Pierangelo Masarati wrote:
hyc@OpenLDAP.org wrote:
Update of /repo/OpenLDAP/pkg/ldap/servers/slapd/overlays
Modified Files: syncprov.c 1.272 -> 1.273
Log Message: More for #6020
Howard,
I'm running a set of private tests loosely derived from test050. There is one which basically consisted in simultaneously running 3 concurrency tests (slapd-tester) with 3 MMR (intensive potentially conflicting write load). It didn't use to pass until this last fix.
Now all modifications seem to apply and replicate consistently. However, in the end the entryCSN of some entries differ (occasionally, the test just passes). Does this give you some useful indication? I can share the tests (they were meant for inclusion in OpenLDAP, all in all, although they are under development). It may require a while, because they need harmonization.
Whenever you're ready. I've also been using slapd-tester but on my faster machines it usually succeeds. At the moment I'm testing on a very slow SPARC/Solaris 10 system and that's been the only way for me to uncover these problems so far.