Mehmet Avcioglu:
> I am getting "too many reverse jump records" messages. Couldn't find
> any information about this message and looking at the source code
> (record.c) it seems to be related with adding recipients and/or
> headers, but I am not adding that many (10000) of either.
>
> There is a milter application before queue that adds a couple of
> headers but that is it. Between start and the final delivery there
> were 509 of these warnings (delays=3548/4125/0.24/0.49). This is a
> very active server and usually there are a few thousand mails in the
> queue.
When this is logged, the record reader returns an error and the
program stops reading the file.
> 2021-07-13T15:26:14.777532+00:00 server01 postfix/smtpd[912469]:
> 4GPKc12345z6n5pV: client=unknown[10.11.22.33], sasl_method=XCLIENT,
> sasl_username=u...@redacted.com
> 2021-07-13T15:26:14.827403+00:00 server01 postfix/cleanup[910108]:
> 4GPKc12345z6n5pV:
> message-id=<i3sb59ki3vilfcdmrq9iure84...@server.redacted.com>
> 2021-07-13T15:26:15.769603+00:00 server01 postfix/showq[785453]:
> warning: incoming/4GPKc12345z6n5pV: too many reverse jump records
The postfix/showq logging repeats the same error for the same file,
every 15 seconds, presumably because you are running 'posqeueue -p"
or "mailq" repeatedly.
Eventually, the Postfix SMTP client readds the file without error.
> 2021-07-13T17:34:06.914021+00:00 server01 postfix/smtp[1788437]:
> 4GPKc12345z6n5pV: to=<u...@rctpdomain.com>,
> relay=rcpt.server.com[10.12.23.34]:25, delay=7673,
> delays=3548/4125/0.24/0.49, dsn=2.0.0, status=sent (250 2.0.0 Message
> accepted for delivery)
Now, Postfix queue files don't repair themselves spontaneously.
Before I go off with speculation, I have a few questions to narrow
the search:
- Is only the showq process affected or other programs, too?
- Is this reproducible?
- Are you using a special malloc library?
I'm suspecting code that I wrote in 20060903 for record loop
detection, and the way that it resets the reverse jump counter when
it switches queue files. But that would not be the cause if
other programs also report "too many reverse jumps".
Wietse
