> On 8 May 2020, at 03:09, Graham Leggett <minf...@sharp.fm> wrote:
> 
> On 07 May 2020, at 18:51, Graham Leggett <minf...@sharp.fm> wrote:
> 
>> I have two servers, an older CentOS7 server running 
>> 389-ds-base-1.3.10.1-5.el7, and a newer CentOS8 server running 
>> 389-ds-base-1.4.1.3-7.module_el8.1.0+234+96aec258, and I want to set up 
>> multi-master-replication between them.
>> 
>> The replication agreement for CentOS7-> CentOS8 works great, replication is 
>> working fine.
>> 
>> The replication agreement for CentOS8 -> CentOS7 doesn’t work, giving the 
>> following strange error:
>> 
>> [07/May/2020:18:42:59.201795217 +0200] - ERR - slapi_ldap_bind - Could not 
>> send bind request for id [cn=Replication Manager,cn=config] authentication 
>> mechanism [SIMPLE]: error -1 (Can't contact LDAP server), system error -5987 
>> (Invalid function argument.), network error 0 (Unknown error, host 
>> “x.x.x:636”)
>> 
>> At the core of the above message is "network error 0”, otherwise known as 
>> “success”.

It could be TLS min versions / max versions setting perhaps? But I think I'd 
want to see more detailed logging to be sure ... or even a packet capture of 
the handshake to see where that's failing. 

>> 
>> Does this ring a bell with anyone?
> 
> Some googling sees me unearth this worrying thread: 
> https://pagure.io/389-ds-base/issue/47536
> 
> Quite a while back I spent an enormous amount of debugging time on an Ubuntu 
> version of 389ds that refused point blank to replicate. We eventually 
> discovered an awful bug where 389ds had been bound to two competing SSL 
> libraries, GnuTLS and NSS, and 389ds was passing NSS parameters (directory 
> paths) to GnuTLS, which was silently failing and then eating error messages. 
> We concluded Ubuntu was too broken to fix in any reasonable time and moved 
> all LDAP servers to CentOS7, which worked.
> 
> Doing an ldd /usr/sbin/ns-slapd shows that on CentOS8 389ds is linked to both 
> NSS and OpenSSL, which looks worryingly like the same bug has crept into 
> CentOS8.

In the future there will be some openssl provided crypto opts for password 
hashing/enc, but no sockets are activated, and the symbol names are seperate so 
I think this may not be the cause. 

> 
> Anyone have any details?
> 
> Regards,
> Graham
> —
> 
> _______________________________________________
> 389-users mailing list -- 389-users@lists.fedoraproject.org
> To unsubscribe send an email to 389-users-le...@lists.fedoraproject.org
> Fedora Code of Conduct: 
> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: 
> https://lists.fedoraproject.org/archives/list/389-users@lists.fedoraproject.org

—
Sincerely,

William Brown

Senior Software Engineer, 389 Directory Server
SUSE Labs
_______________________________________________
389-users mailing list -- 389-users@lists.fedoraproject.org
To unsubscribe send an email to 389-users-le...@lists.fedoraproject.org
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedoraproject.org/archives/list/389-users@lists.fedoraproject.org

Reply via email to