Replication :synchronizing data: error"Systemservernotfound"

[anoops81]

Hi All,

I wanted to replicate budget cube from budget server (adminhost:LAXCOGADPW002) to the reporting server (adminhost:LAXASPW007).
I was able to replicate the cube but with no data in the target server cube . But when i try to syncronize the data , then I am getting an error in processlog saying the following,

20110211043916 [Admin] => Synchronization to "production server x64" started.
20110211043916 [Admin] => ERROR: SystemServerNotFound
20110211043919 [Admin] => ERROR: Disconnected, Replicate Aborted.

Both the tm1 servers have one ADMIN host in common .

The source server cfg file is having the following adminhost entry LAXCOGADPW002;LAXASPW007

The target server cfg file is having the following entry for adminhost
LAXASPW007

The source server is LAXCOGADPW002
The target server is LAXASPW007
Both servers are TM1 9.5.1

I am attaching the errorlog also,

Did i miss something ?

Any help is very much appreceated .


Regards
Anoop

[tomok]

The source server is LAXCOGADPW002. The target server is LAXASPW007

Are you saying you named the TM1 service ("ServerName=" setting) to be exactly the same as the DNS entry for the machine the service is running on? I've never seen this setup before and would not recommend it. Perhaps this is confusing the replication process.

[anoops81]

No Tomok , I was just mentioning that both are different physical servers . The names LAXASPW007 & LAXCOGADPW002 are two seperate physical windows servers. The source tm1 server instance is running on the physical server LAXCOGADPW002 and the target tm1 server instance is running on LAXASPW007.

Regards
Anoop

[MSidat]

HI anoops,

I am pretty certain you cannot have a TM1 Server hooked up to more than one Admin Server. Rep and Sync is flaky or non existant when your servers had different admin hosts. You can keep your TM1 servers installed on different physical servers but in the cfg file of both source and target TM1 Servers have the same admin host.

Hopefully this should solve your issue.

[mattgoff]

I am pretty certain you cannot have a TM1 Server hooked up to more than one Admin Server. Rep and Sync is flaky or non existant when your servers had different admin hosts.

Multiple admin servers are supported. We've been running this way for four years from versions 9.0.2 to 9.5.1. Admin servers are glorified DNS servers, identifying IP addresses and ports to clients upon request; there's no reason you can't connect to more than one and any communication is direct client-to-server (or server-to-server), not through the admin server.

As to Anoop's question, it's hard to say. My guess would be there's a port blocking problem or an unlucky transient issue (try restarting all daemons including the admin server). I would probably also contact support for their advice-- they may be able to enable additional logging that will provide more information. I've seen just about every replication error in the book, but this is a new one to me.

Just had a thought, both servers are on different ports right?

Matt

[anoops81]

Hi All ,

This is resolved . I needed to enter both the TM1 server names as ADMINHOST in the target server and source server as shown below .


cgf file for target tm1 server
ADMINHOST=LAXASPW007;LAXCOADPW002

cfgfile for source tm1 server
ADMINHOST=LAXCOGADPW002;LAXASPW007

Previously laxaspw007 was not added as admin host in the source tm1 server . I was thinking one common admin host (laxcogadpw002) will be enough for both the source and target tm1 servers.


Now when I am replicating , data is also pulled to the target server cube .

Thanks for all the replies which you have provided .

Regards
Anoop

[mattgoff]

Hi All ,

This is resolved . I needed to enter both the TM1 server names as ADMINHOST in the target server and source server as shown below .

[...]

I was thinking one common admin host (laxcogadpw002) will be enough for both the source and target tm1 servers.

I think this is a red herring. Your original thinking was correct: you only need to have one admin server in common for replication to occur correctly. We have nine replicating servers and they share only one admin server (plus each has its own local admin server for local users), so I know this to be correct. It's possible that there was a transient glitch and when you restarted things after changing your .cfg file the error cleared.

Matt