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

Re: Synchronizing two mirror mode clusters

Леонид Юрьев wrote:

I assume that ITS7830 (MDB_MAP_FULL with lots of modifications and
read access) is the same as ITS7904,
and would be solved by a 'dreamcatcher' feature.
Would be best to have an actual reproducible test case so we don't have to assume anything. 

2014-10-22 16:26 GMT+04:00 Meunier, Antonin <antonin.meunier@cgi.com>:

Hello,

I currently have several problems with MMR with lots of modification of data made in the Ldap :
- ITS 7830 (http://www.openldap.org/its/index.cgi/Incoming?id=7830;selectid=7830) --> MDB_MAP_FULL with lots of modifications and read access
- http://www.openldap.org/lists/openldap-technical/201408/msg00152.html --> Pb with Modify in big groups. I reproduce in MDB but not in hdb (perhaps is it a problem with the OS and the management of its memory)


There was already a reply to this, but you never followed up with any feedback.


- http://www.openldap.org/lists/openldap-technical/201409/msg00165.html --> Delay with ADD and DEL synchronisation with at the end no DEL on the replica
- http://www.openldap.org/lists/openldap-technical/201409/msg00171.html --> Full sync of an empty ldap with a full ldap - modified entry on the full ldap are not replicated on the empty (greater than snapshot)

I cannot for the moment test with 2.4.40 but i'll try it soon.


-----Message d'origine-----
De : openldap-technical [mailto:openldap-technical-bounces@openldap.org] De la part de Quanah Gibson-Mount
Envoyé : lundi 20 octobre 2014 21:34
À : Леонид Юрьев
Cc : openldap-technical@openldap.org
Objet : Re: Synchronizing two mirror mode clusters

--On Saturday, October 18, 2014 9:11 PM +0400 Леонид Юрьев <leo@yuriev.ru> wrote:


Sorry, but currently I strongly not recommended multi-master.
We have a big troubles in the same case, but most high load.

I am currently working on the two major issues:
- syncrepl seems unreliable, survives no more than 3-5 min under a
stress test (ITS#7968).
- fix of ITS#7904 does not affect syncrepl, thereby MDB_MAP_FULL may
occur in case of LMDB backed (no ITS now).

But another couple seem be forever:
- replication is significantly slowdown if are more than 3 nodes.
- replication may never be completed until all ldap-write activities
are stopped.


Test cases? ITS reports?

Also, I personally use delta-syncrepl vs straight syncrepl.

--Quanah


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