Re: multiple servers

--- "gshennessy <[email protected]>" <[email protected]> wrote:
> I've configured multiple servers, did the 
> ubhdb, ubhasm, ubhcatalog, and now I'm trying
> to download some files via 'ubh -I "filematch"',
> and it seems to only be pullling from
> one server.

What does the priority column say?  Are all your servers set
up to be the same priority?

You want to set the priority so that higher priority servers
have higher priority numbers.

The concept of "higher priority" in ubh means that if an article
is available on more than one server, then ubh will attempt to
download it from the server with the highest priority; if it is
not available on the server with the highest priority then ubh
will try the server with the next highest priority (and so on)
until all servers have been tried.

I realize that the preceding two paragraphs need to be included
in the setup documentation.


Typically, you will make your free or cheaper servers higher 
priority (eg, your low-retention ISP server), and then put your
expensive servers at lower priority so that they are used for
fills.  You might also want to put your slower servers at a
lower priority as well.

As a performance tip, you will want to set your retry count
on your higher/highest priority servers to be very low (eg, 0 or 1)
so that ubh does not waste time/effort re-trying to download
non-existent articles.  Example: you have a ISP-provided news
server which has ridiculously low retention (say less than 24hrs)
but you still want to try this server once before going on to your
pay server.

> 
> Is that the intended behaviour? 

Yes, as designed.

> When it starts
> up it says it connects to all three servers, 
> but I'm not getting the bandwidth I expect.

Chances are your highest priority server has more articles
than the lower priority server, so the lower priority server
never gets consulted during downloading.

The ubh downloading client is a single-threaded application
which pulls one article at a time from one server at a time,
according to the priority scheme explained above.

You can run multiple simultaneous copies of ubh, but they will
each adhere to the priority scheme.  ubh will *not* distribute
bandwidth load among the connected servers, which if I read
your message correctly that was what you were hoping for.

This was implemented intentionally this way to give you control
over where the articles come from, so as to get articles from 
cheaper (or possibly faster) servers before consulting the more
expensive (or possibly slower) servers.


Hope this helps.  I will fold this info into the documentation
prior to the general release.

-Gerard


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Turn flat surfaces into speakers with the Soundbug.
http://us.click.yahoo.com/QWAVSC/onCFAA/xGHJAA/IHFolB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
[email protected]

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/