%system-f-ASTFLT AST fault

[Ayman Abusaffaqa]

We have 2 Alphaservers ES45 based on OpenVMS 7.3-1 cluster with 2 DS-KGPSA-DA on each,using MSA1000 w/dual controllers & dual MSA switchs 2/8.Disks are configured as Mirrored(12 disks)
the following were done so far for the said configuration:
1.necesary TIMA kita for OVMS 7.3-1.
2.firmware v4.24 for MSA1000.
3.alphaserver firmware v6.5 and
4.fc2391a1 firaware for the HBA'a

RDB is being used as a database & applications are cobol & fms.

every time a single user on each application will login,the application will work fine,when ever another(2nd,3rd & so on) login,it will terminate other users randomly,with error messege
"DCL-S-RETURNED CONTROL RETURNED
%System-f-ASTFLT AST Fault"

Could anybody have any clue for this.knowing that the same OS,Database & Applications work fine on a stand alone AS ES45 with direct attached SCSI disks!!
Regards

[Antoniov.]

Hello,
help from VMS about ASTFLT says:
ASTFLT, AST fault, SP='xxxxxxxx', param='xxxxxxxx', PC='location',
PSL='xxxxxxxx', target PC='location', PSL='xxxxxxxx'

Facility: SYSTEM, System Services

Explanation: The system attempted to deliver an asynchronous system trap (AST) to an image, but the user stack pointer is either invalid or does not have sufficient space for the AST call frame. This message indicates an exception condition and is usually followed by a display of the condition arguments, the registers, and the stack at the time of the exception.

User Action: Examine the parameters displayed in the message to determine the instruction that caused the error. Locate the programming error that modified the stack.

By the help seems a application error but you could increase stack page for kernel modifying KSTACKPAGES as follow:
$ SET DEF SYS$SYSTEM
$ MC SYSMAN
SYSMAN> PARAM SHOW KSTACKPAGES
Example
Parameter Name Current [...]
-------------- ------- -------
KSTACKPAGES 1 [...] Pages
SYSMAN> PARAM SET KSTACKPAGES 2
SYSMAN> PARAM WRITE CURRENT
SYSMAN> EXIT
===>>> Notice: after this operation you need reboot to make active the new value!

Bye
Antoniov

[Åge Rønning]

Have you checked with Oracle?
You don't mention your RDB version, But I've seen this message been reported against several RDB versions some years ago.

Setting BUGCHECKFATAL=1 would give you a dump if more information is needed to trace this down.