cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

We are happy to announce the new Windchill Customization board! Learn more.

Windows 7 and PDMLink 9.1 M050 Workflow Administrator

borourke
9-Granite

Windows 7 and PDMLink 9.1 M050 Workflow Administrator

Our NASA center is pushing Windows 7 out to the community and we have found that when you go to the Workflow Administrator to view or edit a workflow, a blank screen appears. There are no errors thrown in the Java console(we are using 1.6.0_23). We have a call in with PTC, but so far they are stumped. Just curious if anyone else has seen this. Thanks and Happy Mardi Gras to everyone!

Send lawyers, guns, and money...the sh!t has hit the fan!
24 REPLIES 24

Bill,
I have seen that. The workaround that I found was to dumb down my Windows7
display mode from True Color 32 bit down to 256 colors. It's not a great
solution but it will at allow you to work with the Workflow Admin page. And
it should also give PTC the clue they need to resolve it. It seems like the
Java 2D graphics API and Windows 7 display drivers are not happy together.

On Windows 7 to change the display mode it's : RMB on wallpaper > Screen
Resolution > List all modes

Greg Montgomery


On Wed, Feb 9, 2011 at 11:05 AM, Bill O'Rourke
<->wrote:

> Our NASA center is pushing Windows 7 out to the community and we have found
> that when you go to the Workflow Administrator to view or edit a workflow, a
> blank screen appears. There are no errors thrown in the Java console(we are
> using 1.6.0_23). We have a call in with PTC, but so far they are stumped.
> Just curious if anyone else has seen this. Thanks and Happy Mardi Gras to
> everyone!
>
> *Send lawyers, guns, and money...the sh!t has hit the fan! *
jessh
5-Regular Member
(To:borourke)

If there's a Java API vs. graphics driver issue on Windows, then the
solution is to add one or more properties to the JVM options to disable
some of Java's use of Direct* APIs. This is done via the Java Control
Panel.

The primary option to try is

-Dsun.java2d.noddraw=false

This turns off Java's use of DirectDraw and Direct3D completely. See

Hi Bill,

just try to set

-Dsun.java2d.d3d=false

As startup parameter in your Java Runtime Settings

[cid:image001.png@01CBC91B.5D8F1830]

Cheers,

Christian

Good Day All!

I wanted to ask the panel of experts (that's you) what is your order of
start up of Apache, Tomcat and Method Servers? I've always first started
Apache, then Tomcat and lastly PDMLink, but I would always get an error in
my Tomcat catalina.out (see below). For fun this morning I modified my
order to be Apache, PDMLink and then Tomcat and my error went away.

So what is your order of start up?
Option A:
- Apache
- Tomcat
- PDMLink

Option B:
- Apache
- PDMLink
- Tomcat


Error:
Info*Engine servlet: resource URI =
/pdm_data/ptc/Windchill/codebase/WEB-INF/ie.properties
2011-02-10 07:50:48,753 ERROR [main] wt.adapter.LoadBalancer -
wt.util.WTException: wt.util.WTRemoteException: Unable to invoke remote
method; nested exception is:
wt.util.WTRemoteException: Unable to locate method server; nested
exception is:
wt.util.WTRemoteException: Unable to locate method server; nested
exception is:
wt.util.WTRemoteException: Unable to get server; nested exception

Jim,



Appears you have a clustered or load balanced environment. Not everyone has
that, so your process will always be different depending on the
configuration.



Provided Windchill Directory Server is up and running or any other JNDI
connected directory server, MethodServers can run without Tomcat or Apache,
you just can't do much. From a development perspective, there are times
where I know if all I'm doing is writing code that executes on Tomcat JVM or
MethodServer JVM, I'll restart only the desired one.



You will get errors in the log files if users or yourself leave open
browsers with PDMLink in it and apache is still running because it will do
Ajax based session keep alives. I know this because if you leave pdmlink
open in a browser, but clear the HTTP Authentication, it will randomly
prompt you for a userid and password.



The order I choose is:



LDAP/ Directory Server

Apache

Tomcat (many if clustered)

MethodServer(s)



Like I said though, I have tomcat and apache running when editing a workflow
with externally reference code and only the MethodServer needs a bounce.
The answer to this question really just has to keep in mind the JVM process
you are working with and which one uses that code. Because of RMI, this is
possible also. Tomcat being embedded into MethodServer in 10.0, well maybe
Steve Dertien could speak to this more, as I have not seen it, but my guess
is you'll just restart MethodServer(s) and Apache. Would be good for Steve
to give a presentation on this improvement at PTC User.







Dave






RandyJones
19-Tanzanite
(To:borourke)

On 02/10/11 07:00, Jim VanDragt wrote:
> Good Day All!
>
> I wanted to ask the panel of experts (that's you) what is your order of start
> up of Apache, Tomcat and Method Servers? I've always first started Apache,
> then Tomcat and lastly PDMLink, but I would always get an error in my Tomcat
> catalina.out (see below). For fun this morning I modified my order to be
> Apache, PDMLink and then Tomcat and my error went away.
>
> So what is your order of start up?

1. Oracle (including the listener)
su - oracle -c "dbstart wind"
2. WindchillDS (ldap)
/opt/ptc/Windchill/WindchillDS/server/bin/start-ds --quiet
3. Apache
/opt/ptc/Windchill/Apache/bin/apachectl start
4. Tomcat
/opt/ptc/Windchill/Tomcat/bin/wttomcat_start
5. Windchill server manager and method servers
/opt/ptc/Windchill/Windchill/bin/windchill start

> Option A:
> - Apache
> - Tomcat
> - PDMLink
>
> Option B:
> - Apache
> - PDMLink
> - Tomcat
>
>
> Error:
> Info*Engine servlet: resource URI =
> /pdm_data/ptc/Windchill/codebase/WEB-INF/ie.properties
> 2011-02-10 07:50:48,753 ERROR [main] wt.adapter.LoadBalancer -
> wt.util.WTException: wt.util.WTRemoteException: Unable to invoke remote
> method; nested exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to get server; nested exception is:
> wt.util.WTRemoteException: Unable to locate server manager; nested
> exception is:
> java.rmi.ConnectException: Connection refused to host:
> pdm.hermanmiller.com; nested exception is:
> java.net.ConnectException: Connection refused (errno:239)
> Nested exception is: wt.util.WTRemoteException: Unable to invoke remote
> method; nested exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to get server; nested exception is:
> wt.util.WTRemoteException: Unable to locate server manager; nested
> exception is:
> java.rmi.ConnectException: Connection refused to host:
> pdm.hermanmiller.com; nested exception is:
> java.net.ConnectException: Connection refused (errno:239)
> 2011-02-10 07:50:48,756 ERROR [main] com.infoengine.object.ObjectServiceCache
> - load balancing failed for 'com.hermanmiller.pdm.Windchill', selecting a
> service at random.
> 2011-02-10 07:50:48,770 ERROR [main] wt.adapter.LoadBalancer -
> wt.util.WTException: wt.util.WTRemoteException: Unable to invoke remote
> method; nested exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to get server; nested exception is:
> wt.util.WTRemoteException: Unable to locate server manager; nested
> exception is:
> java.rmi.ConnectException: Connection refused to host:
> pdm.hermanmiller.com; nested exception is:
> java.net.ConnectException: Connection refused (errno:239)
> Nested exception is: wt.util.WTRemoteException: Unable to invoke remote
> method; nested exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to locate method server; nested
> exception is:
> wt.util.WTRemoteException: Unable to get server; nested exception is:
> wt.util.WTRemoteException: Unable to locate server manager; nested
> exception is:
> java.rmi.ConnectException: Connection refused to host:
> pdm.hermanmiller.com; nested exception is:
> java.net.ConnectException: Connection refused (errno:239)
> 2011-02-10 07:50:48,771 ERROR [main] com.infoengine.object.ObjectServiceCache
> - load balancing failed for 'com.hermanmiller.pdm.Windchill', selecting a
> service at random.
>
> *Jim Van Dragt*
> Product Development Systems Analyst/Developer - Lead
> Information Technology
>
> _hermanmiller.com_ <www.hermanmiller.com>
> 616 654 5285 OFFICE
> 616 836 8394 MOBILE
>
> *HermanMiller*
>
>
> -----End Original Message-----


--
------------------------------------------------------------------------
Randy Jones
Systems Administrator
Great Plains Mfg., Inc.
1525 E North St
PO Box 5060
Salina, KS USA 67401
email: -
Phone: 785-823-3276
Fax: 785-667-2695
------------------------------------------------------------------------

ArturPonurski
5-Regular Member
(To:borourke)

Option A.



Kind Regards,

Artur


jessh
5-Regular Member
(To:borourke)

Embedding Tomcat into the method servers in 10.0 means that the method
servers are the Tomcats in the picture as well. Thus to [re]start a
Tomcat, you [re]start a method server and when you [re]start a method
server you [re]start the Tomcat therein. In 10.0 you should still have
Oracle and LDAP running first, but beyond that you only have to start
Apache and Windchill. These can be started in either order, but
Windchill will eventually try to ping itself through Apache, so overall
I'd tend to start Apache then Windchill. You can certainly restart
Apache without restarting Windchill or vice versa.

Hello Bill,

What browser are you using? Google Chrome operates horribly with PDMLink. Also if you're using 64-Bit with 64-Bit IE, this will cause issues as well.

Also after Java 6 Update 19, when launching certain functions in PDMLink, you will get a Yes / No prompt. Intuitively you want to select "Yes (recommended)", however, you need to select "No"

_Rudy



jessh
5-Regular Member
(To:borourke)

On 2/10/2011 9:19 AM, Rudy Seidl wrote:
>
> Hello Bill,
>
> What browser are you using? Google Chrome operates horribly with
> PDMLink. Also if you're using 64-Bit with 64-Bit IE, this will cause
> issues as well.
>
A 64-bit OS with a 64-bit browser will work fine -- except initially
obtaining a 64-bit Java Plug-In in MSIE. The classid used by the
out-of-the-box configuration specifies the 32-bit Java Plug-In even
under a 64-bit MSIE -- in large part because there is no classid for the
64-bit Java Plug-In.

The end-user workaround is to manually install the 64-bit plug-in prior
to attempting to use it from the 64-bit MSIE browser.

The server-side solution (found in R10) is to have

wt.taglib.util.plugin.clientmods=MSIE HPUX AIX Win64
wt.taglib.util.plugin.clientmoddef.Win64.agentContains.1=Win64
wt.taglib.util.plugin.attr.type=browserapplet
wt.taglib.util.plugin.clientmod.MSIE.attr.type=applet
wt.taglib.util.plugin.clientmod.Win64.attr.type=browserapplet

in your wt.properties. This will use the <applet> rather than <object>
tag for the applet in 64-bit MSIE, which resolves the issue. The
installation of a 64-bit Java Plug-In won't be quite as seamless as that
of the 32-bit Java Plug-In (this is an intrinsic limitation of the
current Java Plug-In), but it works and is certainly better than having
to know to install the Java Plug-In manually prior to attempting to use it.
>
> Also after Java 6 Update 19, when launching certain functions in
> PDMLink, you will get a Yes / No prompt. Intuitively you want to
> select "Yes (recommended)", however, you need to select "No"
>
There are temp patches for this -- or better yet make sure you're using
R9.1 M060 (which has lots of other good fixes therein, anyway).

There is a separate issue specific to the document management applets
and Update 22 and higher as well -- though this only impacts certain
(network) configurations.

--
Jess Holle

Jesse,
I was wondering if you could provide more details on the Java 22 and up document management applets issues? I personally see such issues with Java 23. We were hoping Java 23 would be more useful due to its 'login' changes, but we are not pushing it because of issue we see with file upload, etc.

Craig Raymond
Windchill / Web Developer

BAE Systems
Intelligence & Security
IT Operations
600 Main St, Johnson City, NY 13790
jessh
5-Regular Member
(To:borourke)

If you see "getCacheDescriptor failed" messages emanating from the
upload applet *and* these messages occur when you use the Java 6 Update
22 (or higher) Plug-In and not when you use the Java 6 Update 21
Plug-In, then you're seeing an issue introduced by Java 6 Update 22.

If you see "getCacheDescriptor failed" messages emanating from the
upload applet when using Java 6 Update 21, then you have a different
issue. [The error message is, unfortunately, far too generic and can
have many wildly disparate underlying causes.]

The Update 22 issue is noted by

access restriction" are found in the Java console logs, despite what
Oracle's release notes state.

If you are impacted by this issue, you can try to examine your DNS, etc,
as suggested in Oracle's release notes. Otherwise a workaround in PTC
code is being worked under SPR #2047075.

--
Jess Holle

On 2/10...





















Thanks Jess, exactly what I was looking for, and very much appreciated.

Craig Raymond
Windchill / Web Developer

BAE Systems
Intelligence & Security
IT Operations
600 Main St, Johnson City, NY 13790

Jim,

I have always started windchill services in the order

Apache
PDMLink
Tomcat.

I have never encountered the error you listed below.

Thanks

Alexius C. Chukwuka
IT Analyst, PDP Systems
John Deere Power Systems
Product Engineering Center
*Voice: 319-292-8575
*Mobile: 319-429-5336
*Fax:319-292-6282
*E-Mail: -
CONFIDENTIALITY. This electronic mail and any files transmitted with it may contain information proprietary to Deere & Company, or one of its subsidiaries or affiliates, and are intended solely for the use of the individual or entity to whom they are addressed, shall be maintained in confidence and not disclosed to third parties without the written consent of the sender. If you are not the intended recipient or the person responsible for delivering the electronic mail to the intended recipient, be advised that you have received this electronic mail in error and that any use, dissemination, forwarding, printing, or copying of this electronic mail is strictly prohibited. If you have received this electronic mail in error, please immediately notify the sender by return mail.

I noticed that several people responded similar as below, with Apache first.
PTC's recommendation is to start Apache last, as you would not want people attempting to get into Windchill until all of the other components are already running.

Am I missing something?

Gerry

In Reply:


I have always started windchill services in the order

Apache
PDMLink
Tomcat.

Jim,

I don't think it matters for Apache, because even if you start Apache last someone can attempt to login to windchill before Apache completely starts, although Apache starts relatively quickly. The only reason why I start Windchill before Tomcat, is because I have seen cases where starting Tomcat first, results in Tomcat running on one of ports designated for the MethodServer, so when you start windchill, it freaks out and dies and spits out "Port already in use" error. I remember way back when Smiley Happy I started working with WIndchill, we had a PTC consultant who advised us to always start Apache first.

Thanks

Alexius C. Chukwuka
IT Analyst, PDP Systems
John Deere Power Systems
Product Engineering Center
*Voice: 319-292-8575
*Mobile: 319-429-5336
*Fax:319-292-6282
*E-Mail: -
CONFIDENTIALITY. This electronic mail and any files transmitted with it may contain information proprietary to Deere & Company, or one of its subsidiaries or affiliates, and are intended solely for the use of the individual or entity to whom they are addressed, shall be maintained in confidence and not disclosed to third parties without the written consent of the sender. If you are not the intended recipient or the person responsible for delivering the electronic mail to the intended recipient, be advised that you have received this electronic mail in error and that any use, dissemination, forwarding, printing, or copying of this electronic mail is strictly prohibited. If you have received this electronic mail in error, please immediately notify the sender by return mail.
MikeLockwood
22-Sapphire I
(To:borourke)

Restarting after a scheduled cold backup for example should not be too sensitive users attempting to log on and start working. But - If you ever need to restart during working hours and do not have a good way to notify all users, it's night and day different depending on when Apache is started.

Best results when users may be attempting to log on are from bringing up Windchill and Tomcat and allow them to stabililze before bringing up Apache.

No technically you are correct. Some people don't even restart apache, they just throw a different resource up there that is redirected to during downtime. Takes some config but your users know it won't be long.

For those that do restart apache and answer this question, it is highly probable its during downtime window or scheduled maintenance and so the likelyhood someone makes a request to apache is significantly low.


Sent from my Verizon Wireless BlackBerry

All running as a service:

1-oracle (rarely restart this)
2-WindchillDS
3-Apache
4-Tomcat
5-Windchill

A lot of systems will experience errors when booting up after a backup or restart if their backup script simply "stops" the Windchill service. I've found it MUCH more stable to include a "C:\ptc\Windchill\bin\windchill stop" command BEFORE stopping the Windchill service (windows service). When starting backup you should only need to start the service.

The issue around this seems to be that stopping the windows server (Windchill) doesn't kill the server manager or method servers all the time.

[cid:image001.gif@01CBCC3E.F287B610]

Steve Vinyard
Application Engineer

That's really interesting because the recommendation I got from at least
some PTC folks was to start Apache first, then Tomcat, then Windchill.
Some suggested it really doesn't matter.

I have seen that error this thread started with though but had not
experienced a problem with startup.

A similar workaround, and maybe less painful than dropping to 256 colors, is to turn off ClearType (Control Panel -> Display -> Adjust ClearType text). This may also help PTC get to the root of the problem.

I didn't have any luck resolving the issue by setting the sun.java2d.noddraw and sun.java2d.d3d properties.

- Jon Langlois

NetIDEAS, Inc.


In Reply to Greg Montgomery:

Bill,
I have seen that. The workaround that I found was to dumb down my Windows7
display mode from True Color 32 bit down to 256 colors. It's not a great
solution but it will at allow you to work with the Workflow Admin page. And
it should also give PTC the clue they need to resolve it. It seems like the
Java 2D graphics API and Windows 7 display drivers are not happy together.

On Windows 7 to change the display mode it's : RMB on wallpaper > Screen
Resolution > List all modes

Greg Montgomery


On Wed, Feb 9, 2011 at 11:05 AM, Bill O'Rourke
@<->wrote:

> Our NASA center is pushing Windows 7 out to the community and we have found
> that when you go to the Workflow Administrator to view or edit a workflow, a
> blank screen appears. There are no errors thrown in the Java console(we are
> using 1.6.0_23). We have a call in with PTC, but so far they are stumped.
> Just curious if anyone else has seen this. Thanks and Happy Mardi Gras to
> everyone!
>
> *Send lawyers, guns, and money...the sh!t has hit the fan! *
jessh
5-Regular Member
(To:borourke)

This issue certainly isn't general to all Windows 7 machines.

I just tested this on my new Windows 7 laptop with 32-bit / True Color
on and without disabling ClearType. This was with R9.1 M060, though I
don't suspect anything is different between M050 and M060 here -- I just
keep the latest MOR installed on my laptop at any given time. This was
also with Java 6 Update 24, but I don't suspect there's a difference
between Update 23 and 24 here, either -- I just keep the latest Java
Update installed as well.

It worked fine. There may be other sun.java2d properties that need to
be tinkered with, but overall this appears to be a Java Plug-In issue
related to specific drivers involved.

--
Jess Holle

Hello

This issue is related to older graphics hardware combined with a newer java version that uses 3D acceleration. On my laptop W7, it works with java 1.6_07, not with 23 and 24. I haven’t tested with other java updates.

I opened a call 7620646 and was provided with these three possible workarounds:

1- The issue that you have is related to an outdated Video Driver and it should be resolved by updating into upgrading to the latest release. => for me, the most recent driver still got the same error
2- Add the following JVM options via the Java control panel: => I believe this also helped.
-Dsun.java2d.d3d=false
-Dsun.java2d.noddraw=true
3- Use previous java Runtime Environment. => this certainly helps. 1.6_07 works.

One of the emails indicates the following:
“please notice that you can safely use the jre is 1.6.0.18 with the Windchill 9.1M040. The issue that you have is related to an outdated Video Driver and it should be resolved by updating into upgrading to the latest release. A similar issue has been also detected with the driver type: ATI Mobility Radeon HD 3400.”

Tony

We also tried all of the java commands with no luck, and I tried the ClearType change, but got the same results. It may well be our video drivers, and going back to the earlier version of Java is not feasible. Since this only affects a hndful of folks(there are only 4 of us who are authorized to use the Workflow Manager), we found a workaround by consoling in to the server, and doing our work there on the workflows. We just found it strange that the Workflow Manager display has problems, yet the Process/Routing displays from the details pages work just fine.

Send lawyers, guns, and money...the sh!t has hit the fan!

Top Tags