RE: Disconnect job / end job question on locks

I concur ref CPF5140. It's the M.O. around here for users to just click the X when they leave for the day, so I see that message in joblogs a lot.

Once the job ends, I wouldn't expect locks to persist. Do her joblogs concur regarding when she closed her 5250, or did she close them quickly after you called?



-----Original Message-----
From: Jeff Crosby [mailto:jlcrosby@xxxxxxxxxxxxxxxx]
Sent: Thursday, May 03, 2018 6:44 AM
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Subject: Disconnect job / end job question on locks

V7R3, IBM i is in the cloud and we connect via site-to-site VPN.

So yesterday I leave work 30 minutes early for an appointment. About 5 minutes out I get a call from the office manager that the "Daily Reports"
batch job (creates and emails a bunch of status reports) is at a standstill, not moving. So I head back to work. Within a minute I get a text that QSYSOPR has a message that needs a response (got a love/hate relationship with message queue monitors).

I arrive at work and the message in QSYSOPR is the wonderful RNQ1218 Unable to allocate a record in file <filename> (R C G D F).

Before replying, I do a WRKOBJLCK <filename> to see who has it and what they're doing. A buyer is updating an open PO. I reply R, then call the buyer. She claims she is not on at all, no 5250 sessions open and she has been signed off for 5-10 minutes. I do another WRKOBJLCK <filename> and the updating a PO lock is gone. Remember, this is all in less that a minute and she claims to have been off at least 5 minutes.

I look at her spool files for joblogs. She started 4 5250 sessions just after 8am. Every one of them has this message, within seconds of each
other:

CPF5140 Diagnostic 70 05/02/18 16:52:36.989193
QWSERROR QSYS 0573 QWSGET QSYS 0683
Message . . . . : Session stopped by a request from device <device>.

Cause . . . . . : The request shutdown was caused by either the user
turning the power off, by a device error, or the ASCII controller inactivity timer expired.
Recovery . . . : Close the files and vary the device off (VRYCFG
command). If the problem occurs again, enter the ANZPRB command to run problem analysis.

Each job ended (per the joblog) within seconds of the above message. And one of the joblogs shows that she was in the PO updating program at the time.

Doesn't this indicate that within the ACS 5250 session (and without signing off), she did one of 1) clicked Communication/disconnect, 2) File/exit (or exit all), or 3) Clicked the X in the upper right corner of the window to close it? I'm pretty sure the answer to this is yes. Furthermore, I note she has done this several days in a row.

The secondary question is would the job locks (like that file) hang around for some minutes? Doesn't seem to me like they would, but . . .

Thanks.