Force Logoff Application Managers

Force Logoff Application Managers

sdt
sdt

January 4th, 2017, 6:40 pm #1

Hi,

We ran into an issue where a user that has 'App Mgr' access was retrieving a spreadsheet that took about 1 min. It just so happens an automated process kicked off. First step in the automated process is to perform a Lev0 data export. This is where the MaxL failed and triggered a failure.

We looked at the logs to determine that at the exact same time the user was retrieving the script tried to export the data.

The error is: ERROR - 1013125 - Cannot export database while user [nstevens@AD_GROUPS] is performing [SpreadsheetOperation] on the database.

Our MaxL to Disable Logins is below. User was logged in prior to this running.

drop lock on application $ApplicationName;
alter system kill request on application $ApplicationName;
alter system logout session on application $ApplicationName force;
alter database $ApplicationName.$DatabaseName unlock all objects;
alter application $ApplicationName disable connects;
alter application $ApplicationName disable commands;
alter system unload application $ApplicationName;
alter system load application $ApplicationName;


Are we missing something that would/should kill the retrieval?
Is it even possible to kill a retrieval that is 'in progress'?
Or should be just try and pause for 60 secs and try again? Some type of loop in script.

Thanks in advance for advice.
Quote
Share

sdt
sdt

January 4th, 2017, 7:10 pm #2

I did a test on the retrieval issue. Unfortunately my results contradict what we thought happened.

1. Logged in as ‘Admin’ and retrieved a big retrieval on Distr cube on Dev.
2. Immediately executed the DisableAppCon.mxl then export database. This is exact same as what was ...happening on AUM3.
3. As my retrieval was going with the hourglass, the second I started the MaxL , I got this popup ...on my Excel: Request [SpreadsheetOperation] from user [nstevens@AD_Group] was terminated by an ...admin user.
4. And my data export continued as desired.
5. Therefore I have zero idea why Ruth did not get here retrieval killed and the automated process ...continued.

Now I'm thoroughly confused. Was this most likely just a hiccup? Is it even possible to troubleshoot?

Not sure why it's such a big issue to begin with.
Quote
Share

Cameron Lackpour
Cameron Lackpour

January 4th, 2017, 8:09 pm #3

See this for an explanation:
http://www.network54.com/Forum/58296/thread/1207316230

And this to kill it:
http://www.network54.com/Forum/58296/thread/1478708731

Make sure you throw disable connects into the script as well.

Regards,

Cameron Lackpour
Quote
Share

sdt
sdt

January 6th, 2017, 2:22 pm #4

What I did is re-arrange the to disable connects and commands first then when I get to 'kill request' I run it three times with a 10 sec pause in between. Only thing I can think of is Essbase did not kill the process that was going on and export failed. Figured I would kill three times to reduce the chances of it happening again. Not sure if this is valid but going to give it a whirl.

Thanks for links to other posts!!!!

set timestamp on;
echo "First attempt to kill any processes on "$AppName"";
alter system kill request on application $AppName;
iferror 'DisableConnectionError';

echo "Pause 10 Seconds";
shell "@ping -n 10 localhost > nul";

echo "Second attempt to kill any processes on "$AppName"";
alter system kill request on application $AppName;
iferror 'DisableConnectionError';

echo "Pause another 10 Seconds";
shell "@ping -n 10 localhost > nul";

echo "Third attempt to kill any processes on "$AppName"";
alter system kill request on application $AppName;
iferror 'DisableConnectionError';
set timestamp off;


Quote
Share


Confirmation of reply: