I haven’t yet rebooted with 2.3.1 installed, but under previous 2.x releases I have noticed reboots getting counted twice on my Windows 7 Ultimate x64 system:
It’s especially weird that they’re counted with slightly different timestamps.
Has anyone else noted this behavior? Is it something that was likely fixed in 2.3 or 2.3.1? I’ll keep an eye out after my next reboot (Windows is bothering me to install updates and restart anyway).
Is this on a virtualised machine, or just a plain Windows machine? Straight to the point; do you notice your Windows have trouble keeping the correct time?
It’s a physical Windows machine. I’ve never noticed any clock issues, but I suppose that doesn’t mean they don’t exist.
Is there a way I can manually delete the duplicates, now or once the cause is found?
Only if you manually edit the sqlite database and delete them from the ‘uptimes’ table…If you’re not familiar with sqlite, that might be a daunting task.
I’ll add something that detects clock skews, instead of trusting the clock.
I’ve done a database hack or two. Worst case I restore from the backup file I make before even opening it.
What I should do when I get a chance is check the Windows Event Log to see if the reboots are displayed once or twice in there. If Windows can’t keep its own records straight, I probably shouldn’t be expecting WhatPulse to do so…
I just had occasion to check the reboot log on my Mac, as well. I’m seeing the same reboot counted up to 25+ times there. The latest reboot (just today) with 2.3.1 seems to have been counted only once, though, so maybe I’ll just have to edit the sqlite DB to fix that.
And now that I look at my Web stats to see if there’s a field that will go out of sync if I edit the DB…I’m curious why reboot stats aren’t visible online.