I always thought Ac_** was the corrupt DB ;-)

 

From: [email protected] [mailto:[email protected]] On Behalf Of Emmitt Dove
Sent: Thursday, August 19, 2010 4:14 PM
To: RBASE-L Mailing List
Subject: [RBASE-L] - Re: Corrupt DB

 

Karen,

 

It always impacted data in either SYS_TABLES or SYS_COLUMNS.  Those are tables.

 

Emmitt Dove

Manager, Converting Applications Development

Evergreen Packaging, Inc.

[email protected]

(203) 214-5683 m

(203) 643-8022 o

(203) 643-8086 f

[email protected]

 

From: [email protected] [mailto:[email protected]] On Behalf Of 
[email protected]
Sent: Thursday, August 19, 2010 15:59
To: RBASE-L Mailing List
Subject: [RBASE-L] - Re: Corrupt DB

 

I disagree.  His email said it's always corruption in one particular table.  
Not in other tables.  Aren't the chances something like 1 in 1,000,000 that a 
hardware/network issue happens to affect only that one table every time, and 
never any other tables?  Granted that's still a small chance of probability ... 
 But the situation that you and Bill describe (seems like the same situation) 
would affect random tables and random columns I'd think.  Or did it always 
change only one single same table/column each time and never any others?

Karen




Not necessarily.

  

Many years ago we were getting periodic corruption of table and column names at 
one location.  After much examination over a period of months I finally figured 
out that characters were always being changed by the same hex value.  For 
instance, an ‘a’ would become a space or some other character.  But the 
difference between them in the ASCII chart was always the same.  I finally 
traced it to one PC that had the non-maskable interrupt turned off.  The NMI’s 
function was to detect memory corruption on the motherboard.  Well, we had 
flaky RAM in one spot.  Since this was DOS, and since things always loaded in 
the same sequence, the particular area of memory was always holding the copy of 
file 1 in that region of memory.  When the PC went to update file 1 and the 
memory was in the wrong state, bad data was written to file 1, either in 
SYS_TABLES or SYS_COLUMNS.

  

Replaced the memory and all was well. 



Point it, it can absolutely be traced to something hardware or network related.

  

Emmitt Dove

Manager, Converting Applicat

 

Reply via email to