Exception OE Has Occured at 0028:C004EA3A in VxD VCACHE(01)
Ugh…. you hate to see those exception OE messages. This particular one happened on an old Windows 98 machine. (Yes, I know, but they want to keep it in service if possible.) The full details are as follows
An Exception OE has Occured at 0028:C004EA3A in VxD VCACHE(01) + 0000019A. … Was Called From 0028:C004BD2E in VxD VFAT(01) + 00009CB6
For starters I loved the detail that got recorded by my client – you feel like you can sink your teeth into that much better than “my computer crashed – it gave some strange error message – could you fix it?” Of course, when I looked at it I couldn’t reproduce the error and was told that it was maybe once per day.
If I remember correctly the OE exceptions with address 0028 seem to be hardware related. So, I see it refers to VCACHE and VFAT (Fat32 filesystem of course on win98) – so I do a few disk checks. Scandisk says everything is okay, so I tried a smart test from a linux livecd. That seemed to declare the hard drive healthy with no current/previous failures, so I moved onto a full read/write test to see if there were any badblocks on the drive. Again, all came up as clean.
So, the message came up again and this time I turned my attention to the system memory. I ran memtest and all was reported as being okay, then I ran the Windows Memory testing utility, again all came up as good and healthy. The system had 64MB of memory.
Now, I’ve seen bad memory manage to pass the memory test utilities before. With this system it sounded like the error message was more likely first thing in the day, then after a reboot it was fine, so I suspected there might be some “cold boot” issue with the memory that I wasn’t getting an opportunity to really test, so I decided the simplest thing to do was to swap the memory and just see if things happened again.
So, I had some used memory (128MB) that we had pulled from another machine there that I used in place of the 64MB installed. It’s been a few months now and from what I hear there aren’t the random startup errors like we saw previously (and of course performance has improved.)
I had told them that if that didn’t solve things we’d just need to look at a replacement machine (which is still on the horizon, but now has been delayed a bit longer.)