|
HP OpenVMS Systemsask the wizard |
|
The Question is:
Dear Wizard!
I have a process running on my VAX system, whose job is to monitor the other
applications running on the system and notfy the support people whenever any
of them is down! I am using the SYS$BRKTHRU system service (not wait).
Our support people log into the VAX using the terminal emulator (PowerTerm).
The problem starts when some of them keep their sessions open for a long
time.. or if does not log off properly. (The terminal device hangs out,
never gets cleared! for example,
if I do a stop/id=
on an interactive process!).The terminal device does not get cleared. In
such cases, when my process tries to notify a user, it goes to RWAST state!
The only solution I am aware of to clear this situation is to schedule a
reboot.. which is very costly in
a real-time system!
I did an analyze/system and found the following:
Current process summary
-----------------------
Extended Indx Process name Username State Pri PCB PHD
Wkset
-- PID -- ---- --------------- ----------- ------- --- -------- --------
------
2A000EDD 00DD CTX-MONITOR GANGA RWAST 6 80B63D40 841E4C00
729
Process index: 00DD Name: CTX-MONITOR Extended PID: 2A000EDD
----------------------------------------------------------------
Channel Window Status Device/file accessed
------- ------ ------ --------------------
0090 8176C880
00B0 00000000 Busy TNA252:
00C0 00000000 Busy TNA252:
00D0 00000000 Busy TNA252:
00E0 00000000 Busy TNA252:
00F0 00000000 Busy TNA252:
0100 00000000 Busy TNA252:
0110 00000000 Busy TNA252:
0120 00000000 Busy TNA252:
0130 00000000 Busy TNA252:
0140 00000000 Busy TNA252:
0150 00000000 Busy TNA252:
0160 00000000 Busy TNA252:
0170 00000000 Busy TNA252:
Press RETURN for more.
SDA>
Process index: 00DD Name: CTX-MONITOR Extended PID: 2A000EDD
----------------------------------------------------------------
Channel Window Status Device/file accessed
------- ------ ------ --------------------
0180 00000000 Busy TNA252:
0190 00000000 Busy TNA252:
01A0 00000000 Busy TNA252:
01B0 00000000 TNA260:
node$ sh dev TNA252:/full
Terminal TNA252:, device type VT200 Series, is online, record-oriented
device,
carriage control.
Error count 0 Operations completed
32433
Owner process "" Owner UIC
[SYSTEM]
Owner process ID 2A001C84 Dev Prot
S:RWLP,O:,G:,W:
Reference count 18 Default buffer size
80
node$
node$ sh proc/id=2A001C84
%SYSTEM-W-NONEXPR, nonexistent process
node$
I will be thankful for any suggestions/work-arounds! Well... this is an old
DSSI cluster with VMS 5.5 running!!
Thanks and Regards,
Ganga
The Answer is : There is unfortunately not much that the OpenVMS Wizard can ascertain from the information provided -- this initially appears that it might related to a lost or mishandled I/O request in a network driver. Please consider an upgrade to a more recent OpenVMS version, please seriously consider an upgrade to a more recent version of TCP/IP Services (with the associated ECO kit) or whichever IP package is installed, and please see the previous discussion here in Ask The Wizard around the HANGUP processing that can be necessary in certain applications. If you wish to pursue this, please contact the Compaq Customer Support Center -- expect to be asked to upgrade to at least OpenVMS VAX V5.5-2 and (depending on the version presently in use) to upgrade and ECO your TCP/IP Services environment.
|