Stats module issue

Moderator: crythias

Post Reply
scottbean
OTRS newbie
Posts: 14
Joined: 30 Jan 2013, 21:08
OTRS Version?: 2.4.7
Real Name: Scott Bean

Stats module issue

Post by scottbean »

I am upgrading our test box to v3.2.1 - I have everything upgraded, but found an issue when trying to add a new report. This happens with either the ADD button on the overview page or when selecting NEW on the drop down. The page either takes a long time (as in 2 minutes or so) to load, times out and doesn't load at all, or acts as if it loads but puts the person right back on the screen they started from. I have searched the forums without much success. Any ideas?

OTRS Test box is RHEL 5.9 and MySQL 5.0.95 running on separate boxes.

Thanks,

Scott
OTRS 2.4.7/RHEL 5.9/MySQL 5.0.95 Production -- OTRS 3.2.2/RHEL 5.9/MySQL 5.0.95 Dev/Test

scottbean
OTRS newbie
Posts: 14
Joined: 30 Jan 2013, 21:08
OTRS Version?: 2.4.7
Real Name: Scott Bean

Re: Stats module issue

Post by scottbean »

***UPDATE***

I have updated our box to v3.2.2 – no luck fixing the issue. I did find the following in the logs for web when I attempted to open the NEW statistics screen:

[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:26 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:30:27 2013] [error] :Apache2 IO flush: (103) Software caused connection abort at /opt/otrs//Kernel/Output/HTML/Layout.pm line 1690
:Apache2 IO flush: (103) Software caused connection abort at -e line 0

This is what I get every time I try to create a new stat - whether from the NEW button in the statistics drop-down or from the ADD button on the overview page.

I was able to get our dev box (also updated to 3.2.2) to replicate the issue of redirecting back to the page I was on when I clicked the link. This is the only time I have been able to do this. Every other time I try it works normally. The error log information from that time is below:

[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
[Tue Feb 26 16:48:41 2013] -e: Use of uninitialized value in sprintf at /opt/otrs//Kernel/System/Stats/Dynamic/ITSMChangeManagementRfcRequester.pm line 90.
:Apache2 IO flush: (103) Software caused connection abort at -e line 0

Any assistance would be greatly appreciated.
OTRS 2.4.7/RHEL 5.9/MySQL 5.0.95 Production -- OTRS 3.2.2/RHEL 5.9/MySQL 5.0.95 Dev/Test

CSL
OTRS wizard
Posts: 159
Joined: 11 Nov 2011, 19:27
OTRS Version?: 3.0.11

Re: Stats module issue

Post by CSL »

This is how I'd go about troubleshooting this. Create new, blank OTRS system running at the same version as the system that has the issue (3.2.2). Enable the MySQL general log. Enable debug level logging in Apache.

First, empty both the MySQL general log and the Apache error_log using (cat /dev/null > (file) ). Then log in to the blank system and create a new stat with a standard set of simple options (that you note down). Take the MySQL general log and the apache error_log and save them to your machine.

Second, log in to your problem system, running on the same version. Empty both MySQL general and apache error_log logfiles. Go through the exact same steps that you did on the test system. Save the logfiles beside the others with a different name (eg production_general.log and test_general.log).

Now, run a compare on the two sets of log files (I like tkdiff ). Go through and see if anything jumps out as being significantly different on the production system. You may have to look inside both databases to check that the data used in the queries captured by the genreal log is the same / consistent.

Looking at the line specified in that ITSMChangeManagementRfcRequester.pm perl module may help, but probably only as a supplement to the logs.

This might be overkill, but it's how I'd start if I had no idea where or why things were breaking. I've not seen that problem before and don't know any quick fixes for it, so I'd go down this route. Someone else may have a 'oh, just look here / do this' answer that's much faster.
Backend: OTRS 3.0.11 RedHat Enterprise Linux 6.2, Apache, MySQL with replication
Frontend: OTRS 3.0.11 RedHat Enterprise Linux 6.2 with SELinux, Apache SSL

scottbean
OTRS newbie
Posts: 14
Joined: 30 Jan 2013, 21:08
OTRS Version?: 2.4.7
Real Name: Scott Bean

Re: Stats module issue

Post by scottbean »

Since the error referred in part to ITSM, it was suggested I uninstall and then reinstall the ITSM module. I did this and now the NEW and ADD links will spin for about 8 – 11 minutes before I get a connection reset error. Before uninstalling and reinstalling the ITSM Module I could at least get into the screen for creating a new stat from time to time.

Anyone out there have any ideas?
OTRS 2.4.7/RHEL 5.9/MySQL 5.0.95 Production -- OTRS 3.2.2/RHEL 5.9/MySQL 5.0.95 Dev/Test

scottbean
OTRS newbie
Posts: 14
Joined: 30 Jan 2013, 21:08
OTRS Version?: 2.4.7
Real Name: Scott Bean

Re: Stats module issue

Post by scottbean »

I have tracked down what I believe is the issue – the Stats module works fine, it just takes too long to collect the info from the database server so the connection is lost. We have a large number of Customer IDs, and once I set the Framework::Frontend::Agent::Stats::CustomerIDAsMultiSelect to “No”, the new stats creation page comes up. Unfortunately this is not a real solution as you can no longer select customer IDs when creating stats. This is something that cannot be missing.

How do I increase the time before the application gives up on hearing from the database server? I have tried changing timeout settings in MySQL, but that does not seem to have any effect on this.
OTRS 2.4.7/RHEL 5.9/MySQL 5.0.95 Production -- OTRS 3.2.2/RHEL 5.9/MySQL 5.0.95 Dev/Test

Post Reply