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

Re: Slurpd can't start...



Jorge Ruão disse:
>
> Thanks, I've found the copy in:
>
> /usr/local/var/openldap-slurp/replica
>
> strange why slurpd isn't maintaining the files in
> /etc/openldap/master-slapd.replog as specified in the replogfile entry
> in slapd.conf.

Not strange at all. That's intended behaviour. (again man slurpd)

>
> One last question regarding recplication: Right after the changes are
> made in the Master, slurpd tries to replicate them to the slaves, if a
> slave is down, slurpd won't replicate. After awhile I've put the slave
> up and some  minutes after that (i can't precise how many) slurpd
> has replicated the changes to the slave. How does slurpd acts when the
> slave is down? He keeps trying from minute to minute until the slave is
> up???

Yes. An intelligent behaviour, isn't it?

>
> Thanks,
> Jorge Ruão
>
> Luca Scamoni wrote:
>
>   find / -name slurpd.replog
>
> slurpd makes a copy of the replogfile before replicating.
>
> man slurpd (especially the -t option)
>
> Jorge Ruão disse:
>
>
>     Actually it works with uri=ldap://apollo.fe.up.pt, but after a
> little
> research i found that the slurpd that was starting was from an older
> OpenLDAP version, a little shortcut bug for my part sorry :(.
> I've cleaned it up and the slurpd is starting ok with replica
> uri=ldaps://apollo.fe.up.pt:636, but it brings me back to the original
> problem:
> "when I make changes to the LDAP directory entries, the
> /etc/openldap/master-slapd.replog and
> /etc/openldap/master-slapd.replog.lock files are generated but are empty
> files... with 0 length"...
> The strange is that all changes made in the master are being replicated
> to the slave by slurpd (I can see the changes being sent by the master
> with the debug option: -d 1), but the replog files are allways empty...
> Shouldn't the replog files store the master changes, even after they are
> transmitted to the slaves???
>
> Hope u can elucidate me,
> Thanks,
> Jorge Ruão
>
>   Jorge Ruão wrote:
>
>
>     Hi,
>
> I'm working with OpenLDAP 2.2.11 now in a Slackware 9.1 machine.
> I've set the replica lines in slapd.conf with an uri as suggested. It's
> done like:
> _______________________________________________________
> backend         bdb
> database        bdb
> suffix          "dc=fe,dc=up,dc=pt"
> rootdn          "cn=admin,dc=fe,dc=up,dc=pt"
> rootpw          {CRYPT}...
> readonly        off
> directory       /usr/local/var/openldap-data
>
> replica uri=ldaps://apollo.fe.up.pt:636
>
>
>
> just to debug the problem, it works if you put something
> like this:
>
> replica uri=ldap://apollo.fe.up.pt
>
> note the lack of "s" in the protocol and the port.
>
>
>
>             binddn="uid=updatedn,ou=admins,dc=fe,dc=up,dc=pt"
>         bindmethod=simple credentials=secret
> replogfile      /etc/openldap/master-slapd.replog
> _______________________________________________________
>
> But every time I start SLURPD i'm still getting the error:
>
> Error: parse_replica_line: unknown keyword
> "uri=ldaps://apollo.fe.up.pt:636" Error: Malformed "replica" line in
> slapd config file, line 78
> Warning: failed to add replica "(null):0 - ignoring replica
> Warning: unknown replica 192.168.53.83:636 found in replication log
> Warning: unknown replica apollo.fe.up.pt:636 found in replication log
> Warning: unknown replica apollo.fe.up.pt:389 found in replication log
>
> I need to use URI because i'm working with ssl through port 636, but if
> i use HOST instead of URI i'm not getting this parse error....
>
> Any help would be gratefull, thanks
> Jorge Ruão
>
>
>
>
>
>
> --
>
>
>
>
>


-- 
Luca Scamoni - e-mail: luca.scamoni@sys-net.it
SysNet snc - Via Dossi, 8 - 27100 Pavia Italy
IT Senior Consultant - mobile: +393471014425




    SysNet - via Dossi,8 27100 Pavia Tel: +390382573859 Fax: +390382476497