Connecting to CMS for Visual Vectors 1

[Stinney]

We have Visual Vectors V11J-04 running on our CMS. I have confirmed that the process is running on the CMS System.

I installed the client software on my computer, but everytime I try to add the CMS server to connect to it, it generates an error message that it can't connect to the host.

I have tried using the IP address and DNS entry and neither work.

I can ping the CMS server and connect to it using CMS Supervisor.

 

[BIS]

Is the CMS on a 10 or a 100 MB Switch? I had this problem, solved it by moving the CMS to a 100MB switch. Just a thought.

 

[Stinney]

It's definitely on a 100MB switch.

We've also added the CMS to our DNS with reverse lookup and nothing.

I've added the CMS IP and DNS to my host file and nothing.

I installed Visual Vectors on a different computer and nothing.

Our maintenance provider is stumped as well.

 

[ukcomms]

Can you see the aom process running on the CMS server?

Type ps -ef|grep aom

If it's not running try starting it with the following command
aom start

Check again to see if the process is running.

If it's not starting you could have a problem with the uname of the CMS server.

Type uname -a

Does the name of the server have ANY capital letters?
If so, this can stop aom registering with Orbix.
Both need to be working for Visual Vectors to start.

If this is the problem I have a document that lists all the files that need to be changed to remove capital letters from the uname.
i.e. TT414707 become tt414707

I can post the details if you need them.

The problem is you will need downtime as the server will need to be rebooted.

 

[Stinney]

Server name is all small, nice to know this information though.

I've tried to ensure that the AOM server has started on numerous occasions. However, your suggestion returned something interesting....nothing. When I typed in the command as you can see below, it simply returned me to the # prompt. So I entered aom start. The system appears to start the AOM server but then when I typed in your command again.....nothing. So I decided to type aom stop, and the system says that the server is already off. So it doesn't look like the start is "holding". Any suggestions?

Here's a copy of what happened:

# ps -ef|grep aom
# aom start
registering server with Orbix daemon
turning on AOM...Sending output to nohup.out
AOM server started.
# ps -ef|grep aom
# aom stop
AOM server is already off
#

 

[BIS]

Wouldn't this be

# ps -ef |grep -i orbix

?

As ukcomms said, there is actually quite a few bugs that could cause this. Capital letters is one thing, length is another..

What version of VV are you using ?. in R12 CMS VV there is a bug, when you do SetupAas and select init_vvs, it only reads the first 6 charters of the VV name, and normally the cms is 10 charters uname. you modify a file in /cms/cc/aas/admin.dat and change the second line to 10 or the length of your server name. If you have done the setup, VV will remember the server name and even if you change the line in admin.dat it will not work. You have to reinstall VV. The admin.dat file is like this.

FT42910020:/cms/cc/aas# vi admin.dat
"admin.dat" 4 lines, 22 characters
100
6
"FT42910020"1

Please also check the logs in /cms/cc/aas/ClientLog folder

There are a ton of other possibilities though, if you would like to go through them all I can post this info (or perhaps ukcomms will..)

If you are not comfortable editing files in Unix I strongly recomend you get your business partner to do this though.

 

[Stinney]

I have v13, so I gues that wouldn't be the issue.

I was looking through some documentation from Avaya and in a Job Aid for configuring the CMS it said to put in the following command:

ps -ef|grep rbi

Said it should return:

root 349 1 0 Apr 03 ? 0:00 usr/bin/sh/opt/OrbixMT/bin/orbixsh

root 350 1 0 Apr 03 ? 0:00 /bin/sh /opt/OrbixMT/bin/nssh

root 349 1 0 Apr 03 ? 0:00 /opt/OrbixMT/bin/orbixd -c

When I used this command it returned the first line with:
root 349 1 0 Apr 03 ? 0:00 usr/bin/sh /opt/OrbixMT/bin/orbixsh

There is an extra space after the sh sub directory. Think there's an issue here?

 

[mojoputter]

Stinney,

Were a BP and one of our customers had the same problem, apparently Avaya has a fix for that issue. I would give the cms group a call.

 

[Stinney]

Good to know, wish my BP knew this. They've been working on it for weeks now.

 

[Stinney]

Okay, we finally have it fixed.

There is a file that can get corrupted that controlls the TCP/IP Sockets the VV needs. I'm trying to find out from my BP which file it is.

All they needed to do was to delete this file and allow it to "respawn" and I was able to connect.

 

[Stinney]

Here's the solution they used:

remove all files in:

/opt/OrbixMT/config/Repositories/NamesRep

 

[BIS]

Thanks for letting us know.