[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: ITS #4198 solved?



Samuel Tran wrote:
Hi All,

I upgraded my production servers (Debian Linux Sarge) to OL 2.3.20.
I have also set up delta-syncrepl. When I first started the provider and
the consumers, I got the following error message in the provider's logs:

Mar 3 16:00:29 info-ldap-003 slapd[11762]: conn=9 op=1 SRCH
base="cn=accesslog" scope=2 deref=0 filter="(&(objectClass=auditWrite
Object)(reqResult=0))" Mar 3 16:00:29 info-ldap-003 slapd[11762]: conn=9 op=1 SRCH attr=reqDN
reqType reqMod reqNewRDN reqDeleteOldRDN reqNewSuperior en
tryCSN Mar 3 16:00:29 info-ldap-003 slapd[11762]: conn=9 op=1 SEARCH RESULT
tag=101 err=4096 nentries=0 text=sync cookie is stale Mar 3 16:00:29 info-ldap-003 slapd[11762]: conn=9 op=2 SRCH
base="dc=example,dc=com" scope=2 deref=0 filter="(objectClass=*)" Mar 3 16:00:29 info-ldap-003 slapd[11762]: conn=9 op=2 SRCH attr=* + Mar 3 16:00:34 info-ldap-003 slapd[11762]: conn=9 op=2 SEARCH RESULT
tag=101 err=0 nentries=103 text=


Apparently the empty accesslog DB caused this. After the first write
occured theses messages disappeared. It looks like the same problem
described in ITS #4198, which was supposed to be solved in 2.3.13?

If the accesslog is empty it has no choice but to fall back to regular syncrepl mode. Once a new write occurs the accesslog will not be empty any more. This is behaving as designed.


--
 -- Howard Chu
 Chief Architect, Symas Corp.  http://www.symas.com
 Director, Highland Sun        http://highlandsun.com/hyc
 OpenLDAP Core Team            http://www.openldap.org/project/