directory-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kiran Ayyagari <kayyag...@apache.org>
Subject Re: Upgrade to M12
Date Tue, 14 May 2013 09:14:31 GMT
On Tue, May 14, 2013 at 2:43 PM, Kiran Ayyagari <kayyagari@apache.org>wrote:

>
>
>
> On Tue, May 14, 2013 at 1:59 PM, Borys <sborys@gmail.com> wrote:
>
>> Perhaps this is what caused deletion. If replication consumer cannot see
>> the partition in the same way as studio cannot then deletion would
>> start....
>>
> looks like an issue with the API (LdapNetworkConnection)
>
BUT I still can't think of any situation how this can happen

>  On 14 May 2013 10:17, "Borys" <sborys@gmail.com> wrote:
>>
>> > I did not move data around. I was just replacing the
>> > lib/apacheds-service.jar file.
>> > Another interesting observation: I cannot see cn=eumetsat partition in
>> the
>> > studio although I can query it using ldapsearch.
>> >  On 14 May 2013 09:46, "Kiran Ayyagari" <kayyagari@apache.org> wrote:
>> >
>> >> On Tue, May 14, 2013 at 1:08 PM, Borys <sborys@gmail.com> wrote:
>> >>
>> >> > No I did not delete anything anywhere. The only what was done is:
>> >> > -Stop both serevrs
>> >> > - update config.ldif on both servers to add missing attribute for M12
>> >> > - replace the jar file
>> >> > - attempt to start one of the servers.
>> >> > - put back old jar file
>> >> > - deleted the attribute from the config.ldif on both servers
>> >> > - start one server server. Check the content - fine.
>> >> > - start second server. Check the content - few entries missing. Check
>> >> > again, even more entries missing.
>> >> >
>> >> I *guess* this is where the problem might have happened, while copying
>> >> files back to this second
>> >> server you might have copied old partition data  triggering a delete
>> >>
>> >> > - check the content on another server - more and more entries are
>>  gone
>> >> and
>> >> > finally, all is gone.
>> >> >
>> >> >
>> >> >
>> >> >
>> >> >
>> >> > On Tue, May 14, 2013 at 9:32 AM, Emmanuel Lécharny <
>> elecharny@gmail.com
>> >> > >wrote:
>> >> >
>> >> > > Le 5/14/13 9:16 AM, Borys a écrit :
>> >> > > > That is exactly the problem, I did add some custom schema.
>> >> > > > Now I have another big problem. I have two servers replicated.
>> The
>> >> > > > replication was setup yesterday and worked fine. Today after
>> server
>> >> > > restart
>> >> > > > the replicated content start to "shrink" . I had fewer and
fewer
>> >> > entries
>> >> > > in
>> >> > > > the database. And now the DB is empty on BOTH servers!!!
>> >> > > > The only message I got when I started second server was:
>> >> > > > [07:05:40] WARN
>> >> > > >
>> >> > >
>> >> >
>> >>
>> [org.apache.directory.server.ldap.replication.provider.SyncReplRequestHandler]
>> >> > > > - received a valid cookie
>> >> > > > rid=001,csn=20130513122431.469000Z#000000#001#000000 but
there
>> is no
>> >> > > event
>> >> > > > log associated with this replica
>> >> > >
>> >> > > My guess is that you deleted all the entries from one server before
>> >> > > injectng the new schema, then  reconnected the server to the other
>> >> one.
>> >> > > As the date of modification on the deleted entries is more recent,
>> >> those
>> >> > > deletions get 'replicated'
>> >> > >
>> >> > >
>> >> > > --
>> >> > > Regards,
>> >> > > Cordialement,
>> >> > > Emmanuel Lécharny
>> >> > > www.iktek.com
>> >> > >
>> >> > >
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Kiran Ayyagari
>> >> http://keydap.com
>> >>
>> >
>>
>
>
>
> --
> Kiran Ayyagari
> http://keydap.com
>



-- 
Kiran Ayyagari
http://keydap.com

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message