|
The 4XTRA system monitor can be started from any PC attached to the 4XTRA network.
| |
 |
To access the monitor sign on as HMW with the required password and answer your first name or OPER to the "who are you" question. |
| |
|
You will then be positioned in the systems menu over the monitor option |
| |
|
Select Monitor and you will be presented with the Monitor screen |
| |
|
Possibly on selecting Monitor, an error will be indicated and you will need to check the error status, even though there may be none. Once this has been check you can return to the top level of the monitor subsystem. |
The Error Log
| |
|
All errors in 4XTRA are logged centrally. This allows support and development to home in on any potential problems and resolve them. The error log will thus contain a mix of user errors and 4XTRA batch errors. As the error log is held as a sequential list, you can go through the log, change direction, or quit. You will gradually begin to recognise errors, and what to do to resolve them. |
Batch Monitor
The Batch Monitor displays a graphic representation of the 4XTRA job steps, highlighting which ones are active, becoming active, completed, or in error. The job steps at Nordea are
| |
|
The Batch monitor display contains the following areas: |
| |
|
A title bar at the top of the screen, showing the area being monitored |
| |
|
A graphic representation of the 4XTRA background tasks |
| |
|
Up to three task status lines showing details on the batch machines activities. |
| |
|
A menu bar at the bottom of the screen showing available F keys (separate displays) and the printer status. |
Each task status lines show the following information
| |
|
Task code |
| |
|
Task description |
| |
|
Task status |
| |
|
Machine number |
| |
|
Start time of job |
| |
|
Expected or actual finish time (estimated from yesterday’s time) |
| |
|
Running time also estimated from yesterday |
| |
|
The status is updated approximately every 20 seconds |
The six displays are
| |
|
F1: Graph: Show a full graph of the dependencies of all batch tasks, plus their current status. A key explains the colour coding used. Selecting a task code in the graph shows its status near the bottom of the screen, together with the last three lines of its message file. |
| |
|
F2: MSG: Shows the message file for the highlighted task in the status display. Use up and down arrows or the mouse to change the selected task. Use Home, End, Page up and Page down to scroll through the message file. |
| |
|
F3: Alerts: Shows all non-information messages in the central message file (REMOTE.MSG). This includes system generated errors, warnings and crashes. New Alerts, which have previously not been seen are highlighted. |
| |
|
F4: Info: Shows all messages in the central message file. New messages are highlighted. |
| |
|
F5: Msgs: Shows the last few lines of all message files of the tasks listed in the task status lines. |
| |
|
F6: Repair: Allows remote repair of batch tasks - see below |
Heartbeat
| |
|
Each 4XTRA batch task regularly writes to a heartbeat file associated with the machine upon which the task is running. 4ARMS checks how long it has been since each task has written to this file, and compares this with a time out value in the Task Static. If the time out value has been exceeded, an intermittent alarm sounds, and the task status line is highlighted in yellow. When and if the heartbeat resumes, the alarm is cleared automatically. |
| |
|
The time out for daytime tasks is set to 60 seconds. The time out for most batch tasks is set at 120 seconds. Some tasks have longer time-outs, since they involve steps that can take longer than 120 seconds. |
| |
|
Occasionally a time out will occur even though a task is still functioning correctly. This has been noticed at the end of DAYRUN'P', when the task waits for 60 seconds to let the final deals clear before shutting down and in some reports like STLOGREP which can sometimes take longer if the activity during the day has been high. |
| |
|
If a time-out occurs, this may be because the machine running the task has mysteriously died. To investigate further, you will need to visit the machine and check it. The clock in the right hand top corner should be working, and various changes on the screen should show if it is still active. If necessary, reboot or re-power on the offending machine. |
Errors
| |
|
When a code error or system generated error or warning occurs during a batch task, and this error is correctly trapped, the 4ARMS display changes to F3: Alerts, and a continuous alarm sounds. The alarm is cancelled by pressing any key. See remote repair below for further information. If an error has occurred on a 4XTRA user machine, it will also be displayed on the monitor screen. When this happens, the user should be instructed to telephone you to ask for help. If necessary, you may need to go to the user machine, break in and investigate the error more deeply. If it is an error that you know, you may be able to instruct the user over the phone. |
| |
|
If a user has an error, his 4XTRA terminal will display the error with an alert box saying OK / HALT SYSTEM. The support team should select Halt System, and then type in the HMW password. (If this fails, hit enter and retype the password until it works). Once 4XTRA has accepted this, it will display the APL stack and the last error message and leave you in APL desk calculator mode. You can now investigate further, and if necessary ask the user what he has been doing. If the error can be recreated, tell HMW and ask for a program update. |
| |
|
Note that the user can not recover from one of these errors. Their only re-course is to boot. Do not encourage this as it can be destructive and you may miss the chance of collecting information on a problem area which could be useful in resolving the problem. |
Remote repair
| |
|
When an error occurs on a batch machine, it can be remedied by using the 4ARMS remote repair facility. This should only be used by HMW personnel or other authorised people. The remote repair facility provides a simulated APL session with some restrictions. |
| |
|
Select F6: Repair on the 4ARMS display, then use the arrow keys or the mouse to select the task to be repaired. Hit enter, and the window in the centre of the screen should display any pending output from the task (e.g. an error report) and prompt for input. |
| |
|
The remote repair facility uses IN and OUT, so in principle any line can be entered, including system command lines. Certain commands are filtered out, however. |
| |
|
)CMD is executed on the local machine, so be warned against entries like )CMD BOOT. CMD can be used, but be careful not to use it in such a way as to loose control of the keyboard. |
| |
|
)EDIT is trapped and causes a copy of the function - only function editing is allowed - to be transferred to the local machine for editing as a variable. Control-E sends the new text back, and a DEFN ERROR will be reported at that stage if either the header is malformed, or the function is pendant. F10 is programmed with )EDIT SI which is interpreted appropriately by the remote repair input handler. |
| |
|
Be warned against using functions which require keyboard input such as INKEY and ARBIN. Also oPUT tends to prompt for who you are, so resist using it, unless the global variable whoami has been manually set. |
| |
|
After rectifying the problem and continuing with #goto LC, ESC will take you out if remote repair. When inputting to remote repair, a blank line (Enter alone) will also end remote repair. |
|
