Hi John,
This is exactly the approach we use at boston.com, a large regional
news site, serving roughly 3 million pages/day.
For historical reasons, most of our web servers for serving static content
are Netscape Enterprise servers, but we use modperl for all of our new
development of dynamic sites (see http://mp3.boston.com/, the code for
which we're planning on releasing as open source "real soon now".)
We typically run the server root off of a local disk, to keep log traffic
and other server accesses off the network. We have a shared document root
that lives on the NetApp, all of the web servers read this. (The document
root is mounted read/write, mostly because I didn't set it up originally.
If I were setting this up from scratch, I'd probably try to have the web
servers mount it readonly, with only the content upload hosts having write
access.)
We're running our perl include tree off of the NFS device also, which may
or may not be a good idea, but it sure makes it easy to keep CPAN modules
up to date.
We've been extremely happy with the convenience and reliability of this
setup.
You *do* need to be aware of the differences in locking semantics across
NFS devices. We're running mostly Solaris, and while I don't have any
proof that it doesn't work, I have seen some strange conditions that lead
me to believe that 'flock' doesn't always behave correctly across the NFS
device. (Anyone who can provide proof otherwise, I'm all ears...) Atomic
renames seem to behave as expected, so you can roll your own lockfile
based mechanism if that's an issue, but we lately have taken to using
MySQL to store shared data so that normal DB semantics can be used to
coordinate distributed processes.
Also, you need to worry a bit about OS versions of the hosts running your
web servers -- binaries on the NFS device compiled for Solaris 7 aren't
going to behave well if one of your web servers is running 2.51 (or even
Linux!), for example.
Good luck.
<Steve Reppucci>
On Mon, 31 Jan 2000, siberian wrote:
> Hi All-
> I am building a pretty in depth architecture for our new service
> using ModPerl. I've done a lot of large scale/high traffic apps in modperl
> before but never in conjunction with a network attached file server. I am
> thinking that it would really make my life easy to have one central
> repository of code, databases and sundry files that all the servers share
> ( making it easy to swap out servers, add servers etc since its a central
> file repository that everyone just hooks into ).
>
> My question is : Has anyone experienced any 'gotchas' in putting perl code
> that modperl handlers use on a Network Attached file server like a network
> appliance box ( www.netapp.com )? I am assuming that there are no real
> issues but before i go blow a ton of cash on this thing I wanted to be
> sure that no one had found a problem.
>
> Thanks for any pointers.
> John Armstrong
>
=-=-=-=-=-=-=-=-=-=- My God! What have I done? -=-=-=-=-=-=-=-=-=-=
Steve Reppucci [EMAIL PROTECTED] |
Logical Choice Software http://logsoft.com/ |
508/958-0183 Be Open |
