PI to SFDC : Error : "ErrText": "java.net.SocketTimeoutException: Read timed out",

[Former Member]

Dear All,

PI messages are failing with error as '""ErrText": "java.net.SocketTimeoutException: Read timed out",",'.

BPM to SFDC is the scenario - sync.

I have checked with SFDC team, looking everything is fine in SFDC.

What is the reason for this issue ? how to solve it.

if I resent the failed message its getting successful. Though it seems a temporary issue, 100+ failed and manually I am resending them.

Thanks,

Let me know if you need any other details.

reg, avinash

[engswee · Accepted solution]

Hi Avinash

This is quite a common issue and one possible reason could be due to temporary high volume of traffic causing the server to be unable to respond to certain requests within the time limit.

Where exactly are you getting this error? Can you provide some screenshots?

Also what adapter are you using to connect to SFDC?

The following is a useful document to troubleshoot timeout related issues. You need to identify where the timeout is occurring and increase the necessary config value to cater for it.

Regards

Eng Swee

[Former Member · Accepted solution]

Hello Eng Swee,

I am seeing errors at two places, 1. when udf makes a call to SFDC via SOAP adapter to get session id

2. When Receiver soap adapter perform Query or Upsert.

Below is the screen shots of error log:

I have read the pdf, it is given that time out can be set using parameter 'XMBWS', Shall I go for this in production ?

Thanks you.

reg, Avinash

[former_member186851 · Accepted solution]

Hello Avinash,

XMBWS parameter should help you.

But before doing in production would suggest to test in DEV/Quality

[engswee · Accepted solution]

Yup, I agree with Raghuraman. If possible, set the XMBWS.Timeout parameter in non production environment first.

However, sometimes it might be difficult to reproduce the scenario as it might required high volume of traffic.

[former_member186851 · Accepted solution]

Avinash how one more checkpoint.

How your calling SFDC from BPM?

If you use WS call from BPM directly using service reference that should increase the efficiency.

[engswee · Accepted solution]

Raghuraman

Can you please elaborate further on your reply? How does WS increase efficiency?

If you use WS call from BPM directly using service reference that should increase the efficiency.

[former_member186851 · Accepted solution]

Engg,

In the BPM if we configure webservice call instead of making it via PI will increase the efficiency right. That's my thought. Since already the BPM processing will take time in this case and again calling PI will increase the processing time.

Let me know your thoughts on this.

[engswee · Accepted solution]

But the timeout is occurring in the receiving application (SFDC).

[former_member186851 · Accepted solution]

Ok Engg..

got a bit confused, I thought its happening in PI because of the delayed processing.

[Former Member · Accepted solution]

We have set the parameter XMBWS to 10min in Prod. Messages are not failing now.

Thank you.

[former_member186851 · Accepted solution]

Great Avinash..

[former_member186851]

Avinash,

Any specific error in BPM monitoring in NWA?

[Former Member]

Hello Raghu,

For the failed PI message corresponding BPM instance is at suspended status.

reg, avinash