IS load time delay in custom port

, ,
1

Hi All,

We configured a custom port for IS and it takes more time(8 sec approximatelt) to even prompt for login .But, the it prompts immediately in case of default port.

Could you please suggest on how to debug this issue.

thanks,
Sathya

2

Hi Satya,

please provide your IS version and Fix Level for further assistance.

Regards,
Holger

3

Hi Holger,

Version 9.0.1.0
Updates IS_9.0_SP1_Core_Fix11

Thanks,
Sathya

4

Did you run a network trace for the new custom port you are accessing and most likely it should be a network related delay for reach/accessing the port in response.

Which browser are you seeing this delayed login prompt and can you try other browsers as well?

Also check the IS server logs once you are able to login the prompt.

HTH,
RMG

5

Thanks for the reply RMG.

I tried in IE, Chrome and I face the same issue in both.

This port behaves strange. Sometimes it prompts in 2 seconds and sometime it takes up to 15 seconds.

We face this issue in both the cluster servers.Also, this port is used as a load balancer port.

We checked with network team and they confirmed that their is no issue in the load balancer URL and they ask to check at server level as the login takes more seconds when invoked separately in individual servers.

Users consume our WSDL and are getting impacted by this delay.

6

Hi Sathya,

please check the following:

Try to login in each cluster member directly and measure the response time.

When being invoked via load balancer URL when it takes different amounts of time this means:
When it is quite fast, the first member answered the request.
When it is quite slow, the request was redirected to a subsequent member.

Additionally you should check the cluster configuration of the different members as well as the cluster cache (Ehcache or Terracotta?)

Regards,
Holger

7

Hi ,

I tried in each cluster member separately and both the nodes sometimes take 2 seconds and sometimes more than 10 seconds .

Should I narrow it down to one server instance now and try trouble shooting? or it is related to cluster ?

Thanks,
Sathya

8

Hi Sathya,

in this case you will definitely go for deeper analysis.

Which OS are you running on?

If it is Unix style, take at least 3 Thread Dumps with an interval of 1 minute for the IS instances:

kill -3 <pid>

Make sure that you use the pid of the IS instance and not the accompanying Tanuki Wrapper.

Is there anything remarkable in IS server.log, IS error log, IS session log or the Wrapper log?

Regards,
Holger

9

We use Windows and I don’t see any strange things in logs too.

I doubt that this port is used by some other process .Also, I have requested our team to create another custom port and try to login .

Thanks,
Sathya

10

Yes that can be worth a try to rule out and post us with the results.

HTH,
RMG

11

Have you seen any error messages in server logs/Terracotta logs? Are you using windows?

12

As per the user response yes it is on Windows and note the default port works fine no issues issue with that particular custom port and hope to hear back with the new results soon.

13

We tried with a different custom port and it prompts the login immediately.

So the issue is only with this particular port in each individual servers .

Any suggestions to narrow it down to find the root cause?

Thanks,
Sathya

14

Latest update is that,this particular port try to do a basic authentication twice even before the page gets loaded and get an access denied error . We identified this by tracing the TCP stream as “401 Authenticate: Basic”

In third attempt it prompts the login page for login.

Note: This behavior is not standard.Sometimes it authenticates immediately. But, when the above scenario happens only when there is a delay .

We checked the ports file in WmRoot and it is identical for all the ports.

Request some help to identify this issue.

Thanks,
Sathya

15

Deleting and recreating the port solved the issue .Will update if at all we are able to find the root cause for the same.

Thanks,
Sathya