Ever used the time parameter on Save option 21?

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Well Steve, the answer has gotta be in whatever is processing your inactive message queue. We set up our system to process the data queue via a CL program I developed. It has numerous conditions in it determining whether a job should be ended or just disconnected. There is special logic there that doesn't touch the console -unless- it is after 8pm. I assumed that if the console is still active after 8, then someone forgot to sign it off and that job should be ended. It sure sounds like you all have the same setup, it explains the scenario perfectly.</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Oh yeah, and stop calling me Lucy. :)</FONT></DIV> <DIV></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> If the IBM owner of save 21 isn't smart enough to realize that, in most situations, the system is gonna kill the job before time for the save to start, what's the point in adding the feature?</P></BLOCKQUOTE>

[tim.shafar@fiserv.com]

It will still run in qctl. The Save works great. The only thing to consider is this. 1. You have to sign on to the System Counsel and then leave it unattended. Any body that can do a sysrq can option 2 and kill the job. Then they have full access to your system as Secofr. Is it worth it? That up to you. Tim

[Brenda Hefel]

We just tried it last weekend and it didn't work. Did option 21 from GO SAVE. The screen was inhibited and got the message, Save submitted at 16:40 Save will start at 11:00. I came in at 17:30 the next day and it was still inhibited and looking at the joblog I can see where the job issued DLYJOB RSMTIME(11:00). It then tried to do an ENDSBS, but not until 17:46. Does anyone have any suggestions?

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Hmm, I haven't tried that command before, but let's try something anyway. I will -assume- that it was doing an EndSbs with the command defaults of Option(*Cntrld) and Delay(*Nolimit) . So, if we take a look at what was really happening, the system couldn't end the subsystems in a controlled manor -and- since it was to wait forever (the result of *Nolimit), you get the results you experienced.</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Oops, I just looked at the help for that menu option and this is what it says: </FONT></DIV> <DIV><FONT face=Arial size=2>

ENDSBS SBS(*ALL) OPTION(*IMMED) SAVSYS SAVLIB LIB(*NONSYS) ACCPTH(*YES) SAVDLO DLO(*ALL) FLR(*ANY) SAV OBJ(('/*) ('/QSYS.LIB' *OMIT) ('/QDLS' *OMIT) UPDHST(*YES) STRSBS SBSD(controlling-subsystem)

</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>So, can you take a look at the joblog again and tell us exactly what the EndSbs command looked like? And maybe even what messages, if any, followed.</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Bill</FONT></DIV> <DIV></DIV> <DIV></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px">We just tried it last weekend and it didn't work. Did option 21 from GO SAVE. The screen was inhibited and got the message, Save submitted at 16:40 Save will start at 11:00. I came in at 17:30 the next day and it was still inhibited and looking at the joblog I can see where the job issued DLYJOB RSMTIME(11:00). It then tried to do an ENDSBS, but not until 17:46. Does anyone have any suggestions?</P></BLOCKQUOTE>

[Guest.Visitor]

I posed this question in preparation for a V4R5 upgrade. My thinking was that I could set the time for the save to start and it would be finished by the time I came in to start the upgrade. Friday 13:41 - Initiated the save. Set the time on the parm to 03:00:00. Got a message that the save was submitted and would start at 03:00. Friday 13:42 - Save process checked for tape in drive. All's well. Friday 16:15 - Checked on job. WRKACTJOB showed DLY 03:00:00. All's well. Went home. Friday 20:20 - "All jobs at work station DSP01 ended." I don't know why this happened. The sysval QINACTITV is set for 15 minutes. QCONSOLE is DSP01. I was signed on as QSECOFR. I have a daily process that begins at 20:00. It sends a BRKMSG to *ALLWS at 20:00, 20:15, and 20:25 before ending QINTER at 20:30. I suppose some interaction occurred between the DLYJOB and the SNDBRKMSG that caused DSP01 to end. Your thoughts? Steve

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Steve,</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Did you save the joblog for DSP01? It should tell you/us everything about what happened.</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Bill</FONT></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> I don't know why this happened. The sysval QINACTITV is set for 15 minutes. QCONSOLE is DSP01. I was signed on as QSECOFR. </P></BLOCKQUOTE>

[Guest.Visitor]

Yep, sure did. (paraphasing) 13:41:34 - Job DSP01 started. 13:41:38 - Go save. 13:42:00 - Volume found on device TAP01. 20:20:36 - All jobs at work station DSP01 ended. 20:20:36 - Job DSP01 ended. That's it. The job log is about half a page long. The "All jobs at work station DSP01 ended" message is CPI1127 which goes into some detail about exceeding the QINACTITV sysval but I don't believe that's an accurate description of what happened. The system was up and running from 13:41 until about 16:15 so I'm pretty sure that the save didn't occur at that time and there is nothing on the (new) tape that I used. Thanks for your help, Steve

[Guest.Visitor]

Hey All, Yes I agree with you on the SAVE option 21, it does have bugs. I developed a CL program that does SAVE 21. It keeps me away from the shop on the Weekends most of the time. Its Interactive. One of The Things I had to do was Disable the QINACTITV & QDSCJOBITV so my job did not end. DLYJOB until it hits the Time. If anyone is interested in Code let me know Steve

[Guest.Visitor]

<BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> That's it. The job log is about half a page long. The "All jobs at work station DSP01 ended" message is CPI1127 which goes into some detail about exceeding the QINACTITV sysval but I don't believe that's an accurate description of what happened. The system was up and running from 13:41 until about 16:15 so I'm pretty sure that the save didn't occur at that time and there is nothing on the (new) tape that I used.</P></BLOCKQUOTE> <FONT face=Arial size=2>Oh but I think that it -is- an accurate description. Think about what happened, your save job sat waiting for the time to begin, but the system said "hey the job at DSP01 has been sitting around inactive (no data interchange), so I'm killing it". Make sense?</FONT> <FONT face=Arial size=2>Bill</FONT></P>

[Guest.Visitor]

Lu-u-u-cy, you've got some 'splainin to do . . . WHY did it wait 6 hours and some change before it killed the job? QINACTITV is 15 minutes. If the IBM owner of save 21 isn't smart enough to realize that, in most situations, the system is gonna kill the job before time for the save to start, what's the point in adding the feature? Disgustedly, Steve

[Guest.Visitor]

We use Option 21 with the time parm every Friday and it works great. I wonder how you are filling in the "Prompt for commands" parm. You will want that to be N, also you will want your "Message queue delivery" to be *notify not break. Those are about the only 2 things that I can imagine that would hose it up..... Bruce Shultz

[Guest.Visitor]

Bruce - Like you, I would think that would hose it up, too, so I entered 'N' for 'Prompt for Commands' and '*NOTIFY for 'Message queue delivery'. What is your QINACTITV sysval? What time do you enter the command on Friday? What time does the save actually start? Thanks, Steve

[Guest.Visitor]

After reviewing the entire thread I have to agree with an earlier post that wondered if there could be a job that runs to kill the inactive jobs, you might want to check your job scheduler.... We enter the command at about 11am on Fridays and set it for about 2300, and it issues the ENDSBS at 2300.Our sysval for QINACTITV is *NONE(so yeah that probably is your problem),though the one you really want to check is QINACTMSGQ because that is the one that determines what action is taken against inactive jobs. If nothing else write a CL to change the time interval for QINACTITV to *NONE(or manually change it for now) and anonther to change it back(at a time when you think your save will have finished) and put them on your Job Scheduler. By the way Wayne Evans book on security has a little program in it that uses the prior mentioned sysvals but allows you to exclude some user profiles (such as QSECOFR) depending on their ACGCDE.

[Guest.Visitor]

Well, I tried again to use the time parm on the SAVE 21 screen and once again I was struck down by the cruel hand of fate . . . Friday 13:17 - Initiated the save. Set the time on the parm to 03:00:00. Got a message that the save was submitted and would start at 03:00. Friday 13:19 - Save process checked for tape in drive. All's well. Friday 16:15 - Checked on job. WRKACTJOB showed DLY 03:00:00. All's well. Went home. Monday 5:00 - "Subsystem ending immediately." (This is caused by a weekly IPL that occurs Mondays at 5:00 AM). Obviously, I didn't get zapped by QINACTITV this time but, once again, I have no idea what happened. Other jobs (PM/400, for instance) continued to run over the weekend and several jobs that run from the job scheduler ran normally. In fact, the normal daily backup, which I had put on 'hold' since I was running the option 21 save, issued a message at 0300 stating that it wouldn't be run because it was held. Any thoughts on this one? Thanks, Steve

[Guest.Visitor]

<DIV><FONT face=Arial size=2>What does the job log say happened at 3am on Saturday?</FONT></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> Friday 13:17 - Initiated the save. Set the time on the parm to 03:00:00. Got a message that the save was submitted and would start at 03:00. Friday 16:15 - Checked on job. WRKACTJOB showed DLY 03:00:00. All's well. Went home. Monday 5:00 - "Subsystem ending immediately." (This is caused by a weekly IPL that occurs Mondays at 5:00 AM).</P></BLOCKQUOTE>

[Guest.Visitor]

The job log shows that absolutlely, positively nothing occurred between 04/27/01 13:19:29 and 04/30/01 05:00:02. Steve

[Guest.Visitor]

<DIV><FONT face=Arial size=2>OK, then paste the joblog leading up to and including the DlyJob command. Specifically, I would like to see the actual DlyJob command.</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Bill</FONT></DIV> <DIV></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px">The job log shows that absolutlely, positively nothing occurred between 04/27/01 13:19:29 and 04/30/01 05:00:02. </BLOCKQUOTE>

[Guest.Visitor]

There is no DLYJOB to see in the joblog, there is only a blank line between 04/27/01 13:19:29 and 04/30/01 05:00:02. When you enter a time on the option 21 screen, it checks the tape and then goes into a DLYJOB that you can see via WRKACTJOB but this does not appear in the joblog. I'm going to write my own CL and forget about the time parameter. Thanks, Steve

[J.Panzenhagen]

Are you using an Async console (Client Access)? I am trying to do a 21 on my 270 using Asysnc console and the console keeps blowing up. Seeing as this is the only console available (no twinax controller) I am at a loss as to how to ge this backup. Any ideas?

[Guest.Visitor]

John - No, sorry . . . good ol' twinax for me! Steve

[Guest.Visitor]

I just can't let this thread die - I wrote my own CL and still had no luck. The first statement in the CL is DLYJOB RSMTIME('03:00:00'). I started it from the console (QCTL) yesterday at 16:45 expecting it to kick off at 3:00 AM this morning and NOTHING HAPPENED! The joblog is attached. I'm starting to wonder if anything works the way it should anymore. Any suggestions? Thanks, Steve

QPJOBLOG095106.txt (http://www.mcpressonline.com/images/fbfiles/files/1e2731ab_QPJOBLOG095106.txt)

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Well unfortunately we can't see the CL statements in steve/savsystem, so this is really no help. Change your CL logging level and try it again.</FONT></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> I wrote my own CL and still had no luck. The first statement in the CL is DLYJOB RSMTIME('03:00:00'). I started it from the console (QCTL) yesterday at 16:45 expecting it to kick off at 3:00 AM this morning and NOTHING HAPPENED! The joblog is attached. I'm starting to wonder if anything works the way it should anymore. </P></BLOCKQUOTE>

[Guest.Visitor]

Allrighty then - I did a RTVCLSRC and attached it . . .

savsystem.txt (http://www.mcpressonline.com/images/fbfiles/files/1e278848_savsystem.txt)

[cwscholbe@dstsystems.com]

Try taking the seconds off. Our DLYJOB looks like this one. DLYJOB RSMTIME('03:00') You never know.

[David Abramowitz]

Try adding the CHGJOB command as your first statement. Make sure CL logging is turned on, the the job log listing set to *SECLVL. This will ensure that you get a job log no matter what happens - assuming that the job is run at all. Dave

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Well Steve, at first glance it looks ok. If you have to remove the seconds part of the RsmTime parameter, I would consider that a glich in the operating system. Which brings up one subject: are you up to date with PTF's?</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Regardless, David's suggestion about changing the job's logging is the one to try next and might as well try Chris' suggestion about removing the seconds.</FONT></DIV> <DIV></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px"> DLYJOB RSMTIME('03:00:00') ENDSBS SBS(*ALL) OPTION(*IMMED) </BLOCKQUOTE>

[Guest.Visitor]

Only available time to run a 21 is late Saturday evenings. Any suggestions besides coming in late Friday night or early Saturday morning?

[Guest.Visitor]

Alan, I would review what option 21 does and then evaluate whether it needs to be run at all. Option 21 runs the following: SAVSYS SAVLIB LIB(*NONSYS) ACCPTH(*YES) SAVDLO DLO(*ALL) FLR(*ANY) SAV OBJ(('/*) ('/QSYS.LIB' *OMIT) ('/QDLS' *OMIT) UPDHST(*YES) I only use option 21 immediately after installing a new release or if it is convenient (that is not coming in on a Saturday night). When I apply a CUM or have a lot of outstanding system changes I do a SAVSYS. Between SAVSYS runs, I log what I have done. The most important thing to save is your data, which does not require a dedicated system. David Morris

[Guest.Visitor]

Has anyone ever tried using the time parameter to start a full system save at a specified time? Does the command still have to run interactively in QCTL when using the time? Have any troubles when using the time parm? Just Curious, Steve

[Guest.Visitor]

Time parm only allows 24 hr advance setup. My only window available to run a 21 is late Saturday evenings. Any suggestions?

[Guest.Visitor]

<DIV><FONT face=Arial size=2>Alan,</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>If your job uses the DlyJob command, the other option besides RsmTime is the Dly parameter which will allow you to delay the job for the specified number of seconds. The maximum value allowed for the Dly parameter equates to over 11 days in waiting. Can you use this?</FONT></DIV> <DIV></DIV> <DIV><FONT face=Arial size=2>Bill</FONT></DIV> <BLOCKQUOTE style="BORDER-LEFT: #000000 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px; PADDING-RIGHT: 0px">Time parm only allows 24 hr advance setup. My only window available to run a 21 is late Saturday evenings. Any suggestions?</BLOCKQUOTE>