Windchill Configuration Assistant

[JIMVANDRAGT]

I just installed and executed WCA on a HP-UX server with 8 CPU's and 32 GB
of RAM. It recommended only 3 method servers. I would have guessed more
than 3.

I'm just wondering what others have found using WCA. Are you making use of
the recommended values?

Jim

Jim Van Dragt
Product Development Systems Analyst/Developer - Lead
Information Technology
hermanmiller.com
616 654 5285 OFFICE
616 836 8394 MOBILE
HermanMiller

[BobLehman2]

I have used it on all of our servers and found it to be better than what most people can do. Did it do 3 front servers and 1 background - that would make more sense to me.

[JIMVANDRAGT]

Yes, 3 foreground and 1 background.

Jim Van Dragt
Product Development Systems Analyst/Developer - Lead
Information Technology
hermanmiller.com
616 654 5285 OFFICE
616 836 8394 MOBILE
HermanMiller

[BobLehman2]

I would guess it gave each of them 4G to work with as well as giving tomcat about the same amount. That should support a fairly sizable installation.

[ShawnLothrop]

The Enterprise Deployment Center did a nice job with the Windchill Configuration Assistant, WCA. We've been using it for a while and have had very good results. The number of recommend method servers is calculated from a lookup table in the Configurator.properties files located in the root folder of the WCA directory.

Here is a section of the file showing the WCA will recommend 3 method servers for 8 CPUs. The WCA won't recommend 4 method servers until you hit 10 CPUs.

# Ratio of CPU to MethodServer

methodServersPerCpu.1=1
methodServersPerCpu.2=2
methodServersPerCpu.3=2
methodServersPerCpu.4=2
methodServersPerCpu.5=2
methodServersPerCpu.6=2
methodServersPerCpu.7=3
methodServersPerCpu.8=3
methodServersPerCpu.10=4


-- Shawn

[PreetiGupta]

I liked WCA very much. Though it is not supported for 9.1 M040 but it run just fine.

There is lot of confusion about setting the various Windchill caches. WCA also helps in determining cache sizes. Now at least we can be sure from Windchill side that it is tuned. Rest is the Oracle tuning.

Thanks,
Preeti

[ShawnLothrop]

The WCA is supported on Windchill 9.0 and higher. It was released in 9.1 M050, but it was field tested by PTC long before the M050 release. We've used it successfully on many releases prior 9.1 M050 including M040.

-- Shawn

[Robert-Altman]

I used the Configuration Assistant previously during a earlier test run
but now I seem to missing something that just does not seem to be
obvious to me.



When I execute even the simplest syntax of



ant -f WindchillConfigAssistant.xml -projecthelp



I get:



c:\PTC\Intralink\Windchill\utilities\wca>ant -f WindchillConfigAssistant
-projecthelp

Buildfile: build.xml does not exist!

Build failed

c:\PTC\Intralink\Windchill\utilities\wca>



Any thoughts??



Thanks in advance,

[MikeLockwood]

Need to set the Windchill shell on the WCA directory. I think this has an error in the path though.

[cid:image002.png@01CD0D99.D9E60420]

[BobLehman2]

On Thu, 2012-03-29 at 13:33 -0400, Robert Altman wrote:
> I used the Configuration Assistant previously during a earlier test
> run but now I seem to missing something that just does not seem to be
> obvious to me.

Did you cut and paste that directly from the pdf document?  If so the
"-" may not actually be a dash, but the longer version.  Replace the
dashes by hand and see what  you get.


>
> 
>
> When I execute even the simplest syntax of
>
> 
>
> ant -f WindchillConfigAssistant.xml -projecthelp
>
> 
>
> I get:
>
> 
>
> c:\PTC\Intralink\Windchill\utilities\wca>ant -f
> WindchillConfigAssistant -projecthelp
>
> Buildfile: build.xml does not exist!
>
> Build failed
>
> c:\PTC\Intralink\Windchill\utilities\wca>
>
> 
>
> Any thoughts??
>
> 
>
> Thanks in advance,
>
>   
>
> cid:012445618@15012007-1FE4  
>
>
> ----------

--
Bob Lehman

Email: -
Mobile: (503)888-1471

[Robert-Altman]

Thanks that’s exactly what happened cut and pasted it from my WordPad file.

Regards,

Robert

Arthrex, Inc.
Manager, Engineering CAD & Tech Support

[ssenthil]

As indicated by Mike, one should navigate to the directory containing WindchillConfigAssistant.xml

The documentation is available in Windchill® Administration - Configuring Your Windchill Environment Guide - under the section General Procedure for Running the Assistant.

HTH
Swamy Senthil
Swasen Inc.

[IgorVarshavsky]

Hi Robert,

You do not type xml in the command

ant -f WindchillConfigAssistant.xml -projecthelp



d:\ptc\Windchill_10.0\Windchill>ant -f utilities\wca\WindchillConfigAssistant.xml -projecthelp
Buildfile: d:\ptc\Windchill_10.0\Windchill\utilities\wca\WindchillConfigAssistant.xml

[Robert-Altman]

Curious what the groups thoughts are on the recommended settings when
running the Windchill Configuration Assistant? For the most part do you
go with the defaults? Or have you found anything that should be set
differently?



These were the software's recommendations:



ant -f WindchillConfigAssistant.xml configure -Dinteractive=true
-Dpropagate=false



[input] Enter Oracle database user for gathering Oracle parameters:
[#######]

[input] Enter Oracle database password for gathering Oracle
parameters: [#######]

[input] skipping input as property pga.aggregate.target has already
been set.

[input] skipping input as property RDBMS.memory.requirement has
already been set.

[input] Percent of physical memory to reserve for Operating System
[20]

[input] Allocate Windchill Memory region from total (8191 MB) or
free (5965 MB) physical memory ([free], total)

[input] Enter % of free RAM (5965 MB) to allocate to Windchill: [80]

[input] Enter % of Windchill Memory (4772) to allocate to
ServerManager Heap: [5]

[input] Enter number of MethodServers to configure (current=1): [2]

[input] Enter MethodServer max heap size (MB): [2266]

[input] Enter MethodServer initial heap size (MB): [2266]

[input] Enter ServerManager max heap size (MB): [238]

[input] Enter ServerManager initial heap size (MB): [238]

[input] Do you want to enable authenticated (non-SSL) remote JMX
access (y/n)? (y, [n])

[input] Do you want to enable anonymous remote JMX access (y/n)? (y,
[n])

[input] Set PolicyAcl Cache Size (max=10000): [200]

[input] Set RemoteObjectInfo Cache Size (max=10000): [1000]

[input] Set ContainerTeam Cache Size (max=2000): [100]

[input] Set WTContainer Cache Size (max=2000): [100]

[input] Set WTPrincipal Cache Size (max=30000): [5000]

[input] Estimated peak number of active users: [10]

[input] Set Session Cache Size (max=2000): [500]

[input] Set RoleAccess Cache Size (max=20000): [1343]

[input] Set Preference Cache Size (max=10000): [1013]

[input] Set AdministrativeDomain Cache Size (max=2000): [2000]

[input] Set PagingSession Cache Size (max=1000): [100]

[input] Set IBADefinition Cache Size (max=10000): [118]

[input] Maximum number of database connections each MethodServer may
open simultaneously: [15]

[input] Minimum number of database connections each MethodServer may
open simultaneously: [5]

[input] Statement Cache Size: [50]

[input] PreparedStatement rowPretch count: [20]

[input] Refresh Cache Size: [300]

[input] Maximum row count to be processed in a paged query (enter -1
for no maximum): [-1]

[input] Search queryLimit (enter -1 for no maximum): [5000]

[input] Search global wt.pom.queryLimit (enter -1 for no maximum):
[-1]

[input] Bind optimization cardinality: [-1]

[input] Use InClause use Bind optimization: [true]

[input] Database connection hard limit (includes pooled and
nonpooled): [25]

[input] Do you want to remove the xterm/cmd terminal windows (y/n)?
(y, [n])

[input] Do you want to enable automatic termination of requests to
avoid OutOfMemory (y/n)? ([y], n)

[input] Minimum percentage of heap available: [5]

[input] E-mail addresses for JMX adminstrators (comma-delimited
list): [-]

[input] Set servlet engine JSP mode: ([prod], dev)




This e-mail and any files transmitted with it are the property of Arthrex, Inc. and/or its affiliates, are confidential, and are intended solely for the use of the individual or entity to whom this e-mail is addressed. If you are not one of the named recipient(s) or otherwise have reason to believe that you have received this message in error, please notify the sender at 239-643-5553 and delete this message immediately from your computer. Any other use, retention, dissemination forwarding, printing or copying of this e-mail is strictly prohibited. Please note that any views or opinions presented in this email are solely those of the author and do not necessarily represent those of the company. Finally, while Arthrex uses virus protection, the recipient should check this email and any attachments for the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted by this email.

[jessh]

Just to be clear, many software's recommendations are depend on the
amount of data in various database tables and the server resources (CPU
and RAM) available -- and thus will be quite different from one
Windchill installation to the next.

--
Jess Holle

[sbugros]

Robert,

I accepted the defaults and all went well. About 1 year after we went live I had this problem in the Server Status window:

[cid:image001.png@01CD1C92.24906C20]

And this was the solution:

I'd like you to increase the max perm size to 256 and see if that clears things up. I would suggest shutting down Tomcat before following these steps.


1) In Tomcat/config.properties change the max perm size to 256
maxPermSize=256

2) Open a Windchill shell

3) Navigate to the Tomcat installation directory

4) Uninstall the Tomcat service:
ant -f config.xml uninstallService -DserviceName=Tomcat

5) Install the Tomcat service:
ant -f config.xml installService -DJAVA_HOME=d:\ptc\Windchill_9.1\Java -DserviceName=Tomcat

Other than that no other issue. (that I can recall anyway...)

Steve


Steven Bugros
Brookhaven National Laboratory
Collider-Accelerator Department
Building 911 B
35 Lawrence Drive
Upton, N.Y. 11973
631-344-3259
bugros@bnl.gov

[SteveVinyard]

I allow this to do the basics but then I modify the number of MS and memory allocation. It does an "OK" job OOTB though.

[cid:image002.gif@01CD1C7E.63D73E40]

Steve Vinyard
Application Engineer

[GaryMansell]

I have a couple of follow up questions on the Configuration Assistant recommendations...

When it suggests 2 Methodservers - does that mean 1 FG and 1 BG Methodserver?

When it suggests 3 Methodservers - does this mean 2xFG and 1 BG Methodserver?

Do you accept the wt.pom.querylimit recommendation of -1 (ie unlimited) as I am sure someone from PTC told me to set this at something like 75000 to prevent massive searches killing the machine?

Rgds

Gary

[jessh]

On 4/18/2012 9:40 AM, Gary Mansell wrote:
>
> I have a couple of follow up questions on the Configuration Assistant
> recommendations...
>
> Do you accept the wt.pom.querylimit recommendation of -1 (ie
> unlimited) as I am sure someone from PTC told me to set this at
> something like 75000 to prevent massive searches killing the machine?

A limiting value for wt.pom.querylimit, i.e. one other than -1, should
only be used as a short-term, temporary workaround to issues like
runaway searches.

The reason is that wt.pom.querylimit limits *all* queries. Thus you run
the risk of preventing queries within the bowels of the software which
are not based on bad user input and that /need /to be performed and left
unlimited for proper operation of the software. Such queries should, of
course, be well constructed to only pull back specific, small columns
and filter those to a reasonable level and may in cases need their own
chunk-based processing logic as well. Simply putting a blanket query
limit on them causes them to err out rather than doing their job.

Instead of a blanket query limit, reasonable configurable
functionality-specific limits need to be placed on functionalities like
search (or some other solution provided, e.g. streaming results with a
capability to interrupt). If such a limit does not exist for some piece
of functionality and it is producing runaway queries, then an SPR should
be produced and wt.pom.querylimit used only as an interim solution until
a fix is delivered.

--
Jess Holle

[GaryMansell]

Great - thanks for the info

[GaryMansell]

I have had a look at this and I can see a couple of Configuration Assistant parameters that need changing (at least on my system):

wt.pom.paging.snapshotQueryLimit - seems to be set to only 10000

The first will surely limit search results too drastically?

And the second prevents things like File Vault content verification working as the numbers of vault files in each folder is higher than 10000...

Also, when running it on my Test Server, where I previously had a FG MS with 3GB and a BG MS with a 1GB allocation, it changed this to one FG with 1.7GB and one BG with 1.7GB. My concern with this is that 1.7GB does not seem enough for a FG MS?

I appreciate this is only my Test Server but it is very representative of my Prod server and we had to increase the FG MS to 3GB to get it to stop crashing?

Any ideas on the better memory configuration?

Rgds

Gary

[SteveVinyard]

Don't trust this thing for memory allocation and ms config.

[cid:image001.gif@01CD1E05.D1E380F0]

Steve Vinyard
Application Engineer

[DanielReid]

For example, at least with the 9.1 version, even if you set MS to
greater than 4GB (we need rather large MS's here), it will still just
set to 4GB. It also defaults to trying to do a range instead of fixed
size. Fixed size has been recommended many times by many sources,
including some top folks within PTC. I bump the sizes to what I need
after WCA completes.



To answer an earlier question that I don't think got answered, when it
suggests 2 MethodServers and you also wanted a Background MethodServer,
one of those two will be BG. So, if you want 2 FG and 1BG, you would
set 3.

[DanielReid]

How about trying the new enhancement request system?

[GaryMansell]

Daniel/Steve thanks for confirming the FG/BG MS calculation; I was not
certain I had understood it correctly.



Forcing the BG MS to have the same amount of memory allocation as the FG
MS seems wrong to me - we find that our two FG MS's require at least 3-4
GB or else we get out of memory crashes, but that we can run our BG MS
with about 2GB (maybe less) with no problems.



If I am wrong about this - I am happy to hear why...



Rgds



Gary

[DanielReid]

Gary, I think you're probably correct about that but I haven't bothered
with it myself. In at least one case, our BGMS hit peak memory usage
and got into distress. I had actually considered giving it more memory
than the FGMS. This just highlights that there is no "one size fits
all" recommendation for this system.



If you looked at the number of users or even the number of objects or
the size of my file vaults, you might think I have too much memory
allocated. It would take a deeper look into the massive dependencies
(we have one hugely variable and very large product) to get an idea that
we are actually just barely covering our needs.



I'd be curious if anyone else is setting FG and BG separately. I hadn't
even looked into how to do this.

[SteveVinyard]

I will typically do as you do. 2GB for BG and 3-4 for FG depending on what I have to work with. I'm partial to giving a lot to the FG if its available and I don't have CPU cores to spare. Especially now that Tomcat lives in each MS too.

[cid:image001.gif@01CD1ED7.58CCEF80]

Steve Vinyard
Application Engineer

[ptc-303881]

Hi,

I wonder how companies handles non each qty, such as meter, kg in PDMLink
with a tight CAD link that drives part structure. I can not see a
good/easy way to manage this.

Below are some examples how it could be done:

1.
Add a bulk-part in assembly to transfer qty and meter to part structure in
PDMLink.
Add one normal part in assembly that have geometry and mass in CAD.
Filter the real part in PDMLink 10 with "Delete Occurrence".
If you use a table in ProE for balloons make a filter to get rid of
bulk-item.

2.
Add 2 parameters in normal ProE part with unit and value.
Adapt views in PDMLink and adapt export program to ERP system to show
parameters instead of each that link gives for a normal proe obejct.

Any more ideas how to handle this? The Ideal solution would be that a ProE
part has the bulk-item functionality when it comes to non each units and
driving part structure. I have filed an enhancement request for this.

Thanks!

Best regards,

Magnus Persson
CAD Specialist

Dynapac Compaction Equipment AB
Part of the Atlas Copco Group

[GaryMansell]

Jess,



I am hoping that you can give me some more help with a previous problem
that were kind enough to advise me on via the list previously...



We are having troubles with Business Admin's accidentally generating
large reports which have been crashing our Windchill System.



PTC Tech Support said to set the following values to prevent it
happening (which has worked):



wt.pom.queryLimit=100000

wt.pom.paging.snapshotQueryLimit=5000

com.ptc.windchill.search.queryLimit=500



These values are recommended on the following Configuration Advice Page
on the PTC Website:

[jessh]

The correct value of wt.pom.queryLimit to avoid the possibility of
inappropriately stopping various internal system queries is -1.

Technical support will often advise setting this to other values in
order to put a immediate stop to crashes from unbounded queries like
those you reported. That should only be a very temporary stop-gap
measure to stop the crashes while a better solution is produced, but all
too often this is treated as the final resolution.

I cannot speak to the other wt.pom.paging.snapshotQueryLimit, but I
believe com.ptc.windchill.search.queryLimit is exactly the sort of
usage-specific query-limit that should be used instead of a blanket
wt.pom.queryLimit setting. If there is not already such a setting for
reports in the product, then one should likely be added.

One other issue with the notion of query-limits in general is the notion
that the limit is simply to a specific number of result rows. If the
result rows are simple, e.g. containing only an object id, then the
appropriate limit is really quite large. If the result rows contain 70
columns of data, many of the long strings or LOBs, then the appropriate
limit is much smaller. On the other hand, a row limit is much easier
for business admin's, etc, to understand than a result "byte limit" or a
"cell limit"...