Upgrade to 7129 problem we encountered

[hakan_carlberg]

For your info(if you want it):

We recently(2 days ago) upgraded to 7.1.2.9 and encountered a problem when storing reports in OD.

All of our reports are stored in OAM.

Last night a lot of our jobs abended, not all but a lot, with message:
"
ARS1107E RAPPFP01 AN UNEXPECTED ERROR OCCURRED.  CONTACT YOUR SYSTEM 
ADMINISTRATOR AND/OR CONSULT THE SYSTEM LOG.  FILE=ARSADMP.C, LINE   
=1059                                                                 
ARS1157E RAPPFP01 UNABLE TO STORE THE OBJECT >144<.  OBJECT SIZE 78543
ARS0088E RAPPFP01 APPLGROUP FAILED LOAD: NAME(T140306) LOADID() FILE(
/ARS/TMP/T140306)                                                     
"
and in ARSSOCKD
"
ARS0114E RAPPFP01 UNABLE TO OPEN FILE >/GLOBAL/ARS/ARSCACHE/ARSDBASE/0
/NMB/RES/144<.  ERRNO=133  SRVR->CSEXTFTP.CS.SHB.SE 192.168.237.205<-
"

errno 133 =  /* No space left on device    */


We're not using cache !!! never specified it in our production.

Finally we found the reason.
All the jobs that abended, was in the process of 'adding' a new resource, as noted in sysprint:
/ars/tmp/T140306.res will be added as resource >144-5034-0<.


looking in the ARSAG-table we noted that all of our reports had 'TYPE = 17',
APPLGRP_TYPE_INTERNAL              0x0001     1                               
APPLGRP_TYPE_SMS_TABLESPACE  0x0010    16

I recalled that in my test-system, when I tested reports against OAM/CACHE, I got TYPE =1553, which means:
APPLGRP_TYPE_INTERNAL              0x0001     1                               
APPLGRP_TYPE_SMS_TABLESPACE  0x0010    16
APPLGRP_TYPE_CACHE_NO_STORE  0x0200   512                       
APPLGRP_TYPE_CACHE_NO_LOOKUP 0x0400  1024

(Got that info from a PMR I opened)
In my test-system I never loaded a report which required a new resource to be added on an old report.


I then went to the Admin Gui-client, did 'update' on the Appl.group(NOT CHANGING ANYTHING), looked in my ARSAG-table
again and it had changed from TYPE=17  to TYPE=1553.
And now I could load the report.
So as a solution we changed every report from TYPE=17 to TYPE=1553(thru SPUFI), and we were able to load everything again


I just wanted to inform you about this, but I think that IBM should have informed us

[geoffwilde]

Thanks. I'll keep that in mind when we upgrade again.

[LWagner]

I am just movng to configuration phase of installation on 7.1.2.9.

That is a very important item to know about. I'll keep that in mind, and in my notes.

Thank You

[isig007]

This is an interesting topic. 

We do cache some of our reports, but most are sent to optical.  When I upgraded to 7.1.2.9 I noticed that my cache file system fills up quicker.  When I run the arsmaint process it will then free unused space.  Could you send me the pmr that you opened or a link?

jcbolton@up.com