Re: SHUTTRAP
BTW, CP seems to accept any virtual system reset or disabled wait as a signal to remove a userid from the list of users in the SHUTDOWN Pending list. That is as I expected. I presume that cancellation is also accepted. :-) Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark Sent: Monday, July 04, 2011 8:18 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP On Monday, 07/04/2011 at 09:12 EDT, David Boyes dbo...@sinenomine.net wrote: SA22-7832-08, pg. 11-16, in the section on Warnings. There's about two sentences referring to power failure. The rest is in the SHUTTRAP source code. I said it was tiny. 8-). Indeed, and vague, too. :-) According to SHUTTRAP source, the LPAR deactivation event (Quiesce) is reflected via an unsolicited Service Signal external interrupt with event 0x1D, not a Warning machine check. So it's definitely in the not published camp. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
On Monday, 07/04/2011 at 09:12 EDT, David Boyes dbo...@sinenomine.net wrote: SA22-7832-08, pg. 11-16, in the section on Warnings. There's about two sentences referring to power failure. The rest is in the SHUTTRAP source code. I said it was tiny. 8-). Indeed, and vague, too. :-) According to SHUTTRAP source, the LPAR deactivation event (Quiesce) is reflected via an unsolicited Service Signal external interrupt with event 0x1D, not a Warning machine check. So it's definitely in the not published camp. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
My watchdog machine as you call it, is the PROP running in OPERATOR. At each logoff of a user, my RELOG EXEC checks if: a) Is RELOGing in effect? If yes, then continue to check if: b) This user must be on at all times unconditionally c) This user must be on IF certain conditions apply - and checks the conditions. The exec that is run when SHUTTRAP catches a shutdown signal starts its work by signaling PROP to turn off reloging. Only after that does it start bringing down the various machines it is responsible to stop. If I were to cancel the SHUTDOWN, I would have to re-enable RELOGing, but to date I have not built in such a function. I have also never cancelled a SHUTDOWN, so if it ever happens, I will need to do the enable manually. (Of course, I better document this, or my successor won't know why RELOG stopped working...) Shimon On Wed, Jun 29, 2011 at 12:21 AM, Schuh, Richard rsc...@visa.com wrote: Given the lack of direction, I suppose that you can do almost anything, so long as you are consistent. There is no right or wrong, so logging off will work so long as the watchdog machine does not bring logged off users back on.
Re: SHUTTRAP
On Wednesday, 06/29/2011 at 10:11 EDT, Shimon Lebowitz shim...@iname.com wrote: If I were to cancel the SHUTDOWN, I would have to re-enable RELOGing, but to date I have not built in such a function. I have also never cancelled a SHUTDOWN, so if it ever happens, I will need to do the enable manually. (Of course, I better document this, or my successor won't know why RELOG stopped working...) I had never cancelled one either, until this note stream popped up. Then I began to play with issuing SIGNAL SHUTDOWN to my 2nd level system (simulating LPAR deactivation) and then cancelled the shutdown. What it revealed is that CP issues a bogus message, telling me that SHUTDOWN REIPL has been cancelled. :-) And I wish CP would decide if he's a Brit or a Yank. Sometimes he says cancelled and sometimes canceled. But maybe that's ok - so do I. :-) Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
SHUTTRAP
I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this: SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh
Re: SHUTTRAP
Richard, I don't see that on my 5.4 system. Here is how I start it. /* */ say 'loading SHUTTRAP as a CMS nucleaus extension' 'nucxload shuttrap (system' say 'enabling SGNLSHUT exec to process shutdown signal' shuttrap exec sgnlshut exit And the sgnlshut exec. /* exec to perform some functions before SHUTDOWN' */ say 'Received shutdown signal -- saving QUERY DASD' 'pipe cp q dasd | shutdown dasd a' say 'done! -- waiting for shutdown' address command cp logoff exit I don't really do anything with this data, just a couriosity thing. Tom On Tue, Jun 28, 2011 at 8:59 AM, Schuh, Richard rsc...@visa.com wrote: I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this: SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh
Re: SHUTTRAP
Hello Richard, I believe the reason for your error message in using SHUTTRAP is that SHUTTRAP can only issue CMS commands and not CP commands. Is the message associated with your SHUTTRAP issuing a CP command? I had the same problem when my message was a CP command. As a test for SHUTTAP I issued the CMS command 'Q DISK' and it worked with no error message. Hope this helps. Raymond E. Noal EMC² where information lives Phone: (508) 249-4076 Ext: 44076 From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Schuh, Richard Sent: Tuesday, June 28, 2011 11:59 AM To: IBMVM@LISTSERV.UARK.EDU Subject: SHUTTRAP I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this: SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh
Re: SHUTTRAP
My RxKernel has ShutTrap support. RxServer uses WAKEUP (EXT to catch the interrupt. No abends then. And, to complete handling a shutdown signal, one is supposed to load an FFF wait state, not to LOGOFF. My first attempt was to end with a LOGOFF. But then our automation saw the server go logoff and started it again. So now I end with the official method: 'CP STORE PSW 000A 8FFF' 2011/6/28 Raymond Noal raymond.n...@emc.com Hello Richard, ** ** I believe the reason for your error message in using SHUTTRAP is that SHUTTRAP can only issue CMS commands and not CP commands. Is the message associated with your SHUTTRAP issuing a CP command? I had the same problem when my message was a CP command. As a test for SHUTTAP I issued the CMS command ‘Q DISK’ and it worked with no error message. ** ** Hope this helps. ** ** *Raymond E. Noal* * * * EMC²* *where information lives* * * *Phone: (508) 249-4076* *Ext: 44076* ** ** *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On Behalf Of *Schuh, Richard *Sent:* Tuesday, June 28, 2011 11:59 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* SHUTTRAP ** ** I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this:** ** SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh -- Kris Buelens, IBM Belgium, VM customer support
Re: SHUTTRAP
Thanks Kris, that is not obvious, but it makes sense.. On Tue, Jun 28, 2011 at 11:07 AM, Kris Buelens kris.buel...@gmail.comwrote: My RxKernel has ShutTrap support. RxServer uses WAKEUP (EXT to catch the interrupt. No abends then. And, to complete handling a shutdown signal, one is supposed to load an FFF wait state, not to LOGOFF. My first attempt was to end with a LOGOFF. But then our automation saw the server go logoff and started it again. So now I end with the official method: 'CP STORE PSW 000A 8FFF' 2011/6/28 Raymond Noal raymond.n...@emc.com Hello Richard, ** ** I believe the reason for your error message in using SHUTTRAP is that SHUTTRAP can only issue CMS commands and not CP commands. Is the message associated with your SHUTTRAP issuing a CP command? I had the same problem when my message was a CP command. As a test for SHUTTAP I issued the CMS command ‘Q DISK’ and it worked with no error message. ** ** Hope this helps. ** ** *Raymond E. Noal* * * * EMC²* *where information lives* * * *Phone: (508) 249-4076* *Ext: 44076* ** ** *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On Behalf Of *Schuh, Richard *Sent:* Tuesday, June 28, 2011 11:59 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* SHUTTRAP ** ** I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this:* *** SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh -- Kris Buelens, IBM Belgium, VM customer support
Re: SHUTTRAP
That code is what CP will load when its shutdown is complete. So, not related to signal shutdown's wait code. I don't remember where I found the code to load, my guess is in the z Series Principles of Operations (the shutdown signal is not a z/VM invention, but a z HW feature that z/VM also virtualizes). 2011/6/28 Schuh, Richard rsc...@visa.com ** No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? Regards, Richard Schuh ** ** ** -- *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On Behalf Of *Tom Huegel *Sent:* Tuesday, June 28, 2011 11:53 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* Re: SHUTTRAP Thanks Kris, that is not obvious, but it makes sense.. On Tue, Jun 28, 2011 at 11:07 AM, Kris Buelens kris.buel...@gmail.comwrote: My RxKernel has ShutTrap support. RxServer uses WAKEUP (EXT to catch the interrupt. No abends then. And, to complete handling a shutdown signal, one is supposed to load an FFF wait state, not to LOGOFF. My first attempt was to end with a LOGOFF. But then our automation saw the server go logoff and started it again. So now I end with the official method: 'CP STORE PSW 000A 8FFF' 2011/6/28 Raymond Noal raymond.n...@emc.com Hello Richard, I believe the reason for your error message in using SHUTTRAP is that SHUTTRAP can only issue CMS commands and not CP commands. Is the message associated with your SHUTTRAP issuing a CP command? I had the same problem when my message was a CP command. As a test for SHUTTAP I issued the CMS command ‘Q DISK’ and it worked with no error message. Hope this helps. *Raymond E. Noal* ** * EMC²* *where information lives* ** *Phone: (508) 249-4076 %28508%29%20249-4076* *Ext: 44076* *From:* The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] *On Behalf Of *Schuh, Richard *Sent:* Tuesday, June 28, 2011 11:59 AM *To:* IBMVM@LISTSERV.UARK.EDU *Subject:* SHUTTRAP I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this: SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh -- Kris Buelens, IBM Belgium, VM customer support -- Kris Buelens, IBM Belgium, VM customer support
Re: SHUTTRAP
I do not find any mention of shutdown in the copy of z/Architecture Principals of Operation that I just searched. Like I said, the documentation is difficult to find. Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Kris Buelens Sent: Tuesday, June 28, 2011 1:05 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP That code is what CP will load when its shutdown is complete. So, not related to signal shutdown's wait code. I don't remember where I found the code to load, my guess is in the z Series Principles of Operations (the shutdown signal is not a z/VM invention, but a z HW feature that z/VM also virtualizes). 2011/6/28 Schuh, Richard rsc...@visa.commailto:rsc...@visa.com No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? Regards, Richard Schuh From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Tom Huegel Sent: Tuesday, June 28, 2011 11:53 AM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP Thanks Kris, that is not obvious, but it makes sense.. On Tue, Jun 28, 2011 at 11:07 AM, Kris Buelens kris.buel...@gmail.commailto:kris.buel...@gmail.com wrote: My RxKernel has ShutTrap support. RxServer uses WAKEUP (EXT to catch the interrupt. No abends then. And, to complete handling a shutdown signal, one is supposed to load an FFF wait state, not to LOGOFF. My first attempt was to end with a LOGOFF. But then our automation saw the server go logoff and started it again. So now I end with the official method: 'CP STORE PSW 000A 8FFF' 2011/6/28 Raymond Noal raymond.n...@emc.commailto:raymond.n...@emc.com Hello Richard, I believe the reason for your error message in using SHUTTRAP is that SHUTTRAP can only issue CMS commands and not CP commands. Is the message associated with your SHUTTRAP issuing a CP command? I had the same problem when my message was a CP command. As a test for SHUTTAP I issued the CMS command 'Q DISK' and it worked with no error message. Hope this helps. Raymond E. Noal EMC² where information lives Phone: (508) 249-4076tel:%28508%29%20249-4076 Ext: 44076 From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Schuh, Richard Sent: Tuesday, June 28, 2011 11:59 AM To: IBMVM@LISTSERV.UARK.EDUmailto:IBMVM@LISTSERV.UARK.EDU Subject: SHUTTRAP I just retrieved SHUTTRAP from the download page and tried it. In my first simpleminded try, the results were mixed. The signal shutdown did indeed trigger the command included with the shuttrap command. Then I got this: SHUTTRAP: Shutdown in 900 seconds DMSITP143T Addressing exception occurred at 81056162 in system routine WAITRD; re-IPL CMS 15:07:56 * MSG FROM RSCHUH1 : DMSDIE3550I All APPC/VM and IUCV paths have been severed. HCPGIR450W CP entered; disabled wait PSW 000A 80F3F752 There is nothing in the documentation that would lead me to believe that this is normal; however, the documentation consists solely of the HELP file and is pretty sparse in this area. I tried reassembling the module with the same results. z/VM 6.1 Service Level 1002 CMS 25, Service Level 002 Regards, Richard Schuh -- Kris Buelens, IBM Belgium, VM customer support -- Kris Buelens, IBM Belgium, VM customer support
Re: SHUTTRAP
On Tuesday, 06/28/2011 at 03:10 EDT, Schuh, Richard rsc...@visa.com wrote: No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? You will not find any published information on the mechanism. I *observe* that if you SIGNAL SHUTDOWN to a 2nd level VM, it will load wait state FFF and the SIGNAL SHUTDOWN command gives a successful termination message. Draw your own conclusions. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
Given the lack of direction, I suppose that you can do almost anything, so long as you are consistent. There is no right or wrong, so logging off will work so long as the watchdog machine does not bring logged off users back on. Is there any way for a user to determine if it is being targeted alone, by a signal shutdown user userid or is included in a signal shutdown all? I suspect not, but I can see where that might be useful at times. Also, CP does notice when a user who has requested the shutdown signal logs off and cancels the request, does it not? It must because a virtual machine can log off at any time, regardless of the shutdown signal, and CP has to handle it correctly. Does CP disallow new logons when it receives signal shutdown all command? Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark Sent: Tuesday, June 28, 2011 1:56 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP On Tuesday, 06/28/2011 at 03:10 EDT, Schuh, Richard rsc...@visa.com wrote: No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? You will not find any published information on the mechanism. I *observe* that if you SIGNAL SHUTDOWN to a 2nd level VM, it will load wait state FFF and the SIGNAL SHUTDOWN command gives a successful termination message. Draw your own conclusions. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
Richard, You can use the CP command - 'Q SIGNALS' to make your determinations, as in - q signals Signalled Timeout UseridSignalSignal Status By Remaining LN192209 SHUTDOWN Enabled - - VMSERVR SHUTDOWN Enabled - - VMSERVU SHUTDOWN Enabled - - VMSERVS SHUTDOWN Enabled - - Happy parsing!! Raymond E. Noal EMC² where information lives Phone: (508) 249-4076 Ext: 44076 -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Schuh, Richard Sent: Tuesday, June 28, 2011 6:22 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP Given the lack of direction, I suppose that you can do almost anything, so long as you are consistent. There is no right or wrong, so logging off will work so long as the watchdog machine does not bring logged off users back on. Is there any way for a user to determine if it is being targeted alone, by a signal shutdown user userid or is included in a signal shutdown all? I suspect not, but I can see where that might be useful at times. Also, CP does notice when a user who has requested the shutdown signal logs off and cancels the request, does it not? It must because a virtual machine can log off at any time, regardless of the shutdown signal, and CP has to handle it correctly. Does CP disallow new logons when it receives signal shutdown all command? Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark Sent: Tuesday, June 28, 2011 1:56 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP On Tuesday, 06/28/2011 at 03:10 EDT, Schuh, Richard rsc...@visa.com wrote: No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? You will not find any published information on the mechanism. I *observe* that if you SIGNAL SHUTDOWN to a 2nd level VM, it will load wait state FFF and the SIGNAL SHUTDOWN command gives a successful termination message. Draw your own conclusions. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
Close, but I am not willing to grant you a cigar. If all, or even most, have the same Signalled by id, then either a signal shutdown all was done or a program did multiple signal shutdown user userid commands in rapid sequence. Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Raymond Noal Sent: Tuesday, June 28, 2011 3:38 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP Richard, You can use the CP command - 'Q SIGNALS' to make your determinations, as in - q signals Signalled Timeout UseridSignalSignal Status By Remaining LN192209 SHUTDOWN Enabled - - VMSERVR SHUTDOWN Enabled - - VMSERVU SHUTDOWN Enabled - - VMSERVS SHUTDOWN Enabled - - Happy parsing!! Raymond E. Noal EMC² where information lives Phone: (508) 249-4076 Ext: 44076 -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Schuh, Richard Sent: Tuesday, June 28, 2011 6:22 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP Given the lack of direction, I suppose that you can do almost anything, so long as you are consistent. There is no right or wrong, so logging off will work so long as the watchdog machine does not bring logged off users back on. Is there any way for a user to determine if it is being targeted alone, by a signal shutdown user userid or is included in a signal shutdown all? I suspect not, but I can see where that might be useful at times. Also, CP does notice when a user who has requested the shutdown signal logs off and cancels the request, does it not? It must because a virtual machine can log off at any time, regardless of the shutdown signal, and CP has to handle it correctly. Does CP disallow new logons when it receives signal shutdown all command? Regards, Richard Schuh -Original Message- From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Alan Altmark Sent: Tuesday, June 28, 2011 1:56 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: SHUTTRAP On Tuesday, 06/28/2011 at 03:10 EDT, Schuh, Richard rsc...@visa.com wrote: No only is it not obvious, so far I have been unable to find it in the documentation. You can specify a wait state code on the SHUTDOWN command and it can be anything between 1 and . I cannot find where the default is FFF. Is that a true system default or is it a code you have adopted for your use? You will not find any published information on the mechanism. I *observe* that if you SIGNAL SHUTDOWN to a 2nd level VM, it will load wait state FFF and the SIGNAL SHUTDOWN command gives a successful termination message. Draw your own conclusions. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
On Tuesday, 06/28/2011 at 07:35 EDT, Schuh, Richard rsc...@visa.com wrote: Close, but I am not willing to grant you a cigar. If all, or even most, have the same Signalled by id, then either a signal shutdown all was done or a program did multiple signal shutdown user userid commands in rapid sequence. You cannot tell the difference between SIGNAL SHUTDOWN USER and SIGNAL SHUTDOWN ALL. If the signal is the result of an LPAR deactivation (or 2nd level equivalent), then the Signalled By user will be SYSTEM. Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: SHUTTRAP
You will not find any published information on the mechanism. There is a tiny amount of information in the POP manual in the external interrupt section on the original LPAR deactivation signal and what's supposed to happen when it triggers, but the end comment is that the effect will be implementation dependent.
Re: SHUTTRAP
On Tuesday, 06/28/2011 at 10:06 EDT, David Boyes dbo...@sinenomine.net wrote: You will not find any published information on the mechanism. There is a tiny amount of information in the POP manual in the external interrupt section on the original LPAR deactivation signal and what's supposed to happen when it triggers, but the end comment is that the effect will be implementation dependent. What POP are you looking at? I can't find any reference to LPAR deactivation. The only machine-dependent external interrupt I can find is Service Signal and that covers a lot of ground, notifying the OS of the completion of some previously requested machine function. The 9730/9221 used to reflect now on batteries! and now on utility power! via an external damage machine check. The logical response to batteries! would be to almost-but-not-quite shutdown and quiesce. That is, go into an enabled wait for utility! and then come roaring back to life with an almost-but-not-quite IPL. (The machine even had a published DIAGNOSE to cause the machine to power down.) Alan Altmark Senior Managing z/VM and Linux Consultant IBM System Lab Services and Training ibm.com/systems/services/labservices office: 607.429.3323 mobile; 607.321.7556 alan_altm...@us.ibm.com IBM Endicott
Re: Wakeup and Shuttrap
I've been reading and testing: - I get WAKEUP's RC 6 too, but when calling WAKEUP again, it directly exist with RC 1. So, the SHUTTRAP signal triggers 2 WAKEUP events: COSN SMSG. Removing the CONS option doesn't do anything. - SHUTTRAP does not work like ADDRESS CMS, (so an eventual CP EXEC wouldn't matter) - Diverting the SMSG delivery method from VMCF to IUCV doesn't help either. - Using SHUTTRAP CP MSG * STOP doesn't change a thing. Bypasses: When getting RC=6, test if another SMSG event is ready too: 'WAKEUP +5 ( RDR SMSG FILE(HOBBIT TIMES *)' if rc=6 then do /* RC 6 caused by SHUTTRAP? */ saveRc=rc /* If yes, WAKEUP also has */ 'WAKEUP +0 (SMSG' /* an SMSG event stacked */ if rc=2 then rc=saveRc end Or, use the approach I ised in RxServer: I run WAKEUP with the EXT option and test for external interrupt 2401 (see subrtn Extrn_event:); SHUTTRAP can be loaded with the following: 'PIPE COMMAND NUCXDROP SHUTTRAP' 'NUCXLOAD SHUTTRAP (SYSTEM' 'SHUTTRAP' Disadavantage: when using EXT in a server that is conencted to an active SFS, WAKEUP will get many EXT 4000 interrupts (your code can ignore them, but the handling is extra overhead anyway). 2007/11/8, Thomas Kern [EMAIL PROTECTED]: The SIGNAL SHUTDOWN command was issued from another authorized user. Ther e is no CP EXEC. I thought that having TERMIAL LINEND OFF might have affect ed it, but I reset it to LINEND # and it did not change the result. I will l ook inside my test copy of VMUTIL to see if I can use SHUTTRAP there and if s o, what the differences are. /Tom Kern /301-903-2211 On Thu, 8 Nov 2007 21:56:31 +0100, Kris Buelens [EMAIL PROTECTED] wrote: I wouldn't know why it would become a console interrupt; my RxServers support SHUTTRAP, and they work well. How did you send the SIGNAL SHUTDOWN command. My guess is that you type d it in on the console of user HOBBIT2, hence the stop with RC 6 of WAKEUP. Or, you have an CP EXEC in HOBBIT2 that stacks something (the CONS on WAKEUP means that WAKEUP should stop when something is in the stack ( I wouldn't be surprized if SHUTTRAP interprets its commands as if they are sent with ADDRESS CMS). -- Kris Buelens, IBM Belgium, VM customer support
Re: Wakeup and Shuttrap
I'd say: test it, look at subroutine Extrn_event: and code a SAY when you get a 4000 interrupt, than you can see how often it happens in your environment. I know we can live with the number. You don't get an EXT 4000 interrupt for each SFS call, far from that. From what I tested, I see that one gets an EXT 4000 when you have an SFS dir accessed and someone issued a GRANT AUTH for you for a file in that subdirectory (the SFS server has to broadcast this to you). Maybe if you'd use Asynch CSL calls to SFS you'll get much more. 2007/11/9, Thomas Kern [EMAIL PROTECTED]: Well, at least I did not make some silly error. When you use WAKEUP to trap the EXT interrupts, do you get alot of the EXT 4000 interrupts queued up when some exec that gets called as a TIMER event needs to query an SFS server via CSL calls? Hobbit client DISK function reports on CP allocations (PAGE, SPOOL, DRCT, TDISK) and optionally on SFS server allocations. /Tom Kern /301-903-2211 Kris Buelens wrote: I've been reading and testing: * I get WAKEUP's RC 6 too, but when calling WAKEUP again, it directly exist with RC 1. So, the SHUTTRAP signal triggers 2 WAKEUP events: COSN SMSG. Removing the CONS option doesn't do anything. * SHUTTRAP does not work like ADDRESS CMS, (so an eventual CP EXEC wouldn't matter) * Diverting the SMSG delivery method from VMCF to IUCV doesn't help either. * Using SHUTTRAP CP MSG * STOP doesn't change a thing. Bypasses: When getting RC=6, test if another SMSG event is ready too: 'WAKEUP +5 ( RDR SMSG FILE(HOBBIT TIMES *)' if rc=6 then do /* RC 6 caused by SHUTTRAP? */ saveRc=rc /* If yes, WAKEUP also has */ 'WAKEUP +0 (SMSG' /* an SMSG event stacked */ if rc=2 then rc=saveRc end Or, use the approach I ised in RxServer: I run WAKEUP with the EXT option and test for external interrupt 2401 (see subrtn Extrn_event:); SHUTTRAP can be loaded with the following: 'PIPE COMMAND NUCXDROP SHUTTRAP' 'NUCXLOAD SHUTTRAP (SYSTEM' 'SHUTTRAP' Disadavantage: when using EXT in a server that is conencted to an active SFS, WAKEUP will get many EXT 4000 interrupts (your code can ignore them, but the handling is extra overhead anyway). -- Kris Buelens, IBM Belgium, VM customer support
Re: Wakeup and Shuttrap
Thanks, I will look into it when I get into the office. I don't think the Hobbit code actually accesses any directories itself and I have mine installed on a minidisk. But I will have to check it when Hobbit is installed in an SFS allocation. /Tom Kern /301-903-2211 Kris Buelens wrote: I'd say: test it, look at subroutine Extrn_event: and code a SAY when you get a 4000 interrupt, than you can see how often it happens in your environment. I know we can live with the number. You don't get an EXT 4000 interrupt for each SFS call, far from that. From what I tested, I see that one gets an EXT 4000 when you have an SFS dir accessed and someone issued a GRANT AUTH for you for a file in that subdirectory (the SFS server has to broadcast this to you). Maybe if you'd use Asynch CSL calls to SFS you'll get much more. 2007/11/9, Thomas Kern [EMAIL PROTECTED] mailto:[EMAIL PROTECTED]: Well, at least I did not make some silly error. When you use WAKEUP to trap the EXT interrupts, do you get alot of the EXT 4000 interrupts queued up when some exec that gets called as a TIMER event needs to query an SFS server via CSL calls? Hobbit client DISK function reports on CP allocations (PAGE, SPOOL, DRCT, TDISK) and optionally on SFS server allocations. /Tom Kern /301-903-2211 Kris Buelens wrote: I've been reading and testing: * I get WAKEUP's RC 6 too, but when calling WAKEUP again, it directly exist with RC 1. So, the SHUTTRAP signal triggers 2 WAKEUP events: COSN SMSG. Removing the CONS option doesn't do anything. * SHUTTRAP does not work like ADDRESS CMS, (so an eventual CP EXEC wouldn't matter) * Diverting the SMSG delivery method from VMCF to IUCV doesn't help either. * Using SHUTTRAP CP MSG * STOP doesn't change a thing. Bypasses: When getting RC=6, test if another SMSG event is ready too: 'WAKEUP +5 ( RDR SMSG FILE(HOBBIT TIMES *)' if rc=6 then do /* RC 6 caused by SHUTTRAP? */ saveRc=rc /* If yes, WAKEUP also has */ 'WAKEUP +0 (SMSG' /* an SMSG event stacked */ if rc=2 then rc=saveRc end Or, use the approach I ised in RxServer: I run WAKEUP with the EXT option and test for external interrupt 2401 (see subrtn Extrn_event:); SHUTTRAP can be loaded with the following: 'PIPE COMMAND NUCXDROP SHUTTRAP' 'NUCXLOAD SHUTTRAP (SYSTEM' 'SHUTTRAP' Disadavantage: when using EXT in a server that is conencted to an active SFS, WAKEUP will get many EXT 4000 interrupts (your code can ignore them, but the handling is extra overhead anyway). -- Kris Buelens, IBM Belgium, VM customer support
Re: Wakeup and Shuttrap
On Fri, 9 Nov 2007 10:15:36 -0600, Thomas Kern [EMAIL PROTECTED] wrote: Now I have to clean it up just a bit more. Which brings up the next tangent on this discussion. When I signal a linux SVM to shutdown, CP reports th at the linux SVM has signalled its termination. When I am logged onto a lin ux SVM and the shutdown signal comes in, linux terminate and loads a disabl ed wait PSW (x0FFF) but does not log off. Is this PSW code (x0FFF) the indicator to CP that SHUTDOWN has completed? If it is, how can I nicely load the same indicator when my SVM has completed its shutdown processing? Use the Linux kernel option VMPOFF=LOGOFF to have the Linux guest logof f after it has shutdown. Brian Nielsen
Re: Wakeup and Shuttrap
Logging off seems like a pretty good signal. :-) You can put any CP command you want in the VMPOFF, so you could use MSG or SMSG to send a signal back to a shutdown SVM, which could wait until all guests have reported in before it issues the real CP SHUTDOWN. Brian Nielsen On Fri, 9 Nov 2007 10:55:12 -0600, Thomas Kern [EMAIL PROTECTED] wrote: My current linux setup works just fine without adding the VMPOFF=LOGOF F. If it is DISCONNECTED when SIGNAL SHUTDOWN arrives, it shuts down, loads th at disable wait PSW and CP logs it off. When I am logged onto it and use th e CLASS G version of SIGNAL SHUTDOWN, linux shuts down, loads the disabled wait and then I get to do what I want with the virtual machine without C P logging me off. What I wanted to know was is there a better way to signal CP that the shutdown processing has been completed besides a CP STORE PSW ? /Tom Kern /301-903-2211 On Fri, 9 Nov 2007 10:34:17 -0600, Brian Nielsen [EMAIL PROTECTED] wrote: On Fri, 9 Nov 2007 10:15:36 -0600, Thomas Kern [EMAIL PROTECTED] wrote: Now I have to clean it up just a bit more. Which brings up the next tangent on this discussion. When I signal a linux SVM to shutdown, CP reports that the linux SVM has signalled its termination. When I am logged onto a linux SVM and the shutdown signal comes in, linux terminate and loads a disabled wait PSW (x0FFF) but does not log off. Is this PSW code (x0FFF) the indicator to CP that SHUTDOWN has completed? If it is, how can I nicel y load the same indicator when my SVM has completed its shutdown processing? Use the Linux kernel option VMPOFF=LOGOFF to have the Linux guest log off after it has shutdown. Brian Nielsen = === = ===
Re: Wakeup and Shuttrap
A CLASS G linux guest only needs to signal CP that ITS own shutdown processing has been completed and apparently the x0FFF in the PSW does th at. An SVM or Operator or Sysprog that issued the SIGNAL SHUTDOWN command wil l get the HCPSIG2113I message when the target SVM has completed its shutdow n (if it has loaded that magic PSW), so they could be running a program to intercept the HCPSIG2113I messages and continue on the rest of the processing if all targets complete before the specified WITHIN value. /Tom Kern /301-903-2211 On Fri, 9 Nov 2007 11:05:04 -0600, Brian Nielsen [EMAIL PROTECTED] wrote: Logging off seems like a pretty good signal. :-) You can put any CP command you want in the VMPOFF, so you could use MSG or SMSG to send a signal back to a shutdown SVM, which could wait until all guests have reported in before it issues the real CP SHUTDOWN. Brian Nielsen
Re: Wakeup and Shuttrap
On Fri, 9 Nov 2007 10:15:36 -0600, Thomas Kern [EMAIL PROTECTED] wrote: Is this PSW code (x0FFF) the indicator to CP that SHUTDOWN has completed? If it is, how can I nicely load the same indicator when my SVM has completed its shutdown processing? Sorry, I mis-read your question the first time around. You can use the class G command CP STORE PSW to change you virtual machines PSW. Brian Nielsen
Re: ? SHUTTRAP ?
For those sites who have it enabled, do you have this IUCV statement in your directory? It appears to be a pre-requisite but isn't mentioned in the discussion here or on the VM list. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ken Meyer Sent: August 3, 2006 12:59 PM To: VSE Discussion List Subject: Re: ? SHUTTRAP ? According to the documentation, there are two things to consider: 1) IUCV ALLOW ANY *SIGNAL in the VM directory 2) CP SET SIGNAL to determine shutdown delay. That is all I have found so far. Ken Meyer CSI Ron Schmiedge wrote: You can add this mix to your scorecard - I am z/VM 4.4 and z/VSE 3.1.1 and Q SIGNALS does not show the VSE guests enabled. snip..