Improvement of the Stop behavior on VMS side (statistics usages on log files)

Would it be possible to have a clean stop of Connx binaries on VMS side (+add of system usage statistics of the stopped processes into the log files) instead of having a simple “stop/id” ?

For now, when stopping Connx from Linux side and when connection were up/established, there is a kind of clean sigkill which log some statistics on Connx Log file, at VMS side.

However, when doing a Connx stop on VMS side (even with connection established), it do a violent kill on VMS (stop/id) and at this moment no statistics are logged on VMS log.

It would be nice to have a statistics, when doing a connx stop on VMS side, a dump should be made of connx statistics before killing the processes.

Indeed, our system team have implemented some statistics dump when a Connx processes is properly closed.

Example:

a/ When connections are going down from JDBC linux server (stop of jdbc server, or simply a connection close), here below what we have into the Connx Log file :

138233730: CNXBATCHBUFFER size reduced to 64000 because CNXCONNECTBACK was set to 0, and max QIO buffer size is 65535

138233731: CNXBATCHBUFFER size reduced to 64000 because CNXCONNECTBACK was set to 0, and max QIO buffer size is 65535

138233732: Process = CNX_4B_BAT01_CO!

*******fin************************

BAT01_CONNX job terminated at 20-MAY-2019 12:16:43.13

de Accounting information:

Buffered I/O count: 52743 Peak working set size: 14656

Direct I/O count: 19818 Peak virtual size: 189856

Page faults: 1164 Mounted volumes: 0

Charged CPU time: 0 00:00:08.77 Elapsed time: 0 12:39:13.62

The highlighted blue part was an adaptation from our system team in order to have statistics of the processes usage.

b/ When we are doing a Connx STOP at VMS side (The stop on VMS side consist to a kill of Connx deamon and main listener), here what we have on the Connx log files :

138122750: CNXBATCHBUFFER size reduced to 64000 because CNXCONNECTBACK was set to 0, and max QIO buffer size is 65535

138122750: CNXBATCHBUFFER size reduced to 64000 because CNXCONNECTBACK was set to 0, and max QIO buffer size is 65535

138122750: Process = CNX_05_BAT01_CO!

[End of file]

Buffer: CNXS_BG33374_BAT01_20190519_OLA3M1.LOG | Read-only | Unmodified

This time, as the process was directly killed by stop/id, there is no dump possibility of system usage into the log files…

This statistics could be nice for debugging or to follow the performances issues (or other kinds of usages…).