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

Deletion of 2000 records on master causes complete resync on Shadow



Hello -

In our infrastructure we have one master and 2 shadow OpenLDAP 2.4.13
servers with db-4.6.21 backend database, The directory contains about
110000 records. Today we had to delete about 2000 records on the
master and that caused a complete refresh on shadows. I mean shadows
started to delete all records in their databases. In the log files it
started with appearing these messages:

Aug 27 14:59:40 berlin slapd[4138]: do_syncrep2: rid=011
LDAP_RES_INTERMEDIATE - SYNC_ID_SET
Aug 27 14:59:40 berlin slapd[4138]: do_syncrep2:
cookie=rid=011,csn=20090827205939.642722Z#000000#000#000000

Then it started deleting records:

Aug 27 14:59:45 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=,,,,,
Aug 27 14:59:46 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=.....
Aug 27 14:59:46 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=.....
Aug 27 14:59:46 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=....
Aug 27 14:59:46 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=....
Aug 27 14:59:46 berlin slapd[4138]: syncrepl_del_nonpresent: rid=011
be_delete cn=...

And deleted records that were deleted records for a whole directory
and then started putting them back. Does anybody know if this is
expected behavior?

-- 
Serge Dubrouski.