New OSgrid system passes test

A new data replication has successfully passed its first round of testing, OSgrid grid administrator Allen Kerensky announced on Monday. However, there was still no word as to when the grid will be coming back up.

In addition, the grid still hasn’t received its recovered data drives from the company that did the recovery work, so there is no news about whether user inventories have been saved or not.

Kerensky promised another update in four to five days.

OSgrid's LBSA Plaza is the cross-roads of the hypergrid.

OSgrid’s LBSA Plaza is the cross-roads of the hypergrid.

New asset servers will replicate automatically

OSgrid has almost as many registered users as InWorldz, the busiest commercial grid, with several times the regions. That adds up to an asset database of about 3.5 terabytes — or over 21 million individual files.

Instead of replacing the old failed asset server with a new asset server, it will be replaced with two — the second one a real-time clone of the first.

That way, if one asset server goes down, the second one is ready at any moment to step in and take over. That failsafe system is now in place, and has been tested.

“Both new asset servers were fully configured and brought online temporarily working as a master-master replication pair,” Kerensky explained inMonday’s update. “This was the first full-scale test of the new database replication… and it was more successful than wed hoped.”

A long, long wait

OSgrid has been down since mid-August. Known as the “crossroads of the metaverse,” OSgrid was the oldest and busiest of the non-profit grids, and was the primary testing grid for OpenSim developers.

Since the outage began, OSgrid’s region owners have had to scramble to figure out what to do with their regions. Some have connected to other open grids, while others have decided to wait it out.

Of those who moved, Metropoliswas the primary destinations. Other open grids that accept self-hosted regions include German Grid, FrancoGrid, Craft,Atek, and WestWorld.

 

maria@hypergridbusiness.com'

Maria Korolov

Maria Korolov is editor and publisher of Hypergrid Business. She has been a journalist for more than twenty years and has worked for the Chicago Tribune, Reuters, and Computerworld and has reported from over a dozen countries, including Russia and China.

  • So THAT’s why I haven’t seen Allen around. He’s been developing the Myriad RP System that I like and use, but I’ve seen nothing new in many months. Now I know why, he’s been helping fix OSG. Didn’t know he’d been made Administrator, either.

    I wonder if he’d look into why I was banned without explanation more than a year ago

    • Han Held

      >I wonder if he’d look into why I was banned without explanation more than a year ago
      Two years ago, and it’s astronomically unlikely.

  • nextreality

    Its been such a long time since OSG went. I don’t think people really care anymore to be honest.

    • Hans Nerido

      to be honest Mike, I don’t believe we osg users want change to drama grids 🙂
      in the last month I meet a lot osg users on other grids,
      they are thankfully other grid owners adopt them, like me on tangle grid, but we cant wait to go back home on our “test” grid

    • Lara Nguya

      Lara cares. She is so hoping that she can persuade her friends to come back to OSGrid when it is up and running.

    • Sammy Greenway

      I care, and have my regions ready to start up on a moments notice. 🙂

  • Tesla

    Guys and Gals …. a online backup and a raid suffer from the same mistake….Don’t make the same mistake twice… http://en.wikipedia.org/wiki/Backup_rotation_scheme … If some installs a malware or your data gets corrupted you will find out about it after it have been thoroughly replicate to all your online backups. Do it right this time… and yes it takes lot of storage but do it right. Secondly encourage and allow everyone to make IAR.

  • Tesla

    By the way, you can get several terabytes of offsite backups for $40 a month from various sources. If you had the same problem again, it should take you days if not hours to recover. NOT months. Always plan for the worst, that is how you run a datacenter/webhosting or other business that others depend on, whether for fun and/or profit.

  • Susannah Avonside

    I’m not sure that I’ll be relocating my regions on OSG once it comes back as I have successfully been running my regions in HG enabled standalone mode since the crash, and as HG works fairly well most of the time there is little reason to re-establish my regions there. I am fortunate in that I have pretty good bandwidth with my internet connection, which makes a huge difference, but I appreciate that this option isn’t open to all.

    I hope that OSG does make sure that it has a robust backup procedure in place, and that they don’t merely rely on the system they are currently testing. Though unlikely. the fail-safe could itself fail when it’s needed; there is no substitute here for belt and braces, plus perhaps also a couple of spares in the wardrobe. Once thing that I have taken from the failure is the need to regularly make backups and store them somewhere safe – I have an advantage as the amount of data I have is nowhere near the 3.5 TB that is the asset database in OSG, but it is nonetheless equally catastrophic loss when all that work disappears for want of an adequate back-up policy.