NSMMReplicationPlugin - Failed to connect to replication consumer

I get the following error:

[15/Nov/2005:17:01:31 -0600] - INFORMATION - NSMMReplicationPlugin - conn=-1 op=-1 msgId=-1 - Could not send consumer in95ldpe.opr.test.statefarm.org:389 the bind request

[15/Nov/2005:17:01:31 -0600] - INFORMATION - NSMMReplicationPlugin - conn=-1 op=-1 msgId=-1 - Failed to connect to replication consumer in95ldpe.opr.test.statefarm.org:389

[15/Nov/2005:17:01:31 -0600] - ERROR<8318> - Repl. Transport - conn=-1 op=-1 msgId=-1 - [S] Bind failed with response: Failed to bind to remote (900).

but only when I use the HOSTNAME. If I use IP ADDRESS, I get no errors.

5.2 sp2.

[677 byte] By [fmodak] at [2007-11-26 6:14:51]
# 1
Try:Adding in95ldpe.opr.test.statefarm.org to /etc/hostsApplying Patch4.Gary
GaryTay at 2007-7-6 13:53:10 > top of Java-index,Java Enterprise System,Java Enterprise System - General Discussion...
# 2

Thank you Gary.

We do not want to add in95ldpe to our /etc/hosts because that is a non-standard configuration at our company, and our HP support folks will not support it.

You said applying Patch 4 should resolve the problem? Do you know of any documentation for Patch 4 where it says it addresses this issue...or similar issues?

Thanks!!

fmodak at 2007-7-6 13:53:10 > top of Java-index,Java Enterprise System,Java Enterprise System - General Discussion...
# 3

If I understand it correctly, pls could someone correct me if I am wrong.

For MMR (Multi-Master Replication) setup, I believe each master acts as "LDAP client" to each other, the LDAP client would look up DNS names using the "hosts:" directive in /etc/nsswitch.conf, which is usually set at "files dns".

Unless you have name service setup for DNS lookup for "ldap", i.e. someything "hosts: files ldap" line in /etc/nsswitch.conf, and you also need ou=hosts subtree in LDAP DIT.

I am not sure if it will work or if it will go into endless loop of DNS lookup, however it seems to me that SUN says in their docs DS5.2 cannot be LDAP name service client to itself, therefore it won't work and you need an entry in /etc/hosts.

I am not sure and did not say Patch_4 could solve your issue and I don't have HP-UX env.

SUN ONE DS5.2 for HP-UX Patch_4, i.e Patch 117669-03 could be obtained from:

<a href="http://sunsolve.sun.com," target="_blank">http://sunsolve.sun.com,</a> click Patch Finder, enter 117669.

Pls read its patch release notes.

Gary

GaryTay at 2007-7-6 13:53:10 > top of Java-index,Java Enterprise System,Java Enterprise System - General Discussion...
# 4

Hello Gary,

Thank you so much for replying to me query. Actually, I did look at the release notes of Patch 4 didn't find anything related to this issue.

Also, I don't have MMR, just one master and several consumers. We have been able to find the problem though. Here's what happened:

The problem we were experiencing occurs when a process memory maps a region, unmaps it, and maps that same region again. This is what was occuring. But it is not always guaraneteed to happen. According to HP they had to run their test program a couple of times on the original problem to recreate things. The reason is that there is no guarantee that the system will give a process the same exact memory region on the subquent calls. If it did, the SunOne product wouldn't replicate on any HP-UX server without this patch.

There's a patch for HP that we are applying.

Thanks!

fmodak at 2007-7-6 13:53:10 > top of Java-index,Java Enterprise System,Java Enterprise System - General Discussion...
# 5
Hi,We are seeing the same issue on a SUN Solaris 5.9 box with the same error messages. Can youplease confirm if any patch needs to be applied to resolve this issue?Thanks.
arun_cts at 2007-7-6 13:53:10 > top of Java-index,Java Enterprise System,Java Enterprise System - General Discussion...