Subject: Re: increasing UT_HOSTSIZE for IPv6?
To: Simon Burge <firstname.lastname@example.org>
From: Greg Hudson <ghudson@MIT.EDU>
Date: 07/26/1999 09:55:40
[For reference, the proposal is to use db for the new version of utmp
> Our current db (v1) has locking problems, and the v2 licence isn't
> "friendly" at the moment...
To the best of my knowledge, db version 1 isn't really intended to
support concurrent access to a database. But that doesn't have to be
fatal for multiple writers; just lock the file before updating it.
How to synchronize writers with readers is a harder problem, since the
readers are not all root and you don't want non-root processes to be
able to interfere with login(1). In most cases the utmp record is
less than one block and there's no race, though.