Popup message "Unable to Process Request - Server Error" on 14.2 after upgrade

[Shalinee Chauhan]

Hi All,

We recently moved on CA PPM 14.2 SP11 from Clarity 12.1.3. We have MS SQL Server 2008 and Microsoft Internet Explorer 11 as Client.

 

We are noticing the attaced popup message with error message "Unable to process request - Server Error" and we notice the below error at same in app-ca.log file.

 

ERROR 2017-05-22 11:45:13,415 [http-bio-85-exec-6] clarity.ui (clarity:VAC12345:49641707__D53C74C1-50D0-438A-8F57-F206599876A7:security.caches) UI_MSG[:Mon May 22 11:42:36 PDT 2017:OnError called--Client Side Exception
java.lang.Exception: VXMLRequest failed. HTTP status code was: 0
at Unknown.Wfb (null:-1)
at Unknown.Om (null:-1)
at Unknown.Xjb (null:-1)
at Unknown.mkb (null:-1)
at Unknown.anonymous (null:-1)
at Unknown.qgb (null:-1)
at Unknown.tgb (null:-1)
at Unknown.anonymous (null:-1)
]

 

Before adding this new question on community I search this in same plateform and I found couple of previous question with same error but I notice in my instance " HTTP status code was: 0" is 0 however in other threads it is some value not 0. We are using SSO and F5 load balancer in our environment. We have total 3 app servers and app service of all three application server with port 85 on F5 load balancer.

 

Any suggestion is much appreciated.

 

Thanks,

Shalinee

 

 

[Nika Hadzhikidi]

Hi Shalinee

 

Does it happen frequently or only on the slowest pages/portlets? Does it happen immediately or a minute after you click on a link?

 

This is a message that occurs to indicate a network interruption. It only occurs in Internet Explorer as this is the way this browser is throwing the client exception messages. So if you use Chrome or Firefox you should not experience this issue.

 

We have fixed most of the reasons for this popup to occur in 15.1, and we have a user story open to address one remaining issue.

 

For now I'd recommend to look into improving the page response time, and use Firefox / Chrome if the issue is frequent. When you upgrade you should see this considerably less often if the problem is not gone away.  

 

Hope this helps -Nika

[Shalinee Chauhan]

Hi Nika,

Thanks for the quick response.

This is happening only when we are accessing our instance via SSO and F5 Load balancer using IE 11. If we are accessing the Clarity URL directly on IE 11 then we did not notice this issue very often . I tried with Chrome as well and did not see the issue at all.

We do not have option to switch on other Browser so I would appreciate if you would suggest some options to avoid this with IE only. Is there any setting which we need to take care from Clarity side? How to improve the page response time. Is there any setting in IE which we can try .

I am wondering if this issue is not related to CA Product then why it is not happening with CA 12.1.3?

 

Any suggestion?

Thanks,
Shalinee

[Nika Hadzhikidi]

Hi Shalinee

 

 

This issue was not happening as we did a redesign of CA Clarity and it is no longer using the same tools when rendering the webpage. This is why the issue did not exist in Clarity 12. It started occurring in release 13.

 

Try the following changes, they helped several customers:

 

1- Apache Reverse Proxy Change:

smproxy.conf - set KeepAliveTimeout to 30 seconds (I believe the default was 5)

 

2- On the F5, modify the One Connect Profile and set it to reuse TCP connections (Connection pooling) with your system administrators. We need F5 not to close those connections and to allow new requests to re-use the existing TCP connections

 

3. Make PPM timeout (in System Settings) smaller than the SSO timeout.

 

Please let me know how it goes.

 

Thank you -Nika

[Shalinee Chauhan]

Hi Nika,

 

Thanks

 

Our F5 team did below 2 change for point 2, 3

1) Timeout changed to 65 mins
2) enabled One Connect Profile 

 

However point 1 is not clear. Who needs to do this change? Is that Clarity Apache Tomact web server side or for SSO team . Is this something which F5 does as health check for Clarity servers in every 5 sec in our side,

 

1- Apache Reverse Proxy Change:

smproxy.conf - set KeepAliveTimeout to 30 seconds (I believe the default was 5)

 

Thanks,

Shalinee

[Nika Hadzhikidi]

HI Shalinee,

 

This change has to be done on your proxy (Apache Web server) in file smproxy.conf. I would suggest to involve your network team to help configure it as this is not a Clarity PPM file. Please let me know how it goes!

 

Thanks -Nika

[Shalinee Chauhan]

Here is the repsonse I get from my F5 team.

I don't know of any Apache Web server that we use for Clarity, so I'm not sure who I can direct you to.
◄	We have two proxies, 1) webseal 2) F5 load balancer, both do not run on Apache, nor are they web servers

 

Can you please guide us step by step so I can appraoch to network team like where this Apache Tomcat webserver is playing the role . I was assuming this Apache Tomcat server which we have on our Clarity Application server.

Please elaborate more here..

 

Thanks a lot for your help.

 

Shal.

[Nika Hadzhikidi]

Hi Shal

 

Although the name is very similar, Apache Web Server proxy is different from the Apache Tomcat used by CA Clarity PPM, it's a different tool. We do not modify anything in Apache Tomcat on the Clarity application side, we only want to modify the proxy.

 

The steps I sent you are for KeepAliveTimeout on an Apache Web Server (reverse proxy), if you're using Webseal, please ask your F5 team to look into modifying its properties similarly. We need to find the corresponding setting to KeepAliveTimeout and increase it to 30. I looked over the parameters and found this article:

 

https://www.ibm.com/support/knowledgecenter/SSPREK_9.0.1/com.ibm.isam.doc/wrp_config/task/tsk_config_webseal_timeout.html

 

It looks like the persistent-con-timeout might be what you are looking for. Can you check with your F5 team to get this modified and see if things are behaving better for you.

 

Thank you -Nika

[Shalinee Chauhan]

Hi Nika,

My F5 team has done the following changes in QA instance which is mirror image of production. 3 Application server and port 85 app service of each server is used in F5 and using SSO home grown portal too.

 

1) TCP timeout increased to 65 mins because Clarity timeout is 60 mins 
2) Enabled One Connect Profile
3) changed persistence from cookie to source address (same as prod in QA instance)

 

We are still getting the same error however the frequency is very low now as compare to previous.

 

Would you like to suggest some additional changes,

 

Thanks,

Shalinee

[Nika Hadzhikidi]

Hi Shal

 

Great to hear the changes worked and the popup is now very low in occurrence!

 

Did you already try as above increasing persistent-con-timeout to 30 and see if that helps? This is the only remaining thing I can think of.

 

Thank you -Nika

[Shalinee Chauhan]

Hi Nika,

 

Thanks for your guidance. However error still exist and that is a showstopper for us to do GO Live in production. We notice one more thing that when we are accessing the OWB and that error message is different.

 

 

However the error shows different message when it appears during navigation on CA PPM application.

 

I have also changed the Clarity system Options setting for inactive login inactivity to 0 minute but no improvement.

 

I will also ask my F5 team to change the persistent timeout to 30 second.

 

Thanks.

[Shalinee Chauhan]

When we changed the persistent timeout setting to 30 sec the frequency got increased. Earlier it was 65 mins.

 

Thanks,

Shalinee

[navzjoshi00]

VXMLRequest failed. HTTP status code was: 0

https://communities.ca.com/message/241901556 

 

NJ

[Shalinee Chauhan]

Hi All,

 

The issue has been resolved by changing the reverse proxy server "KeepAliveTimeout" value higher from IE Default value (60 sec). Reverse Proxy server default value in our environment was from 5 sec and we have modified to to 90 sec . Please refer the below link...

IBM WebSEAL receives POST requests from Internet Explorer without data. - United States 

 

Thanks.

Shalinee