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

[leifj@it.su.se]

On Sat, Sep 29, 2001 at 02:27:26PM -0700, Kurt D. Zeilenga wrote:
> At 02:14 PM 2001-09-29, leifj@it.su.se wrote:
> >On Sat, Sep 29, 2001 at 10:51:10AM -0700, Kurt D. Zeilenga wrote:
> >> At 10:39 AM 2001-09-29, leifj@it.su.se wrote:
> >> >The good news is that there was no index curruption with gdbm!
> >> 
> >> Yes, the problems appear to be limited to use of Berkeley
> >> DB's Concurrent DB API. 
> >> 
> >> >I will try a db-3.3.11 version of ldbm now and report back.
> >> 
> >> If you have problems, try undefining HAVE_BERKELEY_DB_THREAD
> >> (portable.h).  This will disable use of the Concurrent DB API
> >> and should resolve the problem (at the expense of concurrency).
> >
> >Hmm... configure left HAVE_BERKELEY_DB_THREAD undefined so I guess
> >I can't test what would have happened. In any case my 70 ops left
> >the index in great shape!
> 
> I am curious as to why HAVE_BERKELEY_DB_THREAD is undefined
> and, if it were defined, whether the problem exists.

Will try to see if I can break the index tomorrow :-) As for why
configure undefines HAVE_BERKELEY_DB_THREAD the reason is the test
for Berkeley DB thread support. The test code might simply have 
failed because the flag DB_THREAD failed. All other operations are
normal open/close/appinit and I can't see how that could have failed
The other flags used were DB_CREATE | DB_INIT_CDB | DB_INIT_MPOOL |
DB_PRIVATE | DB_MPOOL_PRIVATE | DB_THREAD.

	MVH leifj