Hello,
This is in response to the email from Jesper (see below). As it is
not always obvious, I am not in the least upset in any way. This is
meant to be information about our future direction, and more
directly a response to Jesper's concerns and questions.
I have devoted 12-16 hours a day many times 7 days a week for 12 years
to build Bacula, so I have no intention of see Bacula fade away. I will
be around for
a bit longer (I am well over retirement age). That said, I have to
think about
who is going to take the relay in some years, and my experience with Bacula
Open Source is that the community is *extremely* important in testing,
finding bugs,
sending in small patches, and providing good ideas, but it is poor in doing
major design and programming projects. First, I am not in the least
complaining,
and second, there are some exceptions. We unfortunately reject a
large number of submissions or completely re-write them because the authors
do not follow our standards of coding (checking error returns, having
reasonably
clean code, test cases, documentation, and a good design that won't put
Bacula
design wise in a corner in a few years).
After long thought, my conclusion is that the only way to ensure that
Bacula
continues is to have a commercial company behind it. To have a commercial
company behind it, the company must offer services and be able to sell
those services in sufficient quantity to be at least break-even. For
the first
two years of Bacula Systems existence, we provided services on the same code
base as the community, and we funded a *large* number of projects. Bacula
Systems came close to going under. Customers willing to pay, were not
willing to pay
*only* for service, they want features that are not in the community
version.
Frankly that surprised me.
We changed direction several years ago. We now attempt to differentiate
Bacula Enterprise from Bacula Community by adding high end features that are
Enterprise of nature, and the company is doing well.
A community member does not need SAN shared storage for
example -- or if you do and you consider yourself community, that is
fine with
me, but you will need an Enterprise contract. There are a *lot* of
"enterprises"
out there that given a community version, refuse to pay for an enterprise
contract (mostly ISPs, Universities, governments, and small businesses).
This is fine; it is their right. However, we developers are not
obligated to fix
their sometimes *extremely* complicated bugs or implement their needed
features.
We often do so, but we are not getting paid for that so anything along
that line
that you ("enterprises") get please be happy for. I think everyone of
the Bacula
developers have expressed this feeling at least in some degree. We work on
it mostly for our own needs (mine were satisfied *many* years ago) or
the pleasure.
Most Bacula developers come, and then go (often because they change jobs).
I am pleased with everything they do for Bacula.
You ask about development fees. Well over the 12 years I have worked on
Bacula,
the project has received a total of about $5,000 in cash contributions.
No complaints.
I appreciate those contributions enormously because they represent
something
that very few people do -- thank you contributors.
To have a full time developer on board when everything is considered
about $150-$200 K
per year per developer, and in my opinion, the Bacula project needs at
least 5 developers,
so you can see that the community simply does not have the means to pay
contributers.
That is not even something I am going to try to do again (Bacula Systems
essentially
tried for 2 years unsuccessfully). Now, if you calculate based on the
cost of a developer
a two month project would cost Bacula Systems about $30K without any
margin at
all. One can outsource to various countries such as India at
approximately one
third the price, but then you end up with a pile of code that may be
*very* difficult
to maintain (we have seen this with several of the large community
contributions, which
were developed by experienced people).
For the future: Most of the major features that I personally am
developing are going
into the Enterprise version. All of the bugs fixed by Bacula Systems
are going to
the community. Most all the small or non-Enterprise features are going
in to the
community. All the community patches and other contributions that make
sense to put in
Bacula are going to the community (unless it is from a Bacula partner
who wishes the
code to go only in the Enterprise version).
Bottom line:
Bacula as a project has all the features that Community users need.
Bacula for Enterprise
needs a good number of high end features (some of which have proprietary
interfaces --
Oracle, certain Microsoft code, ...). The Bacula Enterprise features,
unlike the community
features, are far from complete. Over time when the current
differentiating features become
less important, they will be pushed back into the community. If you
are a community
member, then Bacula has everything you need. I know, because I use it
here for my
10-15 computers, and I do not need any of the Enterprise features.
If you are on the boarder line or well over it toward needing Enterprise
features, you
have a number of solutions:
1. Develop the code yourself
2. Encourage the community to develop it
3. Offer a bounty
4. Go without the feature
5. Pay for an Enterprise contract
6. Someone can fork the project, but in my opinion, in that case the fork
will be dead within 5 years, because it is extremely difficult to
write
fast and robust backup software that works across many platforms (I
have been
doing it professionally since 1972).
Note, I no longer have the time or energy to attempt to seek funds from
the community.
I did it, but in general it doesn't work (at least for me). I will do
anything I can to help
such an effort though.
Bacula Systems is not taking anything away from the community. On the
contrary, it is
providing the community with a lot of services that would not otherwise
be there. Just take
a look at the new features in Bacula 5.2.x. Look at how many bug fix
versions we put
out since 5.2.0 -- that takes a lot of work, and is possible only
because there are
salaried persons working on it (free for the community).
I am not sure how contributions of many new features by Bacula Systems in
versions 5.0.0 and again in 5.2.0 can give you concerns about quietly
killing the
community version. Your actual words were to quietly kill the Open
Source version.
For your information, the Enterprise version *is* Open Source. However,
we distribute it
only to customers, which is 100% in conformance with the AGPLv3 Open Source
license it has.
This is in general, not just to Jesper:
You may not agree with me. That is OK. If you do, that is even
better. Please don't
get upset though, there are enough things in the world going wrong that
it is
not worth the effort to complain to someone who has and is giving you
lots of very
useful *free* software.
Best regards,
Kern
On 03/23/2012 07:10 AM, Jesper Krogh wrote:
On 21/03/12 10:27, Kern Sibbald wrote:
We have a project planned for Bacula Enterprise where a Job is blocked
because it is despooling will be able to continue spooling providing
there
is sufficient space. But this is an Enterprise feature, and unless
someone
from the community implements it, it will only be in the Enterprise
version.
In fact, we have a series of plans for adding high end Enterprise
features
to the Storage daemon. Most of them should be completed by the end of
the year (hopefully before).
Hi Kern.
Awesome. This is some of the key features Bacula has been lacking.
Others are: #10 and #29
It is your call (or Bacula Systems), but I really would suggest that
you'd try to push a mail to the Bacula-mailing list of the amount
of direct funding needed to just develop these features directly
in the open source version of Bacula. My feeling is that you're
silently killing of the Open Source bacula by following this path
as strongly as you do.
I like and enjoy working with Bacula. Bacula has been a corner stone
in our setup for more than 7 years, pushing over 1 petabyte of data out
to tape, using different autochangers, LTO-generations and been adoptable
and flexible along the way. 1)
As with all other compontent in the IT-setup the amount of work time
and money there has gone into Backup over the years is significant.
The problem is, I dont think there is a single person on this planet
running bacula in an non-Enterprise context. The amout of work,
hardware and time needed to run
a decent backup system with tapes and autochangers (which
is the corner where Bacula is truly awesome), is highly
overlapping with the enterprise segment, so the money is there.
Now you ask yourself: "What the hell is this dude's problem?"
The problem is, had I evaluated Bacula for out setup today,
compared to 7 years ago. I dont think Bacula had won!.
The one true strength Bacula has compared to commercial alternatives
is the viral community on this list, that creates a true trust in the
software
that a lot of smaller commercial alternatives can't compete with.
The smaller and more restricted the "Open" version becomes, the
more this trust is sacrificed along the way and in the end it is "only"
the pricetag that differes to the commercial alternatives. And, sorry,
even though Bacula is cheaper it is going to loose this one big-time.
In a Backup solution in enterprise context, the price of the software
itself is only one component of the one-to-one comparison and the
commercial alteratives are going to beat the crap out of Bacula on a lot
of others, and the Enterprise version is "looking more and more" like
"Just another commercial alternative".
So, please, keep an eye on it, dont silently kill of the Open Source
version.
A proposed middleground is:
* Try to seek funding for the features, instead of restricting them.
* Push harder on what you are truly good at, support and services.
1)
http://opensourcedays.org/2012/sites/default/files/Bacula%20-%20Experience%20from%20a%20medium%20scale%20real-world%20setup.pdf
Thank you for your attention.
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
Bacula-users mailing list
Bacula-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
