List:Falcon Storage Engine« Previous MessageNext Message »
From:Ann W. Harrison Date:February 18 2009 10:18pm
Subject:Re: Patch for bug#42208
View as plain text  
I wrote:
> 
> There is no need to store index entries if the index key doesn't
> change.  The index still points to the "master" version of the
> record, which may not have the key value for any particular entry -
> entries may represent older versions of the record and require
> walking the record chain to verify their values.
> 

Sorry, I had my head on backward.  You don't need to "verify"
the values, and the index key probably points to the newest
version of the record that has that particular key value.  To
get values for other fields, you do have to walk the chain
to find the right version for you.

Suppose you have two disagreeable connections bickering
over the key value for a record. One insists it should be 'A'
and the other insists it should be 'B'. They alternately
start transactions to make the record look the way they
want it.  In a Falcon index there will be two entries, 'A'
and 'B', both pointing to the same primary record.

In Maria, there would be many entries like this

    insertingTID  removingTID  value recLoc

         1            2          A     34
         2            3          B     67
         3            4          A    102
         4            5          B    136
         5            6          A    158

and so on.  But how likely is that?

Cheers,

Ann
Thread
Patch for bug#42208Lars-Erik Bjørk16 Feb
  • RE: Patch for bug#42208Vladislav Vaintroub16 Feb
    • RE: Patch for bug#42208Vladislav Vaintroub16 Feb
      • Re: Patch for bug#42208Jim Starkey16 Feb
        • RE: Patch for bug#42208Vladislav Vaintroub16 Feb
    • Re: Patch for bug#42208Jim Starkey16 Feb
      • RE: Patch for bug#42208Vladislav Vaintroub16 Feb
      • Re: Patch for bug#42208Lars-Erik Bjørk17 Feb
        • Re: Patch for bug#42208Kevin Lewis17 Feb
          • Re: Patch for bug#42208Ann W. Harrison18 Feb
            • Re: Patch for bug#42208Ann W. Harrison18 Feb
              • Re: Patch for bug#42208Ann W. Harrison18 Feb
                • Re: Patch for bug#42208Kevin Lewis18 Feb
                  • Re: Patch for bug#42208Ann W. Harrison18 Feb
                    • Re: Patch for bug#42208Kevin Lewis18 Feb
          • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
            • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
              • Re: Patch for bug#42208Kevin Lewis18 Feb
                • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
                  • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
              • Re: Patch for bug#42208Jim Starkey18 Feb
                • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
                  • Re: Patch for bug#42208Jim Starkey18 Feb
                    • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
                      • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
            • Re: Patch for bug#42208Ann W. Harrison18 Feb
              • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
                • Re: Patch for bug#42208Jim Starkey18 Feb
                  • Re: Patch for bug#42208Ann W. Harrison18 Feb
                    • Re: Patch for bug#42208Ann W. Harrison18 Feb
                • Re: Patch for bug#42208Ann W. Harrison18 Feb
                  • RE: Patch for bug#42208Vladislav Vaintroub18 Feb
                    • Re: Patch for bug#42208Ann W. Harrison18 Feb
                    • Re: Patch for bug#42208Kevin Lewis18 Feb
                      • Re: Patch for bug#42208Ann W. Harrison18 Feb
      • Re: Patch for bug#42208MARK CALLAGHAN17 Feb
  • Re: Patch for bug#42208Jim Starkey16 Feb