A better open source backup tool

When planning a Linux backup strategy there are a number of options both in open source and from proprietary vendors. However, almost all of them are server-centric. This isn't so surprising, I suppose, considering the server market is much more mature than the desktop market, but Linux on the desktop is become more of a reality with each year.

I have been using a program called Simple Backup. It was created as an Ubuntu project for Google's Summer of Code. It gets the job done, and is pretty good for a project that was started and completed over the summer. However, I think Ubuntu (and other Linux vendors) need to take it a lot further. For instance, when I insert my USB hard drive I should get a notice asking me if I want to start a backup. There should be some built-in intelligence that will not backup the various caches on my system. After I configure (or re-configure) Simple Backup I should be prompted to backup the configuration to somewhere external (web storage, usb, even a floppy). Further if I have never configured Simple Backup before I should be prompted to import the configuration. If I have a system disaster, or I have to re-install, I shouldn't have to go though all the configuration again.

Now that Linux is getting more popular on the desktop we have to be more proactive about protecting our data. This came up for me because I accidentally formatted a partition that had five years worth of data on it (e-mails, contacts, documents, projects...). I was absolutely devastated. I know, I know... I should have had some type of backup, no matter how cumbersome. While I agree, how up to date would that backup be? If I trusted the magic of rsync, would I get the right combinations of command-line arguments and regular expressions? Would I inadvertantly delete something important in an attempt to keep my backups clean and compact? What about the complexity of time stamps and permissions?

My point is that we need to do better. I think Simple Backup is a great start. Now lets build off that.

Oracle can and will hurt Red Hat

I have seen several articles and blogs stating that Oracle could not possibly provide the high level of support that Red Hat provides its customers. These articles and bloggers opine that Oracle will fail horribly. I have to say that I vehemently disagree. Have any of these people talked with any Oracle or Red Hat customers, or read any research from analysts on this topic? It has been known for a long time that Red Hat customers are not very happy with Red Hat support. In addition, it is well known that it is often a bear to deploy Oracle on Linux because the Oracle DB only works with certain versions of the kernel, the system is very volitile on upgrade, and customers are loathe to upgrade unless they have to.

The title of this blog entry is probably a bit too strong, but to say at this stage that Oracle will fail is very premature. Oracle made this move for real reasons, and if they can execute on what they have set out to do they will cause real problems for Red Hat.

Would love to say more on this topic, but will have to leave that for later.

Does it work yet, Does it work yet, Does it work yet?

GNOME has the blog applet that is pretty neat. Anything to get me to blog more is good. ;-) However, the last time I tried it I would not publish to blogger correctly. Haven't tried it in about a year, but we'll see if this works...

Update: Nope. Still doesn't publish the title correctly. :-/ Oh well. Maybe I should look into contributing...