Rebooting Venue

Question: What is the recommendation regarding how often a Venue system should go through the shutdown and startup process? Our console is installed in a permanent venue so it is easy to just turn off the displays and rest of gear at FOH while leaving the FOH rack on 24 hours a day on a UPS. Does Venue run better if it is software reset each day? Its occurred to me the console could be like any other Windows machine where it runs better with daily reboots.

Background: I had a Bombfactory BF-3A inserted on a couple of vocals, using it for the first time, and experienced a strange scenario. When we started the song in rehearsal, suddenly no audio was passing through the channel strip for the vocals with the BF-3A inserted. There was signal showing up on the signal meter, but not hitting the PQ or main busses. Unassigning the channel from any VCAs or mute groups didn’t make a difference. Going to the plug-in screen for said channel caused the screen to not redraw completely with a blank space where the plug-in should be. After trying to cycle to another screen and back, the entire software locked up. I could still control levels from the control surface, but said vocal channels would not pass audio. Scary!!!

Resolution from Digidesign engineer: While the system is designed to be able to run with no end, it is difficult for *any* system to actually *prove* that no bug is going to appear just because the system was up for too long. You’d have to perform full regression suites (which includes all configurations) with weeks of up times. It’s just not realistic. While you can have a few systems kept up & running the same configuration for a large number of days, it’s impossible to do with all configurations SW/HW (how many thousands exists?).

A classic problem for SW is counters rollover. If your system does include counters (and it pretty much guarantied to have some, no matter the OS, in your applications or drivers), you have some level of risk exposure. The textbook example of this was found in Windows 95/98 which crashed after 49.xxx days (http://news.cnet.com/Windows-may-cra…_3-222391.html). This is the type of issues that have given Windows a bad reputation, but the same condition happens in so many different ways. You also have potential memory leaks slowly saturating VM, etc. I’d like to tell you that we keep a bunch of systems running for over 50 days in each SW/HW configuration…

So let’s be practical: if your system has been running 24/7 for more than a week, you are in an increasingly less tested condition. Your system won’t burst into flames, but then, the chances that something is going to misbehave start to grow. So if a daily shutdown doesn’t work in your situation, I’d recommend a weekly restart. I pick a week because it’s the most practical period to remember. That will keep you in a reasonable uptime range, and if you forget it once in a while, it shouldn’t matter.

Additional: I also reinstalled the Bomb Factory plug-ins just to make sure there was nothing corrupt there. I’m using the BF-3A regularly since the incident and haven’t had any more strange behavior. I don’t know for certain that the problem was caused by too much runtime without a restart. Since it was the first time I used the plug-in, it is possible that something was corrupted in the initial install.

Lesson: Restart the console. Don’t leave the mix engine on 24/7. And test plug-ins before using them in a pressure environment since anything is possible in the world of software. I’ve had strange behavior from a PM1D and now I can say I’ve had a strange scenario with Digidesign Venue. Hopefully it doesn’t happen again.