Windchill 503 Service Unavailable

[cannonsl]

I am having a Windchill error when I try to access the database.  The PTC logo comes up but then it times out and says "Service Unavailable".  The Windchill database is on a server and I am accessing from my PC.  We recently had a new security router installed and the IP addresses were changed.  I think this could be part of the issue but I can't find where to fix it.  If I go to the Apache server name in my browser I get the "It Works!" message.  But the log files in Windchill show the method server is not connecting.  I don't know how to fix this.  I have updated the httpd.conf file in the Apache setup folder to show the new IP but it didn't help.

Can someone please help?

[Invisigoth]

Hi,

Windchill isn't ready until you see something like this in the methodServer log:

main: INFO  : wt.method.server.startup  - MethodServer ready (in 704.283 secs).

If you don't see anything like that, work backwards through the log and there should be some reason, a java error, database problem, etc.

Examples of typical issues I have seen:

-There is an incorrect entry for the server in the server's hosts table; or something like this
-The database transaction log is full
-Some other database issue, however full trans log is pretty common.

Not sure about changing IP addresses in httpd.conf.  Not a normal thing I've had to do.  Also somewhat confused about just what/where the IP addresses changed.....? On the router, or on the server(s)?

The Methodserver log is the place to start.

Hope this helps!

[cannonsl]

Thank you.  I will try to dig deeper into the log files.

Is there a folder that holds the database transaction logs that I need to clear out?  I cleared the log files in the Windchill folder when I started this.

Yes, the IP addresses for the server changed.  So I was concerned there was a link somewhere in the system setup that called for the server to connect to that specific IP, in which case I would need to update it.

[cannonsl]

I don't know if you can help with this or not.  I did find some errors but I'm not sure how to decipher the fix from them.  See below.

Thu 7/11/19 00:27:35: main: javax.naming.CommunicationException: serveraddress.com:389 [Root exception is java.net.ConnectException: Connection timed out: connect]

..........

Thu 7/11/19 00:27:35: main: FATAL : wt.method.server.shutdown - Stopping services...
Thu 7/11/19 00:27:35: main: WARNING: The ManagerService is not initialized! This can be caused by:
Thu 7/11/19 00:27:35: main: 1) Attempting to invoke a server only method from a remote client
Thu 7/11/19 00:27:35: main: 2) Attempting to invoke a method on a service from the constructor or static initializer of another service
Thu 7/11/19 00:27:35: main: FATAL : wt.method.server.shutdown - MethodServer stopped

[Invisigoth]

That would be a problem connecting to the WindchillDS (or) LDAP service.

Do you know if the WindchillDS is in use, or an external source or server?

[cannonsl]

The WindchillDS service is started and running when I check the Services app on the desktop.  I tried to login to the LDAP after I started the Apache server but it wouldn't let me in.  It said my username/password was incorrect, but I know they were right.

When I put the server name in the browser without the /Windchill extension I should get the "It Works!" message.  I don't get that when I am on the server, but I do from my PC.  Don't understand that one.  There must be an error with the Apache server then?

[Invisigoth]

 I don't know why Apache "works" remotely but not locally.  Suspect its a name resolution problem.

You should not have to enter the new IP address of the server anywhere in Windchill, WindchillDS or Apache.  I would reverse what you did there. I routinely get VM copies of production systems and have to boot them up in my lab, where they get a different IP address than they did in their native environment. Never have to change anything in Windchill, only ever have to put in a static host entry if the system had a fixed, non-dhcp assigned address originally. [Changing the HOSTNAME of a Windchill server is a whole different thing, with a special utility to help make the change in all the places it needs changing]

I take it this was working before the IP change?  As long as the server can resolve its shortname and longname properly, should be able to take that out of the picture.  I've been able to resolve IP issues by putting an entry in the hosts table:  <IP address>     <Fully qualified longname with domain>     <shortname>

If that all checks out, and the system shows up with the correct hostname and domainname under Windows System widget, then I would stop and restart the Windchill services in this order: DirectoryServer, Apache, then Windchill.

[cannonsl]

I reset the IP addresses in the config files to be the default 0.0.0.0:80.

I stopped and started all of the services.

The server name has not changed, and neither has the domain.  I am still getting the same error.  I have tried to look through different files in the Apache folder to see if there is a name or address somewhere that hasn't been updated.  They are all set to Port 80 and the name is the same for all of them.

[Invisigoth]

There should be logs for the Directory server in E:\ptc\Windchill_11\WindchillDS\server\logs or similar.  The error log in there should have something like:

Started listening for new connections on LDAP Connection Handler 0.0.0.0 port 389

Perhaps an aside: Since you said you can't login to the DS control panel with known good credentials, could be a problem with how a particular Java version enforces security.  I ran into this and used the workaround.

[cannonsl]

The "errors" log file in the DS server logs folder has the following for today.  It says started successfully.

[12/Jul/2019:09:04:13 -0400] category=JEB severity=NOTICE msgID=8847402 msg=The database backend userRoot containing 444 entries has started
[12/Jul/2019:09:04:56 -0400] category=PROTOCOL severity=NOTICE msgID=2556180 msg=Started listening for new connections on Administration Connector 0.0.0.0 port 4444
[12/Jul/2019:09:04:56 -0400] category=PROTOCOL severity=NOTICE msgID=2556180 msg=Started listening for new connections on LDAP Connection Handler 0.0.0.0 port 389
[12/Jul/2019:09:04:56 -0400] category=CORE severity=NOTICE msgID=458887 msg=The Directory Server has started successfully
[12/Jul/2019:09:04:56 -0400] category=CORE severity=NOTICE msgID=458891 msg=The Directory Server has sent an alert notification generated by class org.opends.server.core.DirectoryServer (alert type org.opends.server.DirectoryServerStarted, alert ID 458887): The Directory Server has started successfully

[mmeadows-3]

Did you check the Windchill server's hosts file? Implementation specialists commonly hard code IP addresses in the hosts file so Windchill won't crash if/when the DNS server hiccups.

[cannonsl]

I don't think I have checked that one.  Where do I find this host file?

[Invisigoth · Accepted solution]

For Windows: C:\Windows\System32\drivers\etc

@mmeadows-3 is correct: I've been able to resolve IP issues by putting an entry in the hosts table:

<IP address> <Fully qualified longname with domain> <shortname>

[cannonsl]

Thank you!  I just updated that file and now when I start the Apache service on the server I get the "It Works!" message.  So it at least fixed that problem.

I am going to start Windchill now and see if I can access the server.

[cannonsl]

That worked!!!

Thank you so much for your help!  I really really appreciate it!

[mmeadows-3]

%windir%\system32\drivers\etc\hosts

[cannonsl]

Thank you, Thank you!

I had completely forgotten about that host file.  That did the trick!

[Invisigoth]

Great! You're welcome!

The other symptom of an incorrect/missing hosts entry is that the Database service can't start, and the Windchill startup never gets beyond the main ServerManager.