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.
Exim is not failing this message anew. FAQ Q0608 asks what "retry time not reached [for any host]" means, and why Exim will not try to deliver. The answer starts: "That is not an error. It means exactly what it says." A previous attempt failed with a temporary error, and Exim computed the earliest time to try again. The same rule can apply to local and to remote deliveries. For remote deliveries, each host has its own retry time.
Normally the message stays on the queue. There is one case where it does not. FAQ Q0620 covers the same words appearing after a long failure period: all hosts the message could be sent to have been failing so long that the end of the retry period (typically 4 or 5 days) has been reached. Exim "still computes a next time to retry, but any messages that arrive in the meantime are bounced straight away". The sender then gets a bounce that gives the error "retry time not reached for any host after a long failure period". Q0620 says you can alter that behaviour by unsetting the delay_after_cutoff option on the smtp transport, and then Exim will try most messages for those hosts once before giving up.
Forcing a try before the retry time is a separate decision, and a successful delivery cannot be pulled back.
What the symptom means
Chapter 32 says the retry section controls how often Exim tries to deliver messages that could not be delivered at the first attempt. If there are no retry rules, there are no retries, and temporary errors are treated as permanent. The default configuration contains one general-purpose rule.
Retry processing for a remote host that is down, or unreachable, is per host. Strictly, it is per IP address, not per message. If one message was delayed recently, a new message to the same host waits for that host's retry time. When the retry_defer log selector is set, the message "retry time not reached" is written whenever a delivery is skipped for this reason.
FAQ Q0608 says timing is normally host-based for remote delivery and address-based for local delivery. A new message that arrives early is left on the queue, and the log line is written, with no delivery attempt. When the time has passed, the next queue run tries.
The words "for any host" mean none of the hosts is due yet. FAQ Q0617's related line, "retry time not reached for other hosts", is the case where some hosts failed lookup and another host is not yet due.
Chapter 32 says retry times are hints, not promises. Without new mail, the gap between tries is the gap between queue runners. A five-minute rule does little if the runner is hourly. An immediate delivery still runs the routers and local deliveries. Remote hosts keep their retry times, which is why a new message can log this line instead of connecting.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| The destination was down or unreachable, and the timer has not expired | retry time not reached, and an earlier defer for the same host | exinext on the address, and the earlier log line |
| Several MX hosts each have their own time, and none is due | "for any host" | exinext lists each host |
| One host failed DNS and another's timer has not arrived | "retry time not reached for other hosts", as in FAQ Q0617 | dig the MX targets |
| The queue runner interval is longer than the retry rule | The next real try is the next queue run, not the rule's interval | Chapter 32: the minimum gap without new mail is the queue-runner interval |
| The retry period has ended after days of failure | The sender receives a bounce that gives the error "retry time not reached for any host after a long failure period" (FAQ Q0620) | The earlier defers for that host go back 4 or 5 days |
| The address is unrouteable | A bounce with a different text, not this log line | That is a routing failure, not a retry skip |
Prove the cause before changing anything
All of these are read-only.
1. Find the earlier temporary error. Q0608 says this line follows a temporary failure. Search the log for the same message id or the same host, earlier than this line. The defer reason is the fault to fix. This line is only the wait.
2. Ask when Exim will try. The FAQ says to run exinext on the address:
exinext user@some.domain
Chapter 32 says exinext reports the next retry times for the hosts associated with a mail domain, and for local deliveries that have been deferred. Use the address from the log.
3. See which rule matched. The command-line option -brt tests retry rules. The chapter's example is:
exim -brt bach.comp.mus.example
It writes the rule that matches. The first argument can be a full address or a domain.
4. Check the MX if the earlier error was DNS. FAQ Q0617 shows a domain whose lower-priority exchanger has no address, while the other host's retry time has not been reached. dig the domain's MX and each target before you conclude the timer itself is wrong.
Fix
Fix the temporary error that started the timer: the remote host, the network path, or the DNS the earlier line named. When that is corrected, the next queue run after the retry time will try again. You do not need to delete the retry rule to clear one host.
Save the exinext output before you change DNS or a retry rule. Run exinext on the address from the log. The answer goes to its own file. Copy the earlier defer line from the log into defer-line-before.txt:
exinext <address-from-log> > exinext-before.txt
If the earlier failure was a bad MX and you publish the zone, save that MX as well, to its own file:
dig <domain> MX > mx-before.txt
Risk: the next step changes the MX record of the domain in its DNS zone. It sends the domain's mail to the host you name, and a wrong target delivers mail to the wrong server.
Correct that record, then run exinext again for the same address.
To undo it: put the MX records that mx-before.txt shows back in the zone. If that file shows no MX record, delete only the MX record you added.
Leave the retry rules in place unless exim -brt shows a rule you did not mean to publish. If a rule is one you did not mean to publish, save the retry configuration first. Print the name of the runtime configuration file and copy it:
exim -bP config_file
cp -a <config-file> <config-file>.before
Risk: the next step edits the retry section of the Exim runtime configuration file. It changes the wait for every domain that matches the rule, not only this message. Shortening the intervals makes every failure retry more often. Removing the rules, chapter 32 says, turns temporary errors into permanent failures.
Open the file named by exim -bP config_file in your text editor, change only the rule you did not mean to publish, and save it. Run exim -brt <domain> again to see the rule that now matches.
To undo it: copy the saved file back with cp -a <config-file>.before <config-file>.
FAQ Q0608 also shows a force, exim -M with a message id. The command-line chapter says -M delivers, thaws a frozen message, and overrides retry hints. The caller must be an admin user unless prod_requires_admin is false. For the whole queue the FAQ names the -qf option. One f flag forces each non-frozen message.
To show the exact message first, list the queue. The command-line chapter shows each entry as an age, a size, the message ID and the sender on the first line, then the recipients. The message ID is the third item on the first line. Read it together with the recipient and the host named in the log:
exim -bp
Disruptive step: the next command makes Exim attempt delivery of that one message now, to the real recipient, and overrides the retry hint. It also thaws the message if it is frozen. A message that is delivered has left the queue and cannot be brought back, so there is no back-out. Run it only for the message you identified above.
Force that one message only:
exim -M <message-id>
Do not use exim -qf to test one domain. It forces every non-frozen message in the queue, to real recipients, and has no back-out either.
Hiding the line is not a fix either. Q0608 says -retry_defer in log_selector stops the log text. The mail still waits.
Example (illustrative): the log says retry time not reached for any host for sam@example.net. An hour earlier, delivery to mx.example.net deferred because the connection timed out. exinext sam@example.net shows a next time later this afternoon. The remote host is up again. The administrator leaves the message queued until that time, rather than flushing every deferred message on the server.
Confirm the fix
- The earlier defer reason no longer applies: the host accepts a connection, or the DNS lookup that failed now succeeds.
- After the time exinext printed, a queue run attempts delivery, and the log shows a delivery or a new defer reason, not only "retry time not reached".
- New mail to a host that has passed its retry time is tried. New mail to a host that has not is still logged as not reached. That remaining line is the rule working.
What to watch afterwards
- With
G(geometric) andH(random) rules, a later failure grows the wait. Chapter 32's algorithms areF(fixed),GandH, capped byretry_interval_max, default24h. - The same line for every domain points at this machine's DNS or network.
Read next
Related articles you may want next.
- TroubleshootingExim: Unrouteable address, and MX records that point nowhereExim says Unrouteable address, or that MX records point to non-existent hosts. Test the address, read the MX, and fix the router or the DNS that caused it.Read article
- 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
Risks and limits
- The FAQ applies to Exim 4.43 and later. Its section 6 page shows no date, and the FAQ page shows "Last update: 14-October-2004". The current retry chapter describes the same per-host wait.
- exinext reads hints. Chapter 32 says those times are not a promise that a queue runner is awake at that minute.
Sources
- Exim FAQ, section 6, Q0608 — Exim FAQ, no date shown on this page, accessed 22 November 2026
- The Exim FAQ — Exim FAQ, "Last update: 14-October-2004", accessed 22 November 2026
- Retry configuration — Exim specification, current HTML, no date shown, accessed 22 November 2026
- The Exim command line — Exim specification, current HTML, no date shown, accessed 22 November 2026
This article was originally published on .