Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Postfix Queue ID Lengths

899 views
Skip to first unread message

Joshua E Warchol

unread,
Aug 5, 2002, 10:26:21 AM8/5/02
to
How long should the message ID be in Postfix 1.1.11? I've got 9 servers
where it is 11 characters, and one server where it is 10. They all are
running the same install of postfix (from a modifed smudd RPM). Is this
normal behavior?

Thanks!

--
Joshua Warchol
UNIX Systems Administrator
DSL.net
-
To unsubscribe, send mail to majo...@postfix.org with content
(not subject): unsubscribe postfix-users

Wietse Venema

unread,
Aug 5, 2002, 10:55:40 AM8/5/02
to
Joshua E Warchol:

> How long should the message ID be in Postfix 1.1.11? I've got 9 servers
> where it is 11 characters, and one server where it is 10. They all are
> running the same install of postfix (from a modifed smudd RPM). Is this
> normal behavior?

The queue ID is in part based on the message file's inode number (*).

The length of a queue ID therefore depends on the range of inode
numbers in the file system.

Wietse

(*) This is how Postfix can create a file in the incoming directory
without danger that a file with the same name already exists
in the active or deferred directory. If multiple queue files
had the same name, then one of them would be lost as mail is
moved from incoming to active to deferred queue.

Joshua E Warchol

unread,
Aug 5, 2002, 11:03:30 AM8/5/02
to
Very cool information, I always wondered how it did that. There must just
be something different on my systems then. I'm not sure what, however. They
are all using ext3 spool disks (data=journal), and are all running the same
version of linux. Reminds me of the song "One of these things is not like
the other...".

Thanks for the info

--

Joshua Warchol
UNIX Systems Administrator
DSL.net

Wietse Venema

unread,
Aug 5, 2002, 1:44:51 PM8/5/02
to
The difference depends on the history of the file system. ext3fs,
like ext2fs (which has it from Berkeley FFS) divides a file system
into zones. Each zone contains a number of inodes and data blocks.

When a directory is created, the file system looks for a zone with
lots of unused space. This avoids file fragmentation.

When a file is created, its inode and data blocks are allocated
from the same file system zone as the parent directory. This reduces
disk head movement.

So it depends on the history of the file system where your Postfix
queue directories end up on the disk, and therefore, what typical
incoming queue file inode numbers will be like.

Wietse

Joshua E Warchol:

0 new messages