Data Corruption at ma.gnolia

This blog entry has taken me a while to post, largely because I really didn’t wanted not to have to post it at all.

I have been a huge fan of social bookmarking ever since i first heard of it. For me it just makes so much sense. I store all my bookmarks online, so whether i’m on my macbook, my office pc, my mobile phone or using someone else’s computer i have access to the sites i love.  But its more than that, i’ve signed up to a number of groups of like minded peers, and am now part of those communities. I can see the sites that my friends have discovered, and access those resources, quite often these sites that i come across in this slightly serendipitous way are the real gems.

magnolia logo

Having played with several social bookmarking sites i’d settled on ma.gnolia.com as the site that worked best for me. It was easy to use, worked graphically, had some nice tools, a great community, and i found Larry Halff the founder very likable.

Disasterously in early February magnolia’s suffered catastrophic data corruption. The database which was approaching half a terra bit (500gb) became corrupt. While backed up, the backup was simply backing up live data and not making archived duplicates go could not be roled back to an earlier version. After several attempts to recover the corrupt data, it was decided that there was no hope of recovery of the datastore. Larry managed to come up with some tools to recover individual users datastores from cached local and google data, and suggested other social bookmarking with good communities that users might migrate too.

Typically of Larry,  has been very open and very honest about the whole thing, sharing his experience so other people can learn from it. In the video below Larry Halff talks to Chris Messina of citizen’s garden about the history of magnolia, what happened with the data corruption and what the future holds for magnolia.

Larry Halff Discusses ma.gnolia.com
Larry Halff discusses ma.gnolia.com past,present and future with Chris Messina

I simply can’t imagine how much pain magnolia must be feeling right now, having spend to much time and effort to build up the site and the community surrounding it. While to ma.gnolia’s credit  i’ve never heard them once blaming anyone else or coming up with excuses. as an IT manager i have some idea how much pressure there must have been from users to maximize running speed and performance even at the expense of back-end services like backup and recovery.

I was lucky in that i had a recent backup and i was mirroring my bookmarks on delicious, what i miss mos though is the sense of community. I hope that magnolia does make a recovery and comes back bigger and better than ever, and that i and the rest of the magnolia community get an opportunity to help in whatever way possible.

One lesson that maybe we could all learn from this, is that we all need reliable back up plans of our own, and shouldn’t rely on internet based services too much because you really dn’t know when disater can strike.

Advertisements
Data Corruption at ma.gnolia

Facial Recognition for the masses?

Up until recently facial recognition was something of crime novels, or police super computers, but Google’s Picasa web album, image sharing service introduced its very own facial recognition technology this week.

The new ‘name tagging” feature, identifies faces within uploaded images and asks users to identify those faces with the name and email address of the subject. Picasa very quickly learns to identify faces on its own, and starts to automatically tag them whenever it sees them again.

The technology will undoubtedly take the hard work out of tagging photographs, and will make finding images of people much easier both from within the site and via search engine queries (if the images are marked as public).

But is it all getting a little “big brother”? and what does it mean in terms of our privacy. Do we as individuals have any control over what is being published about us on the internet?

Many employers are now routinely “googling” potential employee before asking them to interview, so it you have a myspace page, a facebook account or a blog you can maybe expect that to be seen, depending on what you’ve choosen to share or keep private. But you have a little less control when it comes to what photos of yourself that others have chosen to share,  maybe its a drunken stag night image, or a fancy dress costume you once wore, but do you really want a potential employer to see it?

In the past those photos may have existed, but they were relatively hard to find, because they were either not tagged, or where somewhere that obscure that only close friends would ever see them, but now there’s a chance they will be alot more visible, firstly because they will be tagged, secondly because picasa is part of google and therefore much more likely to be picked up by the Google search engines, and finally because when you look at the terms and conditions on sites like picasa or the new chrome search engine, you’ll see that all images uploaded give full copyright to the service owner, so should you choose to take an image down you may find its not that simple.

Facial Recognition for the masses?