I've recently started playing around with "delta syncrepl", using OpenLDAP 2.3.21 (after having used slurpd with 2.2.x for years). After a few false starts I finally got it working fairly well (updates to slaves/consumers happen very quickly, and I'm able to populate consumers entirely via syncrepl, it's really cool).
But after a while I started seeing thousands of "sync cookie is stale" errors in the log of my provider. The provider starts rejecting ldapsync traffic and the consumer goes nuts with retries... Here's a small fragment from the log file in my provider: May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10868 SRCH base="dc=reed,dc=edu" scope=2 deref=0 filter="(objectClass=*)" May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10868 SRCH attr=* + May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10868 SEARCH RESULT tag=101 err=0 nentries=0 text= May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10869 SRCH base="cn=accesslog" scope=2 deref=0 filter="(&(objectClass=auditWriteObject)(reqResult=0))" May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10869 SRCH attr=reqDN reqType reqMod reqNewRDN reqDeleteOldRDN reqNewSuperior entryCSN May 3 16:14:46 unagi slapd-master[5653]: conn=0 op=10869 SEARCH RESULT tag=101 err=4096 nentries=0 text=sync cookie is stale This is happening shortly after having freshly loaded my main database and the accesslog database. The consumer (starting with an empty db) appears to pull in the entire database from the provider, but then this trouble ensues. I haven't found a lot of discussion about troubleshooting delta syncrepl configurations. Where do I start? What causes a cookie to get "stale"? Ben
