Yup, GForge rocks. I help admin CougaarForge for a government project; GForge has worked out really well. It's not resource-intensive, either - we run 34 projects and 700 users on one dual proc machine with _plenty_ of horsepower to spare.
Hm, maybe. But I think it's pretty much both - running CVS over NFS, whether by putting the local copy on an NFS drive or by using the local CVS protocol to access an NFS drive, is a bad idea.
Unfortunately I'm not enough of a guru to know who to blame it on - CVS or NFS - but I know how to avoid it:-)
> does not relieve someone from having to > download it
That's correct. But the beauty of BitTorrent is that once a few people download it, newcomers download it from them. So the original server stays available to serve page views and documentation and such, and the bandwidth usage is distributed. Good stuff!
> How much effort is involved in > setting up something like that?
Welll.... I've found it to be a decent amount of work. On the other hand, maybe that's because I'm not really a Unix guru. But you'll need to install PHP, PostgreSQL, Apache, Mailman, Sendmail, and various other components to get things running. It _can_ be done - but it can take a while to sort things out.
There are Debian packages available, and I think folks are happy with those - but I've never used them, so I can't say for sure.
> anyone packaging that for Fedora Core?
There are some spec files in the GForge CVS repository, but I'm not sure how up to date they are. I just installed mine "from scratch".
Splitting them up, I think. This is a good idea since TLM is still up if the Fedora tracker gets overwhelmed... but of course, they don't share peers, then.
Is it possible for trackers to share peer lists? Hm. Seems like it would work, since the pieces hash would be the same... hmmm...
Slashdot Mirror
....right here.
The activity graphs show a lot of recent movement, too... good times!
> GForge has all of this
Yup, GForge rocks. I help admin CougaarForge for a government project; GForge has worked out really well. It's not resource-intensive, either - we run 34 projects and 700 users on one dual proc machine with _plenty_ of horsepower to spare.
Lots of other folks are running GForge servers too...
This always gets asked - and a partial answer is right here.
Eclipse plugins, visualization tools... there's some good stuff there.
...when the man himself signed up for a user account. w00t!
> Thanks for the mirror
No problem, happy to help!
> much more to come
Very cool, rock on!
Hm, maybe. But I think it's pretty much both - running CVS over NFS, whether by putting the local copy on an NFS drive or by using the local CVS protocol to access an NFS drive, is a bad idea.
:-)
Unfortunately I'm not enough of a guru to know who to blame it on - CVS or NFS - but I know how to avoid it
> over the open source counterparts (CVS, etc)
> is primarily due to lack of formal support.
Here's a company that offers support contracts for CVS.
> there are more seeders than leechers ATM
Heh, yup, sweet. I'm seeing 115 seeders... jeepers.
If only that many folks were seeding on The Linux Mirror Project.
Is this a CVS problem or a NFS problem?
Either way, the solution is "don't run CVS over NFS". Use the client-server protocal - either ext or pserver.
> does not relieve someone from having to
> download it
That's correct. But the beauty of BitTorrent is that once a few people download it, newcomers download it from them. So the original server stays available to serve page views and documentation and such, and the bandwidth usage is distributed. Good stuff!
...right here. Only two seeds out there so far...
> NASA isn't dumb enough to use BitTorrent.
Why would using BitTorrent be dumb?
...since it's a 259 MB download.
What a perfect (and missed) opportunity for NASA to use BitTorrent instead of a huge zip file. Ah well.
> about 10K lines of Perl
Ah, true - you're right, it's not that much code compared to a scripting language equivalent.
> Less if we play golf.
Even fewer if we use Ruby! w00t!
...is on RubyForge right here.
It's 400K lines of assembly code... what could be sweeter?
...right here.
It's a bit out of date - it suggests you check out the code from CVS - but lots of good info there nonetheless.
> How much effort is involved in
> setting up something like that?
Welll.... I've found it to be a decent amount of work. On the other hand, maybe that's because I'm not really a Unix guru. But you'll need to install PHP, PostgreSQL, Apache, Mailman, Sendmail, and various other components to get things running. It _can_ be done - but it can take a while to sort things out.
There are Debian packages available, and I think folks are happy with those - but I've never used them, so I can't say for sure.
> anyone packaging that for Fedora Core?
There are some spec files in the GForge CVS repository, but I'm not sure how up to date they are. I just installed mine "from scratch".
> We're using it internally at my workplace
Cool. I help admin RubyForge; GForge has handled the load really well. Good stuff!
...right here.
It's powered by GForge, so it's backed by PHP and PostgreSQL.
There are a bunch of other sites running GForge listed here...
> Articles that fit in a subcategory that
> are determined to be important make it
> to the front page anyway
Hm, maybe... seems like the preferences would be the way to override that, though. But perhaps that explains it.
Or are these stories just not categorized correctly? Shouldn't this be "Politics", not "Internet"?
> Which SuSE version is this ?
9.1-personal
> SuSE aren't happy about their stuff being copied
Hm. Maybe that doesn't apply to the 'personal' version? Guess the TLM folks would know about that...
> the same order as the array is laid out
Cool, that makes sense.
> This allows compilers writers considerable leeway
Sweet. Thanks for the info!
Splitting them up, I think. This is a good idea since TLM is still up if the Fedora tracker gets overwhelmed... but of course, they don't share peers, then.
Is it possible for trackers to share peer lists? Hm. Seems like it would work, since the pieces hash would be the same... hmmm...