and above queue_minfree if you set it. If they are not, repeat this change with another file.
To undo it: copy the file back from <backup-directory-on-another-filesystem> to its original path with cp -a, if the filesystem has room for it again.
Change 2: set message_size_limit back to the default
Use this only when the disk has room at the default and the size limit was raised far past it. The values you saved with postconf -n are the way back.
Risk: the next command edits main.cf and reloads Postfix. Lowering message_size_limit makes Postfix reject larger messages with 552 and Message size exceeds fixed limit, which is a different check in smtpd_check_size. Mail that users could send before may now be refused.
postconf(1) writes name=value with -e:
postconf -e 'message_size_limit=10240000'
postfix reload
10240000 is the documented default. Use another number only if you have calculated 1.5 times that number against the free space df showed. postfix reload re-reads configuration.
To undo it: run postconf -e with the message_size_limit value from the saved postconf -n output, then postfix reload. If that parameter was not in the saved -n output, remove it with postconf -X message_size_limit and reload, which returns the built-in default.
Raising queue_minfree makes 452 more likely, not less, so do not raise it to fix this error.
Example (illustrative): message_size_limit is 102400000, so the check wants about 153600000 bytes free. df shows 80M available on /var/spool/postfix. The log says not enough free space in mail queue: 80000000 bytes < 1.5*message size limit. The administrator moves old compressed logs off that filesystem. df then shows more than 153600000 bytes available. The size limit is left unchanged because the disk can hold it. The next MAIL FROM is accepted.
Confirm the fix
- df on the queue directory shows available space greater than 1.5 times
message_size_limit, and greater thanqueue_minfreeif set. - A new SMTP session gets past MAIL FROM without
452 4.3.1. postqueue -pstill lists any mail that was already queued. This fix does not discard it.
What to watch afterwards
- Alert on free space on the queue filesystem before it falls under 1.5 times
message_size_limit. - If you raise
message_size_limit, raise the disk budget with it. postconf's note onqueue_minfreeis the formula. - Any data on the same filesystem as the queue can use up the free space that this check looks at. The SMTP reply does not name what filled the disk.
- Do not run
postqueue -fto clear the backlog. It flushes the whole deferred queue, and the manual warns that flushing undeliverable mail often hurts delivery of everything else. Free the disk first.
Read next
Related articles you may want next.
- TroubleshootingPostfix mail stuck in the queue as deferred: read the reasonPostfix mail stuck as deferred? Read the reason in postqueue -p and the mail log, group by destination with qshape, fix the cause, then flush once.Read article
- TroubleshootingPostfix: Host or domain name not found. Name service errorPostfix logged Host or domain name not found. Name service error. Read the name and record type in the log, then compare dig on the mail server with the DNS it should be using.Read article
- TroubleshootingPostfix 550 5.1.1 User unknown: local recipient table or virtual mailbox tablePostfix rejected a recipient with 550 5.1.1 User unknown. See whether the table is local or virtual, then add the address or relax the reject code only while you check the map.Read article
Risks and limits
- The 1.5 factor is
smtpd_space_multfin the Postfix 3.9.1 source and the figure in postconf(5), which says it applies to Postfix 2.1 and later. - df reports the filesystem of the path you give it. A queue on a different mount needs that path.
message_size_limitis the message size limit, not a mailbox size limit.
Sources
- smtpd_check.c — Postfix tag v3.9.1, no date shown, accessed 22 November 2026
- Postfix Configuration Parameters — Postfix, no date shown, accessed 22 November 2026
- postconf(1) — Postfix, no date shown, accessed 22 November 2026
- postfix(1) — Postfix, no date shown, accessed 22 November 2026
- postqueue(1) — Postfix, no date shown, accessed 22 November 2026
- df(1) — GNU coreutils 9.11, April 2026, accessed 22 November 2026
This article was originally published on .