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

Help: Unstable PDMLink 9.1 M030, Unable to Login about once a week


Help: Unstable PDMLink 9.1 M030, Unable to Login about once a week

Hello All,

We have PDMLink 9.1 M030 under a cluster environment with F5 BigIP.

Ever we started using 9.1 for about 8 month now, we constant have system that is able to login, but stuck at the PTC logo page.

It does not got to the default page at all. This happens usually once a week, sometimes more sometimes less.


- Method server and backgroundhas no log at all, AP server CPU is very low

-My engineer told me that DB server have high CPU usage, and one SQL statement (attached below)consumes about 85% CPU Time when this happen, and the Status is "Waiting".

- Restart all method server include background would solve this problem. (so it shouldn't be aphelion problem)

So what is the next step I can do to solve this, because it seems to happen randomly.

UPDATE WfAssignedActivity SET adSQEN=:1, alertTime=:2, blob$context=:3, changeStateTime=:4, classnamekeycontainerReferen=:5, idA3containerReference=:6, context=:7, dSQEN=:8, deadline=:9, deadlineDuration=:10, description=:11, classnamekeydomainRef=:12, idA3domainRef=:13, endTime=:14, eventConfiguration=:15, eventSet=:16, inheritedDomain=:17, inputVariableMap=:18, instructions=:19, name=:20, outputVariableMap=:21, parentProcessRefIsNull=:22, classnamekeyparentProcessRef=:23, idA3parentProcessRef=:24, pdSQEN=:25, priority=:26, routerType=:27, securityLabels=:28, startTime=:29, state=:30, suspendTime=:31, templateIsNull=:32, classnamekeyA5=:33, idA3A5=:34, markForDeleteA2=:35, timeToStart=:36, tripCount=:37, userEventList=:38, updateStampA2=:39, modifyStampA2=:40, updateCountA2=updateCountA2+1 WHERE ((idA2A2 = :41) AND (updateCountA2 = :42))

Joe Chen


Did you insert the numbers into the SQL statement your having a problem with
to shorten it and have value substitutes?

Top of my head - you didn't say if you altered the SQL posted so I'm
guessing you did. sounds like you have a oracle tuning problem. In the
database itself or connection pooling/cache. Areas I would focus on at
first. You may not be getting past splash screen in browser because the
process is hanging without sending back the actual http redirection but
simultaneously executing a servlet. You should monitor HTTP headers with a
tool like wireshark and send them back to the list or PTC as well as monitor
both Tomcat, Apache, and log4j logs to get as much verbosity about
connections and requests.

PTC probably would want to see all of your properties.

Good luck.

Learn more about how we are using our browser poll here