Re: (ITS#3616) Syncrepl and rootless installations

[rob@dsvr.net]

Hi,
I mirrored my config to yours. There main difference was the syncrep 
method being used. We were using refreshAndPersist. When starting a 
consumer with -c "csn=XXXXXXXXXXXXXX,rid=1,sid=1", slapd segfaults. When 
starting slapd with the refreshOnly config, shortly after startup slapd 
is scanning the whole tree, presumably doing a full context check.

I was also using the syncprov-checkpoint/sessionlog in the slapd.conf of 
the provider, as that's where the docs say to put it. Your example has 
it in the consumers slapd(?).

I don't want to abort it just now, but it'll probably take another half 
hour on these boxes to complete. Hence I'll write now to mention the 
segfault. AFAIK 'sid' had been dropped from 2.2 -> 2.3 however I don't 
see this change in CVS logs for the man page for slapd(8) so maybe I 
dreamt it(!).

Have you tested refreshAndPersist ?


provider:
####
database    bdb
suffix      ""
rootdn      "cn=root"
rootpw      foobar
directory   /drbd/opt23/var/openldap-data

index       objectClass eq
index       cn,sn,uid   pres,eq,sub
index       entryUUID,entryCSN  eq

overlay syncprov
#syncprov-checkpoint 10 1
#syncprov-sessionlog 10

####


consumer:
####
database    bdb
suffix      ""
rootdn      "cn=root"
rootpw      foobar
directory   /home/rob/ldap/var/openldap-data

index       objectClass eq
index       cn,sn,uid   pres,eq,sub
index       entryUUID,entryCSN  eq

syncrepl rid=1
                 provider=ldap://10.7.20.14:389
          type=refreshOnly
          interval=00:00:00:10
                # type=refreshAndPersist
                 searchbase=""
                 filter="(objectClass=*)"
                 scope=sub
                 attrs="*"
                 schemachecking=off
                 bindmethod=simple
                 binddn="cn=root"
                 credentials=foobar
                 retry=5,5,300,+

syncprov-sessionlog 100
####