On Thu, 20 Feb 2003, Giulio Orsero wrote:
> On Thu, 20 Feb 2003 09:56:43 -0500, "Billy Davis" <[EMAIL PROTECTED]>
> wrote:
>
> > (S4), then I would be simply changing the line
> > '/var/log/mgetty.log.tty* {' to be '/var/log/mgetty.log.ttyS4 {' which
> > would still allow error messages to be created for port S4. But,
> > these are the errors that I was >getting. What I really want to know,
> > is what causes the S4 port to start creating bunches and bunches of
> > empty log files in just a few seconds. Any idea?
>
> A bug in logrotate handling of the config file or in the making of the
> config file.
It's a feature 8^). Packagers who use logrotate need to be aware that
wildcards are wildcards, and will be treated as such.
>
> What happens is that the first time you have ttyS4, logrotate uses tty* and
> it matches ttyS4, it rotates it to ttyS4.1. So you now have ttyS4 ttyS4.1.
> Next time logrotate uses tty* and it matches both the current log and the
> rotated one, so it rotates both instead of just the current, ie it sees
> ttyS4.1 ad a valid log file not as a rotated log file.
>
> Don't know if recent logrotate versions can handle this better.
They don't. It's up to the individual packager not to rely on trailing
wildcards for matches. After raising this issue a couple of years back, I
am convinced that this is the most reasonable behavior. The bug is in the
packages, not in logrotate.
--
Matthew Saltzman
Clemson University Math Sciences
[EMAIL PROTECTED]
http://www.math.clemson.edu/~mjs
--
redhat-list mailing list
unsubscribe mailto:[EMAIL PROTECTED]?subject=unsubscribe
https://listman.redhat.com/mailman/listinfo/redhat-list
