I can reproduce it easily and consistently in my environment.  First off, I'm running 1.2.11.30.  I didn't have this issue until I upgraded to this version (or at least it never came up).  And I have a couple of masters and a couple of consumers.  When everything is in sync, I do the following:

1.  Init via the console from a master to a consumer on one of the suffixes.
2.  While the init is still happening (i.e. it is somewhere in the middle of it), I add an entry to the master.
3.  When the init is done, the newly added entry does exist on the consumer, but the errors show up indicating that the master is trying to send that ADD to the consumer.

Not sure if that helps?

Thanks!

-- Shilen


From: thierry bordaz <tbordaz@redhat.com>
Date: Friday, August 22, 2014 1:10 PM
To: Shilen Patel <shilen@duke.edu>
Cc: "mreynolds@redhat.com" <mreynolds@redhat.com>, "General discussion list for the 389 Directory server project." <389-users@lists.fedoraproject.org>
Subject: Re: [389-users] Replication error after initializing consumer

Hello,

I am still fighting to reproduce this issue.
Have you a reproducible test case, so that I could investigate locally ?

thanks
thierry
On 08/22/2014 06:47 PM, Shilen Patel wrote:
While the issue is happening, if I add another entry to the same suffix, it does not replicate over it seems.  The same error continues.

Thanks!

-- Shilen

From: thierry bordaz <tbordaz@redhat.com>
Date: Friday, August 22, 2014 12:02 PM
To: Shilen Patel <shilen@duke.edu>
Cc: "mreynolds@redhat.com" <mreynolds@redhat.com>, "General discussion list for the 389 Directory server project." <389-users@lists.fedoraproject.org>
Subject: Re: [389-users] Replication error after initializing consumer

Hello Shilen,

I wonder if it is not a transient issue.
If you add an other entry on the supplier, is it replicated  and the warning messages do no long occur ?

thanks
theirry
On 08/22/2014 03:59 PM, Shilen Patel wrote:
Once the init is done, the following is repeated on the consumer (when logging is turned up).  Note that the ADD was done while the init was happening and after the init is done, the entry does exist on the consumer but, from what I can tell, the supplier wants to still send it to the consumer and the consumer is rejecting it.

Also, the other issue that I'm noticing is that a couple of my consumers seem to be crashing (about once a day).  Right now, they are not getting any traffic at all other than updates from replication.  I'm debugging that at the moment and will report what I find.

[22/Aug/2014:14:48:34 +0100] - import test: Import complete.  Processed 282 entries in 145 seconds. (1.94 entries/sec)
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - multimaster_be_state_change: replica ou=test,dc=duke,dc=edu is coming online; enabling replication
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - _replica_configure_ruv: No ruv tombstone found for replica ou=test,dc=duke,dc=edu. Created a new one
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - conn=0 op=0 repl="ou=test,dc=duke,dc=edu": Released replica held by locking_purl=conn=15 id=4
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - replica_enable_replication: replica ou=test,dc=duke,dc=edu is relinquished
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - conn=15 op=287 repl="ou=test,dc=duke,dc=edu": Replica not in use
[22/Aug/2014:14:48:34 +0100] NSMMReplicationPlugin - conn=15 op=287 Relinquishing consumer connection extension
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=69 Acquired consumer connection extension
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=69 repl="ou=test,dc=duke,dc=edu": Begin incremental protocol
[22/Aug/2014:14:48:36 +0100] - csngen_adjust_time: gen state before 53f74a330001:1408715314:0:1
[22/Aug/2014:14:48:36 +0100] - _csngen_adjust_local_time: gen state before 53f74a330001:1408715314:0:1
[22/Aug/2014:14:48:36 +0100] - _csngen_adjust_local_time: gen state after 53f74a350000:1408715316:0:1
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=69 repl="ou=test,dc=duke,dc=edu": Acquired replica
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=69 repl="ou=test,dc=duke,dc=edu": StartNSDS50ReplicationRequest: response=0 rc=0
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=69 Relinquishing consumer connection extension
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - ruv_add_csn_inprogress: successfully inserted csn 53f749bb000000050000 into pending list
[22/Aug/2014:14:48:36 +0100]  £ - urp_add (uid=shilen3,ou=people,ou=test,dc=duke,dc=edu): an entry with this uniqueid already exists.
[22/Aug/2014:14:48:36 +0100] NSMMReplicationPlugin - conn=13 op=70 csn=53f749bb000000050000 process postop: canceling operation csn
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 Acquired consumer connection extension
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 repl="ou=test,dc=duke,dc=edu": Begin incremental protocol
[22/Aug/2014:14:48:37 +0100] - csngen_adjust_time: gen state before 53f74a350000:1408715316:0:1
[22/Aug/2014:14:48:37 +0100] - _csngen_adjust_local_time: gen state before 53f74a350000:1408715316:0:1
[22/Aug/2014:14:48:37 +0100] - _csngen_adjust_local_time: gen state after 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:37 +0100] - csngen_adjust_time: gen state after 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 repl="ou=test,dc=duke,dc=edu": Replica in use locking_purl=conn=13 id=69
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 replica="ou=test,dc=duke,dc=edu": Unable to acquire replica: error: replica busy locked by conn=13 id=69 for incremental update
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 repl="ou=test,dc=duke,dc=edu": StartNSDS50ReplicationRequest: response=1 rc=0
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=16 op=63 Relinquishing consumer connection extension
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 Acquired consumer connection extension
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 repl="ou=test,dc=duke,dc=edu": Begin incremental protocol
[22/Aug/2014:14:48:37 +0100] - csngen_adjust_time: gen state before 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:37 +0100] - csngen_adjust_time: gen state after 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 repl="ou=test,dc=duke,dc=edu": Replica in use locking_purl=conn=13 id=69
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 replica="ou=test,dc=duke,dc=edu": Unable to acquire replica: error: replica busy locked by conn=13 id=69 for incremental update
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 repl="ou=test,dc=duke,dc=edu": StartNSDS50ReplicationRequest: response=1 rc=0
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=18 op=4 Relinquishing consumer connection extension
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=13 op=71 Acquired consumer connection extension
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=13 op=71 repl="ou=test,dc=duke,dc=edu": Released replica held by locking_purl=conn=13 id=69
[22/Aug/2014:14:48:37 +0100] NSMMReplicationPlugin - conn=13 op=71 Relinquishing consumer connection extension
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - conn=17 op=63 Acquired consumer connection extension
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - conn=17 op=63 repl="ou=test,dc=duke,dc=edu": Begin incremental protocol
[22/Aug/2014:14:48:38 +0100] - csngen_adjust_time: gen state before 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:38 +0100] - _csngen_adjust_local_time: gen state before 53f74a360000:1408715317:0:1
[22/Aug/2014:14:48:38 +0100] - _csngen_adjust_local_time: gen state after 53f74a370000:1408715318:0:1
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - conn=17 op=63 repl="ou=test,dc=duke,dc=edu": Acquired replica
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - conn=17 op=63 repl="ou=test,dc=duke,dc=edu": StartNSDS50ReplicationRequest: response=0 rc=0
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - conn=17 op=63 Relinquishing consumer connection extension
[22/Aug/2014:14:48:38 +0100] NSMMReplicationPlugin - ruv_add_csn_inprogress: successfully inserted csn 53f749bb000000050000 into pending list
[22/Aug/2014:14:48:38 +0100]  £ - urp_add (uid=shilen3,ou=people,ou=test,dc=duke,dc=edu): an entry with this uniqueid already exists.

Thanks!

-- Shilen


From: thierry bordaz <tbordaz@redhat.com>
Date: Thursday, August 21, 2014 10:20 AM
To: "mreynolds@redhat.com" <mreynolds@redhat.com>
Cc: "General discussion list for the 389 Directory server project." <389-users@lists.fedoraproject.org>, Shilen Patel <shilen@duke.edu>
Subject: Re: [389-users] Replication error after initializing consumer

Hello,

This logs are coming from the supplier. It confirms what you said. The supplier retrieves a starting point for sending updates (53f5e4b5000000050000) then sends the next update 53f5e647000000050000 (uid=shilen3,ou=people,ou=test,dc=duke,dc=edu). For some reason the consumerhost:636 is unwilling to apply the update (is the backend readonly ?).
Have you got the error logs from consumerhost (with replication logging enabled) ?.

thanks
theirry
On 08/21/2014 02:45 PM, Shilen Patel wrote:
Here's the additional logging for the init issue:

[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - Finished total update of replica "agmt="cn=test5toSing1" (consumerhost:636)". Sent 278 entries.
[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Cancelling linger on the connection
[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Disconnected from the consumer
[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): State: start -> ready_to_acquire_replica
[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Trying secure slapi_ldap_init_ext
[21/Aug/2014:13:31:55 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): binddn = cn=xxx  passwd = xxx 
[21/Aug/2014:13:31:56 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): No linger to cancel on the connection
[21/Aug/2014:13:31:57 +0100] - _csngen_adjust_local_time: gen state before 53f5e64c0000:1408624203:0:1
[21/Aug/2014:13:31:57 +0100] - _csngen_adjust_local_time: gen state after 53f5e6be0000:1408624317:0:1
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Replica was successfully acquired.
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): State: ready_to_acquire_replica -> sending_updates
[21/Aug/2014:13:31:58 +0100] - csngen_adjust_time: gen state before 53f5e6be0001:1408624317:0:1
[21/Aug/2014:13:31:58 +0100] - _csngen_adjust_local_time: gen state before 53f5e6be0001:1408624317:0:1
[21/Aug/2014:13:31:58 +0100] - _csngen_adjust_local_time: gen state after 53f5e6bf0000:1408624318:0:1
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - changelog program - _cl5GetDBFile: found DB object 2b9c0c0 for database /var/lib/dirsrv/slapd-idms-authdir/changelogdb/e9c04c06-1c0711e4-a066d327-58be45f0_53e0f14e000000050000.db4
[21/Aug/2014:13:31:58 +0100] - _cl5PositionCursorForReplay (agmt="cn=test5toSing1" (consumerhost:636)): Consumer RUV:
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replicageneration} 53e0f14e000000050000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 5 ldap://xxx:389} 53e0f3cc000000050000 53f5e4b5000000050000 00000000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 8 ldap://xxx:389}
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 7 ldap://xxx:389} 53f3a401000000070000 53f3a401000000070000 00000000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 6 ldap://xxx:389} 53e3b64f000000060000 53f3a68e000000060000 00000000
[21/Aug/2014:13:31:58 +0100] - _cl5PositionCursorForReplay (agmt="cn=test5toSing1" (consumerhost:636)): Supplier RUV:
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replicageneration} 53e0f14e000000050000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 5 ldap://xxx:389} 53e0f3cc000000050000 53f5e647000000050000 53f5e646
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 8 ldap://xxx:389}
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 7 ldap://xxx:389} 53f3a401000000070000 53f3a401000000070000 00000000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): {replica 6 ldap://xxx:389} 53e3b64f000000060000 53f3a68e000000060000 00000000
[21/Aug/2014:13:31:58 +0100] agmt="cn=test5toSing1" (consumerhost:636) - clcache_get_buffer: found thread private buffer cache 7fac100773d0
[21/Aug/2014:13:31:58 +0100] agmt="cn=test5toSing1" (consumerhost:636) - clcache_get_buffer: _pool is 2b8b6f0 _pool->pl_busy_lists is 7fac140012e0 _pool->pl_busy_lists->bl_buffers is 7fac04126e60
[21/Aug/2014:13:31:58 +0100] agmt="cn=test5toSing1" (consumerhost:636) - session start: anchorcsn=53f5e4b5000000050000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - changelog program - agmt="cn=test5toSing1" (consumerhost:636): CSN 53f5e4b5000000050000 found, position set for replay
[21/Aug/2014:13:31:58 +0100] agmt="cn=test5toSing1" (consumerhost:636) - load=1 rec=1 csn=53f5e647000000050000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): replay_update: Sending add operation (dn="uid=shilen3,ou=people,ou=test,dc=duke,dc=edu" csn=53f5e647000000050000)
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): replay_update: Consumer successfully sent operation with csn 53f5e647000000050000
[21/Aug/2014:13:31:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Consumer failed to replay change (uniqueid b92e2101-292e11e4-a066d327-58be45f0, CSN 53f5e647000000050000): Server is unwilling to perform (53). Will retry later.
[21/Aug/2014:13:31:58 +0100] agmt="cn=test5toSing1" (consumerhost:636) - session end: state=0 load=1 sent=1 skipped=0 skipped_new_rid=0 skipped_csn_gt_cons_maxcsn=0 skipped_up_to_date=0 skipped_csn_gt_ruv=0 skipped_csn_covered=0
[21/Aug/2014:13:31:59 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Successfully released consumer
[21/Aug/2014:13:31:59 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Beginning linger on the connection
[21/Aug/2014:13:31:59 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): State: sending_updates -> start_backoff

Thanks!

-- Shilen

From: Mark Reynolds <mareynol@redhat.com>
Reply-To: "mreynolds@redhat.com" <mreynolds@redhat.com>
Date: Tuesday, August 19, 2014 3:32 PM
To: "General discussion list for the 389 Directory server project." <389-users@lists.fedoraproject.org>, Shilen Patel <shilen@duke.edu>
Subject: Re: [389-users] Replication error after initializing consumer


On 08/19/2014 03:05 PM, Shilen Patel wrote:
Hi Mark,

So with this issue, I'm not adding a password at all to the entry.  See the LDIF in (2) below.
No, I was referring to your first email, not this issue.  Sorry I should of responded separately. 
 I'll open a ticket.
For what issue?  For the initialization issue I want you to grab the logging first.  For the first issue(userpassword not being replicated), if you can reproduce it when only adding already hashed passwords, then please file a ticket with a reproducible testcase. 

I hope that clears things up.

Regards,
Mark



Thanks!

-- Shilen

From: Mark Reynolds <mareynol@redhat.com>
Reply-To: "mreynolds@redhat.com" <mreynolds@redhat.com>
Date: Tuesday, August 19, 2014 2:58 PM
To: "General discussion list for the 389 Directory server project." <389-users@lists.fedoraproject.org>, Shilen Patel <shilen@duke.edu>
Subject: Re: [389-users] Replication error after initializing consumer

Shilen,

A few things, you should not be adding a prehashed password (e.g.  {SSHA}DMK4S6PK6+rKSLNOL1Hl01mVJmgGi5jH) - but that should not break replication.  Can you confirm that only prehashed passwords are causing the issue?  If so, please files a ticket with a reproducible testcase:  https://fedorahosted.org/389/newticket

As for this replication issue, enabling replication logging might provide more information:

https://access.redhat.com/documentation/en-US/Red_Hat_Directory_Server/9.0/html/Configuration_Command_and_File_Reference/error-logs.html#error-logs-levels

Then post the logging output back on this list.  You should also undo this logging level once you're done - as this logging level can impact performance.

Thanks,
Mark


On 08/19/2014 01:08 PM, Shilen Patel wrote:
Hi,

I'm not sure if this is related to my previous email, but I'm also seeing issues when adding an entry while a suffix on a consumer is being initialized.  Again, I'm running 1.2.11.30.  Here are the details:

1.  Using the console, I started to initialize a suffix.

[19/Aug/2014:17:46:35 +0100] NSMMReplicationPlugin - Beginning total update of replica "agmt="cn=test5toSing1" (consumerhost:636)".

2.  While that was happening, I added an entry to the suffix on the master.

dn: uid=shilen3,ou=people,ou=test,dc=duke,dc=edu
changetype: add
objectclass: top
objectclass: person
objectclass: inetorgperson
objectclass: organizationalperson
cn: test
sn: test
uid: shilen3


[19/Aug/2014:17:47:04 +0100] conn=155 op=1 ADD dn="uid=shilen3,ou=people,ou=test,dc=duke,dc=edu"
[19/Aug/2014:17:47:04 +0100] conn=155 op=1 RESULT err=0 tag=105 nentries=0 etime=0 csn=53f37f89000000050000

3.  The init later finished.

[19/Aug/2014:17:48:55 +0100] NSMMReplicationPlugin - Finished total update of replica "agmt="cn=test5toSing1" (consumerhost:636)". Sent 270 entries.

At this point, the entry does exist on the consumer.  Presumably it was added as part of the init.


4.  However, the master still wants to send the ADD to the consumer.  On the consumer, I have the following repeated every few seconds:

[19/Aug/2014:17:48:54 +0100] conn=2002 op=56 ADD dn="uid=shilen3,ou=people,ou=test,dc=duke,dc=edu"
[19/Aug/2014:17:48:54 +0100] conn=2002 op=56 RESULT err=53 tag=105 nentries=0 etime=0 csn=53f37f89000000050000

And on the master, I have this:

[19/Aug/2014:17:48:58 +0100] NSMMReplicationPlugin - agmt="cn=test5toSing1" (consumerhost:636): Consumer failed to replay change (uniqueid 5d458a02-27c011e4-a066d327-58be45f0, CSN 53f37f89000000050000): Server is unwilling to perform (53). Will retry later.

A cl-dump shows the following:

changetype: add
replgen: 53e0f14e000000050000
csn: 53f37f89000000050000
nsuniqueid: 5d458a02-27c011e4-a066d327-58be45f0
parentuniqueid: bcd34401-21ba11df-80799838-3cd697e9
dn: uid=shilen3,ou=people,ou=test,dc=duke,dc=edu
change::
add: objectClass
objectClass: top
objectClass: person
objectClass: inetorgperson
objectClass: organizationalperson
-
add: cn
cn: test
-
add: sn
sn: test
-
add: uid
uid: shilen3
-
add: creatorsName
creatorsName: cn=directory manager
-
add: modifiersName
modifiersName: cn=directory manager
-
add: createTimestamp
createTimestamp: 20140819164704Z
-
add: modifyTimestamp
modifyTimestamp: 20140819164704Z
-
add: nsUniqueId
nsUniqueId: 5d458a02-27c011e4-a066d327-58be45f0
-
add: parentid
parentid: 2
-
add: entryid
entryid: 277
-
add: entrydn
entrydn: uid=shilen3,ou=people,ou=test,dc=duke,dc=edu
-


If anyone has any thoughts on this one, that would be appreciated.

Thanks!

-- Shilen


--
389 users mailing list
389-users@lists.fedoraproject.orghttps://admin.fedoraproject.org/mailman/listinfo/389-users



--
389 users mailing list
389-users@lists.fedoraproject.orghttps://admin.fedoraproject.org/mailman/listinfo/389-users



--
389 users mailing list
389-users@lists.fedoraproject.orghttps://admin.fedoraproject.org/mailman/listinfo/389-users