2979907 - SIgnal 11 with empty stack traces - SAP ASE

•  ASE errorlog reports an "infected with Signal 11" error with a stack trace:

Current process (0x9d004f) infected with signal 11 (SIGSEGV)
Address 0x2020203d20202020 (), siginfo (code, address) = (50, 0x2020203d20202020)

• The stack trace includes the following functions:

pcstkwalk
ucstkgentrace
ucbacktrace
terminate_process

  Note: A complete stack trace is available under Attachments.

• On AIX**, if the "secondary signal handling" configuration parameter set to 0, it is followed by the message and the ASE server keeps running:

Task 10289231 is made zombie because it cannot be cleaned up during its termination. This task may have resources allocated to it.

• On AIX**, if the "secondary signal handling" configuration parameter set to 1 (default), it is followed by the looping messages with similar stack traces:

FATAL UNHANDLED EXCEPTION: signal 11 hit while handling a previously hit signal. The server cannot continue and will shut down.

  The ASE server ends up with detecting a stack overflow and a stack guardword corruption, and it crashes.

• The terminated user process was attempting to run the following SQL:

DECLARE      @client                 Nvarchar(100)
SET @client = 'xxxxx'
Declare @sql Nvarchar(3000)
Set  @sql = '
CREATE OR REPLACE PROCEDURE dbo.procname _' + @client  + '
( @my_id       INT
...
) 
AS' +
/*
...
...

*******************************/
+ '
BEGIN
….. ‘ 

• ** SAP's repro on Linux shows that the user process is terminated cleanly. No zombie tasks are left behind and ASE does not crash.

SAP Adaptive Server Enterprise (ASE) 16.0

CR823126, CR#823126, 823126, SEGV, segmentation, storage, access, violation, fault, segfault, kisignal, stacktrace,  registers, literal, autoparam, literal_autoparam , KBA , kill query , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Bug Filed