gromox-mbsize

Name

gromox-mbsize — Mailbox size analysis

Synopsis

gromox-mbsize directory

Description

Shows a detailed view of how a mailbox size translates to on-disk usage. Explanation of the columns/rows follows.

Apparent size: This is the exact size of the object, or simply the sum of sizes of objects.

On FS: This is the space that is used on the filesystem, and is subject to fs block sizes. Details about this behavior may be found on <https://en.wikipedia.org/wiki/Block_(data_storage)>. As a result, the on-disk size may be larger than the apparent size.

RFC5322/Mbox: For the sake of IMAP, RFC5322 copies of messages and some metadata is retained.

RFC5322 Received: Applies to messages received via delivery(8gx).

RFC5322 Sent: Applies to any other message.

Body analysis: A set of 4 MAPI properties that usually get stored as files on disk rather than inside the sqlite database: PR_BODY, PR_HTML, PR_RTF_COMPRESSED and PR_INTERNET_TRANSPORT_HEADERS. In gromox-mbsize, these are considered "body".

Attachment analysis: What it says. Not all MIME parts are or stay an attachment; for example, calendar items/meeting requests are usually converted to MAPI objects.

Missing items/Apparent: The number of MAPI properties/attachments which have a dangling reference into the filesystem.

Missing items/FS: The on-disk number of files that seem to be absent. This number can be lower than Apparent due to internal data deduplication that is transparent to MAPI/exmdb clients.

Informational content: The logical amount of data that is represented by those four MAPI properties / by MAPI attachments.

After deduplication: The logical amount of unique bodies/attachments.

Dedup ratio/gains: For the "body" group, there are usually little gains to be observed in practice; bodies are just very unique. Messages like "test" are ironically the ones that benefit. Attachments dedup a little better, owing to many people sending/receiving redundant information, such as company logos.

After compression: Besides deduplication, Gromox can also compress before data goes to disk. This is also the final form and so there is an apparent and an on-disk value. The on-disk value may be higher due to aforementioned filesystem block sizing.

File compress ratio/gains: The earnings going from Dedup to Compressed.

IFC compress ratio/gains: The earnings going from Informational Content to Compressed.

MAPI reported sizes: The PR_MESSAGE_SIZE property of a store, folder, message, and the PR_ATTACH_SIZE property of an attachment, all give a close approximation to the amount of data needed to transfer the object(s) over a MAPI connection.

NTS deviation: how much the Network Transfer Size is off from the on-disk size.

Provisioning factor: The ratio between on-disk usage and the logical mailbox size reported inside MUAs.

See also

gromox(7)