Slashdot Mirror
Andrew Tridgell Talks About The Future Of Samba
Spud writes "Andrew Tridgell tells us what Samba is up to in a new interview." Specifically, he talks about several new features planned for Samba 4.0, and says that the release of 3.0 will happen "soon."
Andrew obviously know his stuff when it comes to networked filesystems. In previous interviews he has let his disgust with CIFS/SMB be known. I wonder if he has any plans to design and write a new FS to compete with the defacto standard NFS.
One thing at a time. Windows needs to die first, then Samba will wither away through lack of interest.
-- Alastair
I especially liked the anecdote about Linus hacking a crashed server to find out what went wrong. Just goes to show how damn proficient the leader of the Linux world is.
:). Maybe they'll have Fairy penguin nipping sessions there 'specially for all the visiting geeks. Beats getting mauled by a koala (those things are real killers!).
I have now added the National Zoo & Aquarium to my list of must-see Canberra places (wow, my list has doubled
Linuxworld seems to be Slashdotted, so my knowledge of Tridgell is limited to the fact that he's reverse-engineered two well-known protocols: SMB, and the Tivo subscription thingee. What else has he down? Or should I ask what else has he reverse-engineered?
...most readers have missed the significance of Samba4. Samba3 is already a better Windows fileserver than Windows, in several ways. Samba4 will give Tridge and team the flexibility to nail down every corner when it comes to compatibility and performance. It wouldn't surprise me at all to see features like distributed file stores appear out of this; y'know, rather than buy a $30,000 obsolete-tomorrow monster (and spend another $30-60,000 for MS-seats - "here's your frame, motor, panels and wheels, sir; and how many people will you be buying seats for?"), you buy three reasonable $5000 boxes and get better performance, automatic load-sharing and failover.
Got time? Spend some of it coding or testing
Frankly. you're right (and don't deserve the flamebait mods). And the ironic part is that Samba is one of the best run projects around. I have tremendous admiration for the entire team, I think they're some of the sharpest folks in the community. I admire Jeremy Allison almost more than I admire Linus, he's savvy and quotable. It's an excellent project that is solid, reliable very useful, and gains a lot of admiration for OSS in general. And what they're doing is hard. And yet I agree with you that Samba needs to die. (After helping to kill Windows, of course.)
But I'm not really worried. These guys are sharp, and after it becomes time for Samba to die (which will not be as soon as most of us might like), I'm sure they'll all find other interesting things to work on. And in the mean time, they seem quite happy to work on this currently-still-very-important software for us. Kudos to all the Samba team.
Remember the Microsoft TCO bullshit campaign? It was mainly targeted against Linux as a replacement for Windows SMB servers. Each computer running linux/Samba is:
1. A computer running linux.
2. A computer not running windows.
Now, 1. gets the management to see that Linux does well in their environment. This doesn't have a measurable financial effect, but influences future decisions. Samba gets the penguin's foot (paw?) in the corporate door. Point 2. cuts directly into Microsoft income.
To sum it all up: Go, samba!
Every now and then, I go looking at the network/distributed filesystems (particularly looking for disconnected operation) and every time I find there's still something majorly wrong with all of them. I can't think of anyone I'd trust more than the Samba team to bring that kind of thing out of the research/toy department into the real world. Not just because they know network filesystems, but because they have all this experience in maintaining a solid product of such great importance to its users.
So I'm hoping that what these guys work on after Samba is ...Samba!
Fuck the system? Nah, you might catch something.
Huh? I always thought samba was the name of a snake. Or at least it was in "Temple of Apshai" :-D
And to think, it required a whopping 256K of RAM! But it did support the bleeding edge CGA technology...
Government IS the problem.
I'll feed this troll enough to point to this definition of samba which includes the program Samba. I've always associated it with the dance beat samba.
Please move on, nothing to see here.
I am, and always will be, an idiot. Karma: Coma (mostly effected by
You're thinking of sambo , moron.
I've just completed a Samba3 Winbindd setup to synchronise some pop3 users+pws with an NT PDC. I have to admit it was a non-trivial task and i found myself scurrying to the 3.0 Head docs frequently.
I had been hoping to implement a samba BDC to an NT pdc but i was not successful in my efforts. It is trivial to configure a Samba filer but DC Replication is a still a work in progress. I will wait for the efforts to mature and contemplate the move to a Samba PDC.
I cannot but extend massive kudos to the Samba team et all for getting this far without access to the source code.
What are you people thinking? What will replace Samba? What will replace M$ if Samba kills it before it goes away?
Samba is far more stable and much easier to use/secure than NT4 ever was and W2K ever will be.
I can only assume you have no idea what your talking about and your experience with PDC/BDC's extends only as far as a /. article.
go away, you make us look bad.
----
--- Just say no to negativity.
Yeah, a troll but I'll bite.
1) The racist term you're referring to is Sambo, not samba.
1b) This is more of an american term, and Tridge is Aussie, so he's probably unaware of Sambo anyway.
2) samba is a dance style and a snake.
He named it by thinking "well, this does smb" and picking what he liked from
grep s.*m.*b
in a dictionary file.
...SaMBa uses more memory but will handle more sessions on the same box.
Got time? Spend some of it coding or testing