Thursday, May 10, 2012

Innovator-C: Approaching 400 Days of Up Time


IBM Informix Dynamic Server Version 11.70.FC2IE -- On-Line (Prim) -- Up 398 days 01:11:10 -- 2084080 Kbytes

Put into production over a year ago in April 2011, Innovator-C has just been cruising along with no problems and no signs of slowing down.

Just a single socket, quad core Intel machine with 2GB allocated to Informix and a couple of internal drives mirrored RAID1 using cooked chunks, Innovator-C is able to handle an average of 700,000 new session requests, 4 million commits, 900K rollbacks and 3.5 billion ISAM calls per day. This is just what our production load throws at it, I don't know what the actual limits are. Judging from CPU usage and current response times, Innovator-C running on this lower end hardware can handle much more.

Makes me question why someone would choose a different free or low cost engine for smaller installations, especially considering there is a well defined upgrade path to Growth and Ultimate as your need for speed increases. It definitely can't be because of technical reasons. Most likely they don't even know Innovator-C is a free for production option. Heck, there were people at IIUG2012 that didn't know Innovator-C was a free for production option and these are the Informix Fanatics!

You might have noticed the On-Line (Prim) in my 'onstat -' above, Innovator-C comes complete with full HDR Primary/Secondary support so even the smallest of systems can have enterprise class redundancy, even though I haven't had a need for it since the engine is so stable.

May 14, 2012 Update:

IBM Informix Dynamic Server Version 11.70.FC2IE -- On-Line (Prim) -- Up 402 days 00:41:51 -- 2084080 Kbytes

Made it past the 400 day mark during Mother's Day weekend, our busiest time of the year. Of course, Informix handled the 2x spike in volume without any problems. Yawn.

6 comments:

  1. Hey! what did you expect?
    Others RDBMS are hardly trying to reach 400 hours of continuous uptime on much bigger configs, not to talk about performance.
    Let's talk again at 800 days, and BTW, according to my calculations, 400 days should be today ;-)

    Champagne!

    ReplyDelete
  2. what is the hardware specification?

    ReplyDelete
  3. @Eric

    Unlikely we will reach 800 days. Not because of Informix, but because of our internal forced server reboot policy. Not sure how this set of machines has flown under the radar, but I'm sure they'll be pinging me for a failover/reboot soon. It stinks that these reboots kill what could be years of uninterrupted Informix engine online times, but it is a good practice that forces us to failover to the secondary so we know that the A and B side of the HDR pair are configured correctly in case there is a real outage (hardware related, of course :) and we need to perform a real failover to the secondary.

    @Anonymous

    "Just a single socket, quad core Intel machine with 2GB allocated to Informix and a couple of internal drives mirrored RAID1 using cooked chunks", what more do you need to know? I could probably run Innovator-C on a toaster and it would still be able to handle the load I'm throwing at it.

    ReplyDelete
  4. any tutorial for connection manager for failover?

    ReplyDelete
  5. I gave a presentation at IIUG 2009 on the Connection Manager, check out the member's area of iiug.org for the slides until I can get a blog post up.

    ReplyDelete
  6. @andrew, Thanks, I can't find the tutorial in IIUG. Waiting for your blogpost.

    ReplyDelete