Re: Index corruption

[Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>]

Marc Munro <marc(at)bloodnok(dot)com> writes:
> Tom,
> we have a newer and much smaller (35M) file showing the same thing:

Thanks.  Looking into this, what I find is that *both* indexes have
duplicated entries for the same heap tuple:

idx1:

 Item 190 -- Length:   24  Offset: 3616 (0x0e20)  Flags: USED
  Block Id: 27806  linp Index: 2  Size: 24
  Has Nulls: 0  Has Varwidths: 0

  0e20: 00009e6c 02001800 01000000 0057d123  ...l.........W.#
  0e30: 1a781200 00000000                    .x......        

 Item 191 -- Length:   24  Offset: 3592 (0x0e08)  Flags: USED
  Block Id: 27806  linp Index: 2  Size: 24
  Has Nulls: 0  Has Varwidths: 0

  0e08: 00009e6c 02001800 01000000 0057d123  ...l.........W.#
  0e18: 2e781200 00000000                    .x......        

idx2:

 Item 127 -- Length:   16  Offset: 6144 (0x1800)  Flags: USED
  Block Id: 27806  linp Index: 2  Size: 16
  Has Nulls: 0  Has Varwidths: 0

  1800: 00009e6c 02001000 0057d123 00000000  ...l.....W.#....

 Item 128 -- Length:   16  Offset: 6128 (0x17f0)  Flags: USED
  Block Id: 27806  linp Index: 2  Size: 16
  Has Nulls: 0  Has Varwidths: 0

  17f0: 00009e6c 02001000 0057d123 00000000  ...l.....W.#....

However, the two entries in idx1 contain different data!!

What I speculate right at the moment is that we are not looking at index
corruption at all, but at heap corruption: somehow, the first insertion
into ctid (27806,2) got lost and the same ctid got re-used for the next
inserted row.  We fixed one bug like this before ...

			regards, tom lane