HPUX 11.11/Progress 10.1B03
We have had a situation for the last four weeks where a user or process generates a number (maybe 8 or 10) of logins to an SQL server process - always between 4 and 5 in the on Monday mornings
We believe that since our users have been set up with ODBC using a the Data Direct 5.2 driver, that the connection is being initiated in this manner.
The SQL server process crashes and then the database crashes with a locked buffer error. The syslog error message is:
1090012 May 9 04:11:27 a300sue2 vmunix: 1090013 May 9 04:11:27 a300sue2 vmunix: Pid 29340 killed due to trashed stack. 1090014 May 9 04:11:27 a300sue2 vmunix: Pid 29340 was killed due to failure in writing the signal context. 1090015 May 9 04:11:27 a300sue2 vmunix: 1090016 May 9 04:11:40 a300sue2 su: + tty?? root-mfgeb 1090017 May 9 04:11:41 a300sue2 vmunix: 1090018 May 9 04:11:41 a300sue2 vmunix: Pid 22208 killed due to trashed stack. 1090019 May 9 04:11:41 a300sue2 vmunix: Pid 22208 was killed due to failure in writing the signal context. 1090020 May 9 04:11:41 a300sue2 vmunix: The Progress log shows: [2011/05/09@04:11:46.373-0400] P-2715 T-1 I BROKER 1: (1153) BROKER detects death of server 29340. [2011/05/09@04:11:46.373-0400] P-2715 T-1 I BROKER 1: (1153) BROKER detects death of server 22208. [2011/05/09@04:11:46.373-0400] P-2715 T-1 I BROKER 1: (8839) No SQL servers are available. Try again later. [2011/05/09@04:11:50.863-0400] P-2715 T-1 I BROKER 1: (1153) BROKER detects death of server 29340. [2011/05/09@04:11:50.863-0400] P-2715 T-1 I BROKER 1: (1153) BROKER detects death of server 22208. [2011/05/09@04:11:50.863-0400] P-2715 T-1 I BROKER 1: (8839) No SQL servers are available. Try again later. [2011/05/09@04:11:59.902-0400] P-219 T-1 I BROKER 0: (2526) Disconn ecting client 102 of dead server 6. [2011/05/09@04:11:59.902-0400] P-219 T-1 I BROKER 0: (2523) User 10 2 died with 1 buffers locked. [2011/05/09@04:11:59.902-0400] P-2801 T-1 I AIW 14: (2520) Stopped . [2011/05/09@04:11:59.907-0400] P-4274 T-1 I SRV 4: (2520) Stopped . All active processes are then stopped and the database halts with an abnormal shutdown code. We've been trying to troublemshoot this by approaching users and our sql report base - and we're looking at putting a sniffer on the net to see who or what is accessing the sql server. But we're wondering if anyone has any other ideas. TIA.