Re:

Bad Guy wrote:

Dear all,

I am running the openldap 2.4.11 with 4 way masters  (SID=001 to 004) configured.  (my suffix is empty in slapd.conf)

The data can be synced initially.  I add records in 1 server and all the other 3 servers will have the new record added.  However, I found that after running for some time, one server will have corrupted contextCSN in SID=001.  

dn:
contextCSN:: sCttCIio0wAxNTQzMTMuMDQ1Mjk3WiMwMDAwMDAjMDAyIzAwMDAwMA==
contextCSN: 20081107061013.853051Z#000000#001#000000
contextCSN: 20081107073602.911356Z#000000#003#000000
contextCSN: 20081107061028.825773Z#000000#004#000000

The contextCSN for SID=002 in server 1 is corrupted.  So, whenever there is an update in SID=002 server, the SID=001 server will never get the update,
however, when there is update in SID=003 or SID=004 server, the records will get updated in SID=001.

We have a background cron job in each server running at 1 minutes interval to retrieve the records and set some user defined attributes if it meet some certain criteria.

What's the cause to this corruption ?  Is there any way to recover the corrupted contextCSN by command or script without rebuild the data ?

Looks similar to http://www.openldap.org/its?findid=5661. Can you post your configuration? Also, can you try re24 code from the CVS (or wait until 2.4.13 is out)?

p.

Ing. Pierangelo Masarati OpenLDAP Core Team

SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it ----------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Fax: +39 0382 476497 Email: ando@sys-net.it -----------------------------------