OSNews

Someone has used Amanda and Bacula and can tell us which is better? I've read about both and Bacula seems more professional, but it's only my feeling. I know Amanda is widely used, but that not means better product.

One thing about Amanda is that it is rock-solid. Sometimes it has worked for me even in face of hardware errors. It doesn't have a pretty GUI (but I don't use GUI's anyway). I have used Bacula a bit, and didn't see a reason to switch to it.

One big advantage that Amanda has over Bacula is that it uses unix native utilities for storage of data (tar and dump). So, even if Amanda engine is not working you can use your age old tools to recover. Bacula uses its own (somewhat strange) format.

This is considered a big enough advantage at our site.

But Bacula's founder, Kern Sibbald, is committed to format stability. The format is documented and therefore I think you can read your old taped even in 10,20 years (if your hardware is in place, that is).

There are statically linked restore tools and even a rescue disk so you can do a bare-metal restore.

fs

Can Amanda backups Windows volume? It seems critical on my site.

Yes Amanda can backup Windows. It uses samba to access the windows volumes to be backed up.

Using Samba for Windows is not what I would consider a "Windows backup". What about sparse files? Backing up windows databases with native tools (Exchange for example)?

fs

I don't know if this is still the case but in the past Amanda was not able to perform a backup if the data doesn't fit on one tape. Bacula will handle this case and requests more tapes (or uses your autochanger to change the tape).

fs

Amanda 2.5 has tape spanning feature to go beyond one tape.

In my expirience, Bacula is far more mature project. I spent whole month making Amanda work something, untill one guy suggested to try Bacula.

Amanda is ok for simple backups (when I say simple I don't mean backup your home dir), but when you need full hard time working backup (couple of sets of tapes, prexec & postexec scripts, different conditional backups, etc...) it's Bacula. Bacula is even better then some fancy solutions that cost a lot of $$.

Yes, it is a energizer bunny. Good to see some new activity in the project.

Did you even read the full article?

Anyway, lion share of open source programmers are paid software developers.

"Ok, first off, should the title of a section be what's tossed on at the end as addendum?"

I think you forgot to click on "Read More".

You can make money on:
1) Selling support services/contracts
2) Selling custimizations to the software
3) Selling traning class'
4) Selling books

These are just a few off the top of my head. FOSS, has a different revenue model. Its based on selling support; rather than shrink wrapped goods.

Some examples would be:
1) Snort
2) Sugar CRM
3) Red Hat

Any ways, just my .02 cents worth.

1, 3 and 4 are all predicated on the assumption that your software is difficult to use, which is hardly a big attraction for people when there are easy-to-use commerical alternatives with just a once-off payment.

2 is also difficult, as a customisation could be considered a derivation of your product, which would put you in trouble if you used the GPL.

You could choose not to use the GPL, but then someone else could take your product, make it easier to use, and sell it one themselves under a proprietary license, essentially destroying your market.

I think open-source is great, I think being able to create dervied works is great, but I think free-of-charge distribution makes it very hard for small programming houses to make money. My ideal license would be open-source, prevent source-distribution, but still allow the distribution of patches and add-ons. That way the original author would be paid by every new user, enabling them to continue working, while a cottage industry could grow up around the product selling add-ons.

The difficulty is legally distinguishing between an add-on and an entirely new product, especially if it becomes possible to re-create the product's entire source solely from the patches/add-ons.

Open-source is good for infrastructure, indeed it's preferable. But as for end-user products I'm not convinced. The Nessus story is pretty interesting in this regard: http://it.slashdot.org/article.pl?sid=05/11/26/1629237&tid=172&tid=...

I don't know anything that compares with Veritas except TSM, that even surpases it IMHO. But both are comercial products.

Veritas has gone downhill while Amanda is going up.

Most Veritas customers use just basic functionality but pay ridiculous money for licensing and support. Amanda is already rock solid and good enough from functionality perspective. With commercial support for Amanda there will be no need to buy IBM or Veritas backup.

Obviously you have no experience with enterprise class backup systems. There is no comparing Amanda to TSM or Veritas Netbackup period.

That’s what Sun was saying about Linux just a few years back. Today even Sun admits that 90% of their own x64 servers are running Linux and not Solaris.

Thanks. I was just reading these comments in hope of good information. I will give rsnapshot a try because I really need a backup system now that my workspace is 600 MBs, without considering other uses which could make it handle a bigger 3 GBs files or more.

I really like Rsync and rsnapshot should be good.

Now, on the "I use Windows and I love it you open source zealots". Well, we don't really have a problem with you preferring Windows. But you have a problem with we developing Open Source tools or using them. So the problem lied in you and still lies in you. You behave as your worst nightmare users for us. You know, those users that sometimes the developers like to complain about.

Few developers bite the bullet and open their firms. So it's bogus to pretend that you rule your proprietary world.

I think you are ignoring that robustness is the most important thing to consider while picking a backup software. This is where Amanda excels. BTW, I find Amanda community to be very helpful as well.

I agree with you 100%. Amanda is definitely the way to go. I also second the comment from another user about people paying too much money to Veritas for limited functionality needs.

Not to forget the transport encryption feature: All traffic between all Bacula daemons can be encrypted with TLS. Data encryption is being worked on.

fs

Amanda 2.5 has encryption as well...