Since there is no evidence of profile interrupt failures during the last several days, we've decreased the frequency of adnaup commands to every 3 hours, rather than every hour.  The adnaup sequence appears to remove 16s of data.

Update (Militzer): I'm not sure where Steve got the '16s' other than when we did test the script a couple days ago.   The script was not running at all because the full path wasn't included in the crontab statement.....fixed now.   Also, the kernel log does show regular spurios interrupts since we looked at the box on the 11th.  Apparently they aren't enough to cause the lock-up problem....ie the adam has been staying up on it's own.   Was this the result of reinserting the resistor pack on the viper; not sure but could be that is true.

Now the crontab will run and we'll see those '16s' outages beginning at 18:15 every 3 hours...