[webmin-l] Timestamp in config files - was Re: BIND serial number style suggestion

[Carlton T. <ca...@gi...>]

Good idea adding the date/time based serial number to BIND config files.
But, as Danny said, what would be more useful is getting Webmin to show
the last modification time of the files.

That brings me on to an idea I had some time ago but never got round to
posting. I feel it would be very useful to have Webmin/Virtualmin add
timestamps to all configuration files that are created or changed.
These can be added in as comment lines where possible and if the file
structure does not permit comments, then maybe a supplementary file
could be created and comments added to that instead.

The date timestamp could include other useful information such as the
name of the Webmin user who made the change(s) and where possible, an
indication of the change(s) made.

These timestamp comments could be left in the files and would provide
a change history. It would also be nice to have a "show-change-history"
feature in all Webmin modules which create/modify system config files.
I know that it is possible to check the Webmin logs and glean most of
this information, but there is nothing like having it there in black
and white while you are looking at the file.

OK, my dream is to have a config management system built into Webmin
which would allow rollback of changes and other change management
goodies, but, like I said, thats just a dream.

Regards !

--
Carlton
=============================
GIFFORD INTERNET SERVICES
Bristol, United Kingdom
Tel: 0845 868 2245
Fax: 0845 004 6843
Email: ad...@gi...
Web: http://www.gifford.co.uk
=============================

On Wed, 2 Nov 2005, Danny Sauer wrote:

> On Thu, Nov 03, 2005 at 12:38:17PM +0000, David Derrick wrote:
> > Can an option be added to the BIND module to use the Unix timestamp as
> > the serial number rather than a date based serial number or a running
> > number? Sometimes it's useful to see the time a zone was updated.
>
> Not to downplay the usefulness of such a timestamp, but 1) unix timestamps
> mean almost nothing to me when *I* look at them and 2) ls -l shows the
> modification time in human-readable format... :)
>
> But hey, the field's size is 32 bits, and our favorite time_t is also 32
> bits, so I'd endorse it as another valid option (until January, 2038).  I
> think it'd be kinda neat to just show the last modification time in Webmin,
> though.  While we're on the topic.
>
> --Danny