Contributor "unrecoverable error" & "itk server communications timed out"

[jeffowentn]

I have a couple users (out of almost 400) who have extremely slow performance once they click on a Contributor application in Cognos Connection.  Loading the hierarchy AND opening the elist node web grid takes a really long time, and sometimes the web grid fails entirely, with the error message "itk server communications timed out." 

We are on Cognos 8.4.1 SP3 for BI and SP2 for EP, on a Windows environment.  The user is on IE9 on a Win 7 laptop. 

Multiple contributor applications yield the same result for this user.  If I log into Cognos Connection (still his login on his laptop), I get pretty close to normal response times to produce both the elist hierarchy and the web grid.  Some of the applications are server persist and some are client persist.

I also tried deleting the workbench.xml files from the user's profile, but there was no improvement there.

Any thoughts?

Jeff

[ericlfg]

Hey Jeff,

Sounds like something specific to these user's cognos connection profile information or security configurations for the applications.  A few questions:

1. Is there anything unique about these few users?  Different geography?  Etc.
2. Are you using AD security?
3. If AD security is being used, do you apply contributor rights mappings pulled directly from AD or do you use a cognos group populated with the AD group(s) or user(s)?
4. Do you have the same level of access to the same number of nodes as the affected users?
5. In the planningerrorlog, does it have anything that might be of use?

In your version of EP, when an end user is opening an application, the product goes through and recursively builds the elist workflow.  This is what the hour glass delay at the beginning is attributed to.  As for the delay / failure of the grid opening, I'm not sure what might be causing this right off.  Try clearing all the cache files for planning and IE.  You can go through IE to clear the temp files and cookies; do a search on the system for *.at and *.store and remove any files you find.

Cheers

[jeffowentn]

Eric,

Thanks for the feedback...I just posted this to another user's thread due to what I saw in the PlanningErrorLog.csv from one particular user:


I have this error for one user, but I have no resolution.  I also get "Server Error: ServerError: ITK_RPC_TIMEOUT [recoverable==false]" in the PlanningErrorLog.csv and this message as a prompt "ITK Server Communications Timed Out."  If anyone has any solutions, I would greatly appreciate it. 

We are on Cognos 8.4.1 SP2 for Planning and SP3 for BI.  This particular user is on Win 7, and we are using SQL Server 2005/8 in our environment.


In answer to your other questions, we use Access Manager to manage security, but we leverage OS authentication for about 95% of our Contributor web users.

These users are in our corporate office, and the only difference (although I've tested for this), for one user that I'm spending the most time with, is that he is using WiFi in the building instead of a LAN connection.  Interestingly, while logged onto the laptop as the user, I can login to Cognos Connection with my basic signon and see drastically improved performance.

My access is significantly greater than the user's access, although I can mirror the user's access with a test account on his machine to see what impact that has.

For whatever it is worth, here is essentially what's in the PlanningErrorLog.csv:

Error Number   Error Description
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   [C8ITK:-15] SOAP proxy timed out
-2147467259   Process exception
0   Server Error Rxd: src=Server.cpp:RequestResponse:60, errNum=0, params=, desc=[C8ITK:-15] SOAP proxy timed out
0   Server Error: ServerError: ITK_RPC_TIMEOUT [recoverable==false]

I'll try clearing the cache for IE and Planning, although this is a brand new user who has only had problems over the last week or so we've been trying to get him working properly.

Thank you, again, for your assistance.

Jeff

[jeffowentn]

I followed your steps, which did not seem to resolve the issue.  I did try accessing the same content using a test account with exactly the same permissions as the user, and the performance was normal and NOT similar to the user's experience.

So, I simply blew away his profile in Access Manager and recreated it.  Problem solved.  I have no clue what "corrupted" his profile in Access Manager/Cognos Planning, but this appears to be the fix.  Instead of it taking roughly 1 min and roughly 3 min to show the hierarchy and then open a single elist budget, it now takes roughly 5 seconds and 15-20 seconds, respectively.  This is in line with our normal expectations.

Thank you for your help, and have a great weekend!

Jeff

[jeffowentn]

Ok, so not sure that rebuilding the profile in Access Manager solved the problem.  The user continued to have issues after I made that change.  So, here are the other things I did, in case any of them help:

1. Delete IE cache (user on IE 9)
2. Delete Planning cache (.at and .store files)
3. Disable Symantec Intrusion Prevention add-on in IE (not sure this is the issue as it has been enabled on my IE8 for quite some time with no issues)

I also learned that he user was opening a multi-elist item view in Contributor and explained model size to him and what his options might need to be if that causes a problem.  We are good for now, so we'll see...

Hope this helps someone.

Jeff

[jeffowentn]

Hmmm...still no resolution.  Any other suggestions?