Segmentation violation STPX_LP

[gamotte]

Hi,

I work on a SAS web application. We have deployed new programs on the production server yesterday. These programs worked fine on the other environments but make the stored process server crash in production. The STP log file indicates a segmentation violation for the STPX_LP task (see log extract below) :

2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl - ERROR:  An exception has been encountered.
2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl - Veuillez contacter le support technique et transmettez-lui les informations suivantes :
2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl -  
2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl - Le nom de la tâche SAS est [STPX_LP ]
2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl - Violation de segmentation
2016-06-01T11:04:58,959 INFO  [00035171] 32:glesaulnier-cp@fdl -  
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - Trace de l'exception :
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl -  
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - _ptrgl+0xfffffffffffffffc
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - zdincr+0xa0
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - zdincr+0x38 -- bridge stub
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yjautol+0x310
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yjautol+0x38 -- bridge stub
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - IPRA.$yhslin+0x18c4
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yhsget+0x68
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - ytscgl+0x1bc
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - ytscch+0x40
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - ytscan+0x1b50
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yjdelta+0x5c
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yjnext+0x110
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - ytokap+0xc0
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yttknx+0x2604
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yalexec+0xf9c
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yalinln+0x6a8
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - yaltask+0x100
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - vvtentr+0x194
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - _pthread_body+0xf0
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl -
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - ERROR: Generic critical error.
2016-06-01T11:04:58,960 INFO  [00035171] 32:glesaulnier-cp@fdl - =============== STP Context 15 Log End ===============

The problem seem to occur on specific programs.

We work in SAS 9.4 with Oracle 10g.

Our application is a stored process web application that runs in the SAS Information Delivery Portal.

We have to find the source of the problem as soon as possible because of strong calendar constraints.

Any help is appreciated.

p.s. The code is confidential so I cannot post the programs.

Thanks in advance

[gamotte · Accepted solution]

We have found the source of the problem. Our stored process called a macro that contained a libname instruction for accessing an Oracle database. That libname was executed each time a user runned the stored process. Preventing the libname instruction to be re-executed solved the problem.