archivemail (1) - Linux Manuals

archivemail: archive and compress your old email


archivemail - archive and compress your old email


archivemail [ options ] MAILBOX ...


archivemail is a tool for archiving and compressing old email in mailboxes. By default it will read the mailbox MAILBOX, moving messages that are older that the specified number of days (180 by default) to a mbox(5)-format mailbox in the same directory that is compressed with gzip(1). It can also just delete old email rather than archive it.

archivemail supports reading IMAP, Maildir, MH and mbox-format mailboxes, but always writes mbox-format archives.

Messages that are flagged important are not archived or deleted unless explicitely requested with the --include-flagged option. Also, archivemail can be configured not to archive unread mail, or to only archive messages larger than a specified size.

To archive an IMAP-format mailbox, use the format imap://username:password@server/mailbox to specify the mailbox. You can omit the password from the URL; use the --pwfile option to make archivemail read the password from a file, or alternatively just enter it upon request. If the --pwfile option is set, archivemail does not look for a password in the URL, and the colon is not considered a delimiter. Substitute 'imap' with 'imaps', and archivemail will establish a secure SSL connection. See below for more IMAP peculiarities.

archivemail has some support for being run as the root user on user mailboxes. When running as root, it will seteuid(2) to the owner of the mailbox it is reading, creating any archive files as that user. Warning: this automatic seteuid feature is insecure and deprecated. It will be removed from later versions of archivemail.


-d NUM, --days=NUM
Archive messages older than NUM days. The default is 180. This option is incompatible with the --date option below.
-D DATE, --date=DATE
Archive messages older than DATE. DATE can be a date string in ISO format (eg '2002-04-23'), Internet format (eg '23 Apr 2002') or Internet format with full month names (eg '23 April 2002'). Two-digit years are not supported. This option is incompatible with the --days option above.
-o PATH, --output-dir=PATH
Use the directory name PATH to store the mailbox archives. The default is the same directory as the mailbox to be read.
-P FILE, --pwfile=FILE
Read IMAP password from file FILE instead of from the command line. Note that this will probably not work if you are archiving folders from more than one IMAP account.
-F STRING, --filter-append=STRING
Append STRING to the IMAP filter string. For IMAP wizards.
-s NAME, --suffix=NAME
Use the suffix NAME to create the filename used for archives. The default is _archive. For example, if you run archivemail on a mailbox called exsouthrock, the archive will be created with the filename exsouthrock_archive.gz.

NAME is run through the python(1) time.strftime() function, which means that you can specify any of the following special directives in NAME to make archives named after the archive cut-off date:

%a Locale's abbreviated weekday name.
%A Locale's full weekday name.
%b Locale's abbreviated month name.
%B Locale's full month name.
%c Locale's appropriate date and time representation.
%d Day of the month as a decimal number [01,31].
%H Hour (24-hour clock) as a decimal number [00,23].
%I Hour (12-hour clock) as a decimal number [01,12].
%j Day of the year as a decimal number [001,366].
%m Month as a decimal number [01,12].
%M Minute as a decimal number [00,59].
%p Locale's equivalent of either AM or PM.
%S Second as a decimal number [00,61]. (1)
%U Week number of the year (Sunday as the first day of the week) as a decimal number [00,53]. All days in a new year preceding the first Sunday are considered to be in week 0.
%w Weekday as a decimal number [0(Sunday),6].
%W Week number of the year (Monday as the first day of the week) as a decimal number [00,53]. All days in a new year preceding the first Sunday are considered to be in week 0.
%x Locale's appropriate date representation.
%X Locale's appropriate time representation.
%y Year without century as a decimal number [00,99].
%Y Year with century as a decimal number.
%Z Time zone name (or by no characters if no time zone exists).
%% A literal "%" character.
-S NUM, --size=NUM
Only archive messages that are NUM bytes or greater.
-n, --dry-run
Don't write to any files -- just show what would have been done. This is useful for testing to see how many messages would have been archived.
-u, --preserve-unread
Do not archive any messages that have not yet been read. archivemail determines if a message in a mbox-format or MH-format mailbox has been read by looking at the Status header (if it exists). If the status header is equal to 'RO' or 'OR' then archivemail assumes the message has been read. archivemail determines if a maildir message has been read by looking at the filename. If the filename contains an 'S' after :2, then it assumes the message has been read.
Do not mangle lines in message bodies beginning with "From ". When archiving a message from a mailbox not in mbox format, by default archivemail mangles such lines by prepending a '>' to them, since mail user agents might otherwise interpret these lines as message separators. Messages from mbox folders are never mangled. See mbox(5) for more information.
Delete rather than archive old mail. Use this option with caution!
Copy rather than archive old mail. Creates an archive, but the archived messages are not deleted from the originating mailbox, which is left unchanged. This is a complement to the --delete option, and mainly useful for testing purposes.
Normally messages that are flagged important are not archived or deleted. If you specify this option, these messages can be archived or deleted just like any other message.
Do not compress any archives.
Warn about duplicate Message-IDs that appear in the input mailbox.
-v, --verbose
Reports lots of extra debugging information about what is going on.
-q, --quiet
Turns on quiet mode. Do not print any statistics about how many messages were archived. This should be used if you are running archivemail from cron.
-V, --version
Display the version of archivemail and exit.
-h, --help
Display brief summary information about how to run archivemail.


archivemail requires python(1) version 2.3 or later. When reading an mbox-format mailbox, archivemail will create a lockfile with the extension .lock so that procmail will not deliver to the mailbox while it is being processed. It will also create an advisory lock on the mailbox using flock(2). archivemail will also complain and abort if a 3rd-party modifies the mailbox while it is being read.

archivemail will always attempt to preserve the mode, last-access and last-modify times of the input mailbox. However, archive mailboxes are always created with a mode of 0600. If archivemail finds a pre-existing archive mailbox it will append rather than overwrite that archive. archivemail will refuse to operate on mailboxes that are symbolic links or create tempfiles or archives in world-writable directories.

archivemail attempts to find the delivery date of a message by looking for valid dates in the following headers, in order of precedence: Delivery-date, Date and Resent-Date. If it cannot find any valid date in these headers, it will use the last-modified file timestamp on MH and Maildir format mailboxes, or the date on the From line on mbox-format mailboxes.

A conversion from other formats to mbox(5) will silently overwrite existing Status and X-Status message headers.


When archivemail processes an IMAP folder, all messages in that folder will have their \Recent flag unset, and they will probably not show up as 'new' in your user agent later on. There is no way around this, it's just how IMAP works. This does not apply, however, if you run archivemail with the options --dry-run or --copy.

archivemail relies on server-side searches to determine the messages that should be archived. When matching message dates, IMAP servers refer to server internal message dates, and these may differ from both delivery time of a message and its Date header. Also, there exist broken servers which do not implement server side searches.


archivemail's IMAP URL parser was written with the RFC 2882 (Internet Message Format) rules for the local-part of email addresses in mind. So, rather than enforcing an URL-style encoding of non-ascii and reserved characters, it allows to double-quote the username and password. If your username or password contains the delimiter characters '@' or ':', just quote it like this: imap://"username [at]":"password" You can use a backslash to escape double-quotes that are part of a quoted username or password. Note that quoting only a substring will not work, and be aware that your shell will probably remove unprotected quotes or backslashes.

IMAP servers supporting subfolders may use any character as a mailbox path separator, that is, as an equivalent to the slash character on Unix systems. If you are archiving an IMAP subfolder, first archivemail will try to open a given mailbox name unchanged; if this fails, it will interpret any slashes in the URL as path separators and try again.


To archive all messages in the mailbox debian-user that are older than 180 days to a compressed mailbox called debian-user_archive.gz in the current directory:

bash$ archivemail debian-user

To archive all messages in the mailbox debian-user that are older than 180 days to a compressed mailbox called debian-user_October_2001.gz (where the current month and year is April, 2002) in the current directory:

bash$ archivemail --suffix '_%B_%Y' debian-user

To archive all messages in the mailbox cm-melb that are older than the first of January 2002 to a compressed mailbox called cm-melb_archive.gz in the current directory:

bash$ archivemail --date'1 Jan 2002' cm-melb

Exactly the same as the above example, using an ISO date format instead:

bash$ archivemail --date=2002-01-01 cm-melb

To delete all messages in the mailbox spam that are older than 30 days:

bash$ archivemail --delete --days=30 spam

To archive all read messages in the mailbox incoming that are older than 180 days to a compressed mailbox called incoming_archive.gz in the current directory:

bash$ archivemail --preserve-unread incoming

To archive all messages in the mailbox received that are older than 180 days to an uncompressed mailbox called received_archive in the current directory:

bash$ archivemail --no-compress received

To archive all mailboxes in the directory $HOME/Mail that are older than 90 days to compressed mailboxes in the $HOME/Mail/Archive directory:

bash$ archivemail -d90 -o $HOME/Mail/Archive $HOME/Mail/*

To archive all mails older than 180 days from the given IMAP INBOX to a compressed mailbox INBOX_archive.gz in the $HOME/Mail/Archive directory, quoting the password and reading it from the environment variable PASSWORD:

bash$ archivemail -o $HOME/Mail/Archive imaps://user:'"'$PASSWORD'"' 

Note the protected quotes.


Probably the best way to run archivemail is from your crontab(5) file, using the --quiet option. Don't forget to try the --dry-run and perhaps the --copy option for non-destructive testing.


Normally the exit status is 0. Nonzero indicates an unexpected error.


If an IMAP mailbox path contains slashes, the archive filename will be derived from the basename of the mailbox. If the server's folder separator differs from the Unix slash and is used in the IMAP URL, however, the whole path will be considered the basename of the mailbox. E.g. the two URLs imap:// and imap:// will be archived in subfolder_archive.gz and folder.subfolder_archive.gz, respectively, although they might refer to the same IMAP mailbox.

archivemail does not support reading MMDF or Babyl-format mailboxes. In fact, it will probably think it is reading an mbox-format mailbox and cause all sorts of problems.

archivemail is still too slow, but if you are running from crontab(5) you won't care. Archiving maildir-format mailboxes should be a lot quicker than mbox-format mailboxes since it is less painful for the original mailbox to be reconstructed after selective message removal.


The archivemail home page is currently hosted at sourceforge <URL:>


This manual page was written by Paul Rodger <paul at paulrodger dot com>. Updated and supplemented by Nikolaus Schulz <microschulz [at]>


python(1), gzip(1), mutt(1), procmail(1)