Re: index corruption (1164) still present in 2.0.15 with db 3.1.17 (ITS#1359)

[Kurt@OpenLDAP.org]

At 09:29 AM 2001-10-08, Leif Johansson wrote:
>On Fri, Oct 05, 2001 at 09:23:16AM -0700, Kurt D. Zeilenga wrote:
>> At 07:51 AM 2001-10-05, leifj@it.su.se wrote:
>> >On Wed, Oct 03, 2001 at 08:48:01AM -0700, Kurt D. Zeilenga wrote:
>> >> I haven't found anything yet in my digging...  still looking.
>> >
>> >Do you have tools to verify the consistency of the index (apart from
>> >slapindex) and report problems?
>> 
>> Unfornately not.  A simple verification (make sure all index
>> lists contained only unique ids in sorted order) wouldn't be
>> too hard to write (this is what id_check() basically does in
>> slapd).  Verification that the lists were consistent
>> with the entries would be much harder.
>> 
>
>This is getting really strange!  Just to be on the safe side wrt clients
>I had my user management scripts produce ldif which I ran through ldapmodify
>instead of letting Net::LDAPS do the modifications directly. Using this
>method I *did not* see any index corruption. Maybe I just got lucky. I 
>will try this method for a few days and see what happens. Incidentally
>ldapmodify chokes the ldif generated by Net::LDIF. I don't remember if
>
>changetype: modify
>add: foo
>foo: bar
>-
>replace: baz
>baz:foo
>replace: frotz
>frotz:foogaa
>
>is ok or if you have to do
>
>changetype: modify
>add: foo
>foo: bar
>-
>replace: baz
>baz:foo
>-
>replace: frotz
>frotz:foogaa
>
>I.e should I kick Graham or you guys :-)

Technically (RFC2849), each mod-spec ends with "-".  So
it should be:

changetype: modify
add: foo
foo: bar
-
replace: baz
baz:foo
-
replace: frotz
frotz:foogaa
-

ldapmodify(1) is liberal and doesn't require the final "-"
(for compatibility with slapd.replog(5)), but it does
require them for separation purposes.  

Kurt