The sys admin's daily grind: Archivemail

Box of Legends

By Charly Kühnast

Occasionally I need to delve into my mailboxes for an antique message, just to prod my memory. Because my mail client has a full text index of the inbox, I usually find the information I'm looking for within a couple of seconds. Unfortunately, this flush of success is often a delicate thing: With 300 messages flooding in every day, not counting mailing lists, the indexer often gives up the fight. And because many messages have bulky attachments, I genuinely need to clean out my modern day equivalent of the Augean stables.

Archivemail [1] can help me clean up my mailbox; in fact, it has a couple of neat options for doing so. The dry run switch makes sure my mail survives the initial experiments. After setting it, Archivemail simply simulates the process without actually writing. The simple command line

archivemail --dry-run /home/charly/mailbox

packs messages that have lived in the mailbox for more than 180 days into a separate mailbox_archive file, then unbloats the file with Gzip.

Daily

Archivemail creates a new mbox-formatted file so that I can easily read the file with a mail client if I need to do so. Of course, you can configure the expiry date: If I want to get rid of mail before the 180-day limit, I could specify half this value,

archivemail --dry-run --days=90 /home/charly/mailbox

as shown in Figure 1. But in fact, I prefer to keep my messages far longer: days=730 to be precise. To keep my inbox lean despite this, I need to call Archivemail again:

archivemail --dry-run --size=1000000 /home/charly/mailbox

This sends all the messages greater than 1MB in size to the archive.

Figure 1: In this example, Archivemail takes messages of 90 days or older from an overflowing mailbox and archives them in a Gzip file.

Archivemail has a couple of other useful options. If I use the archive mbox regularly or need permanent access, --no-compress prevents Gzip compression. If you get behind and can't keep up with all your messages, -preserve-unread excludes unread mail from the archiving process, even if time- or size-based criteria normally would apply.

The one thing you really need to organize your life is a big trash can. This adage also applies without restriction to mailboxes. If you are of the same mind, you might also want to try the -delete switch. This moves any messages that match the search criteria directly to the trash can rather than going through the archiving step.