Ed Brown wrote:
> Greetings,
> 
> With the release of 5.3, I'm looking to improve on my first mirroring 
> effort, and to really understand the different trees.  I've studied the 
> documentation on the website, and searched mail archives, but am still 
> confused about what's best or what's necessary.  This page,
> 
> https://www.scientificlinux.org/download/mirroring/mirror.rsync
> 
> suggests just the 5N tree, while the ftp guide,
> 
> https://www.scientificlinux.org/download/mirroring/mirror.ftp
> 
> shows 5x AND 5rolling being mirrored.
> 
> The excludes on the latter page, by the way, I think need to be updated, 
> they show "errata/obsolete" and "errata/debuginfo" when I think it 
> should or could be simply "archives" for 5rolling and "archive" for 5x 
> (yes, they're different!?, and they only contain obsolete and debuginfo, 
> so no reason to list separately).  I'm also thinking "SRPMS" and "iso" 
> aren't really necessary for now.
> 
> So (using lftp) I started with (for 5rolling):
> --delete --exclude archives --exclude SRPMS --exclude iso --exclude 
> sites/Fermi
> and for (5x):
> --delete --exclude archive --exclude SRPMS --exclude iso -exclude 
> sites/Fermi
> 
> But there were problems with 5x because the 53/ directory did not exist 
> and the links were therefore broken.  And if I was excluding archive and 
> SRPMS anyway, a directory with just symlinks to another didn't seem very 
> useful.  So now I'm thinking I really only need/want to mirror 53...
> 
> Is this the idea: if I mirror 5x, then I don't have to change my 
> mirroring script when a new update level is released, or is there more 
> to it than that?
> 
> And why mirror 5rolling, is it to have access to the early release rpms 
> in testing/, and/or early access to update level releases?
> 
> 
> thanks for SL!
> Ed

Hi Ed,
First off, the example on those pages are just that, examples.  But I 
will change the ftp one, I don't know why it's pointing at 5rolling, 
because there really is no reason to mirror it unless you are testing a 
beta.  And, now I've changed it to use 4x and 5x as the examples, and 5x 
is fixed so that the exludes are for the right things.

Second off, rsync is a much better way to mirror in my opinion that the 
ftp method.  It is much easier on the network, and usually faster.  Just 
remember to use rsync.scientificlinux.org.

Third off, it's up to each mirror site to determine what they want to 
mirror and what not to mirror.  On
https://www.scientificlinux.org/download/mirroring/mirror.tips
I give a list of the most common, but there certainly are more.  SRPMS 
and sites/example are two other common ones.  But it really is up to you 
and your users.

Troy
-- 
__________________________________________________
Troy Dawson  [log in to unmask]  (630)840-6468
Fermilab  ComputingDivision/LCSI/CSI LMSS Group
__________________________________________________