Re: current 1.3.1 testing

Adrian -

        I've embedded replies to your specific questions below...

On Sun, 30 Jun 2002, Adrian Miranda wrote:

> I'm starting to use current 1.3.1.  Since we're in the process of
> rolling out fairly large numbers of Linux hosts, the features of 1.3.x
> are a big help.  I had some trouble with the docs, but I eventually
> got it working fine.  In particular, I'd suggest expanding the section
> about generating ssl keys for apache to explain that you'll need to
> enter a password every time you start it (or need to generate an
> unencrypted key).
        This is probably something I should have thought of earlier,
actually.  I've definitely got to flesh out the docs a bit more...

> Anyway, one thing I noticed is that some stuff seems to want to write
> to /var/log/httpd/current.log as indicated in the docs, but some other
> stuff seems to want to write to /var/log/current.log?  I'm seeing
> messages like the following in httpd/error_log:
> 
> 	Using log_file /var/log/current.log
> 
> Is this configurable somewhere?
        This may not be as thoroughly documented as it will be.  The
mod_python version looks forst for an entry in the global section of the
config file called "apache_log_file" and tests for it's writability.  If
it can't write to that file, or the entry doesn't exist, it looks for
the old "log_file" entry, and tests it.  If it can't write to that file,
the default logfile is /var/log/httpd/current.log.

> Also, can you continue to use current 1.3.1 in the old stand-alone
> mode?  I assume you can't use both modes at the same time.
        You should be able to do so, but it will require some magic on
the front-end in setting up the same SSL keys in multiple places on the
server.  I don't know that I've tested the standalone frontend in recent
1.3 work, but I didn't deliberately set out to break it.

> I understand I still need to run "cadmin create channel" when I add
> new RPMs.  Is there something I should do to prevent current from
> getting run during that time?  Probably I should shut the apache
> server down?
        Right now, shutting the Apache server down is the only way to
ensure there aren't issues with cadmin and the server running at the
same time; we're trying to figure out how to best do some sort of
process locking now.

> Any suggestions on how I might go about setting up redundant servers?
> I suppose I could just make sure one host mirrors the other, and move
> the DNS entry if/when the primary one fails.  I'm just curious if there
> are some nicer solutions that aren't too hairy.
        There's not really a facility for redundant server built into
Current, and if there ever is going to be, it's a long way down the
road.  The best thing to do right now is what you mentioned - mirror the
server hosts, and make sure the databases on both hosts are updated at
the same time, then use DNS to initiate a manual failover if necessary.

-- 
Thank you,
John Berninger

Systems Administrator		[email protected]
Department of Mathematics	Box 8205, Harrelson Hall
NC State University		Raleigh, NC 27695
Phone:  (919)515-6315		Fax:	(919)515-3798

GPG Key ID: A8C1D45C
        Fingerprint: B1BB 90CB 5314 3113 CF22  66AE 822D 42A8 A8C1 D45C

"Non illegitimi carborundum."
--