Re: z/VM page space
But since Linux does know how to evacuate swap space, you can define a new VDISK for swap and free the old VDISK. Once you detach it, the slots on page space will be freed up. Whether it makes sense to shake up your memory reference patterns like that is an entirely different issue... | Rob Ok, I had forgotten Linux could do that, so that could get all of the vdisk pages moved (assuming there are no other vdisk uses), but it does stir the pot in terms of recency of reference (all of those pages have to become resident and then age back out through the paging hierarchy (which isn't ideal), and it won't get other types of CP owned pages, but it might be enough for some shops. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On Wed, Jun 22, 2011 at 4:07 PM, Bill Holder hold...@us.ibm.com wrote: But since Linux does know how to evacuate swap space, you can define a new VDISK for swap and free the old VDISK. Once you detach it, the slots on page space will be freed up. Whether it makes sense to shake up your memory reference patterns like that is an entirely different issue... | Rob Ok, I had forgotten Linux could do that, so that could get all of the vdisk pages moved (assuming there are no other vdisk uses), but it does stir the pot in terms of recency of reference (all of those pages have to become resident and then age back out through the paging hierarchy (which isn't ideal), and it won't get other types of CP owned pages, but it might be enough for some shops. Right, you'd have to be careful since you (briefly) increase your memory requirement. I only recommend that when the cause of the swapping is gone already and Linux has released most slots on the swap disk again. CP would continue to care for the pages even though Linux does not need the contents. In that case there's little stirring and the benefit is that CP can then release the backing of the VDISK. PS Instead of the discussed wibni's, I'd rather have DIAG10 for VDISK on the wish list (call it CMM-3 if that looks better on your PBC). The things in Linux to exploit that are mostly there (aka TRIM support for SSD devices). | Rob -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Right, you'd have to be careful since you (briefly) increase your memory requirement. I only recommend that when the cause of the swapping is gone already and Linux has released most slots on the swap disk again. CP would continue to care for the pages even though Linux does not need the contents. In that case there's little stirring and the benefit is that CP can then release the backing of the VDISK. PS Instead of the discussed wibni's, I'd rather have DIAG10 for VDISK on the wish list (call it CMM-3 if that looks better on your PBC). The things in Linux to exploit that are mostly there (aka TRIM support for SSD devices). | Rob It would be perhaps a bit underhanded to use it this way (in terms of it being an undocumented side effect), but we already do have a Diag 10 type Trim function for vdisks - a format write is essentially converted into a page release call when simulated. I'd have to check with our I/O guys exactly which CCW that is, and whether there are any restrictions on when it results in the page release (probably needs to be at least 4K in size, since vdisks are 512 byte FBA devices, but Linux usage is 4K anyway, correct?). And if folks really think an official, documented Trim function for vdisks is required, by all means, open a requirement. It likely won't get implemented otherwise. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Just a idea, but what if we put the back link on the DASD page (I don't know much about the structure of the page, but I think there's a header there somewhere), not on central memory? We would only have to read the back link on the page header on DASD and we would know where the page is, migrate it and release the disk. Mauro http://mauro.limeiratem.com - registered Linux User: 294521 Scripture is both history, and a love letter from God. On Tue, Jun 21, 2011 at 2:09 AM, Alan Altmark alan_altm...@us.ibm.comwrote: On Monday, 06/20/2011 at 08:56 EDT, Clovis Pereira gclo...@br.ibm.com wrote: At least, we seeded the idea. The requirement to provide this function has been on the books and on VM Development's collective conscience since 1990. :-( *Everyone* runs into a dasd migration/reclamation project at some point that forces an IPL. (IBMers, too, btw.) The good news is that this conversation has re-focused the Eye of Sauron once again upon the requirement. (In this case, that's a Good Thing.) If anyone would like to express their support for this requirement, contact the Support Center or your rep/BP and ask to have your name added to the Interested Parties list of requirement MR00025368, which covers both spooling and paging. Alan Altmark Senior Managing z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Just a perhaps crazy idea. - Add new page volumes - drain all old page vlumes to stop adding new data there - for each Linux, one at a time change the reserved memory to the same as memsize (for this to work, z/VM now must restore all paged memory into real memory) - wait until this is done (hehe, have no clue here about the time or how to check it out, sorry) - remove that reserved memory (or is it dedicated?) - and all new paging will we on the new page volumes Any chance this can work anyone ? :) Cordialement / Vriendelijke Groeten / Best Regards / Med Vänliga Hälsningar Tore Agblad Tore Agblad System programmer, Volvo IT certified IT Architect Volvo Information Technology Infrastructure Mainframe Design Development, Linux servers Dept 4352 DA1S SE-405 08, Gothenburg Sweden Telephone: +46-31-3233569 E-mail: tore.agb...@volvo.com http://www.volvo.com/volvoit/global/en-gb/ From: Linux on 390 Port [LINUX-390@VM.MARIST.EDU] On Behalf Of Mauro Souza [thoriu...@gmail.com] Sent: Tuesday, June 21, 2011 15:20 To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space Just a idea, but what if we put the back link on the DASD page (I don't know much about the structure of the page, but I think there's a header there somewhere), not on central memory? We would only have to read the back link on the page header on DASD and we would know where the page is, migrate it and release the disk. Mauro http://mauro.limeiratem.com - registered Linux User: 294521 Scripture is both history, and a love letter from God. On Tue, Jun 21, 2011 at 2:09 AM, Alan Altmark alan_altm...@us.ibm.comwrote: On Monday, 06/20/2011 at 08:56 EDT, Clovis Pereira gclo...@br.ibm.com wrote: At least, we seeded the idea. The requirement to provide this function has been on the books and on VM Development's collective conscience since 1990. :-( *Everyone* runs into a dasd migration/reclamation project at some point that forces an IPL. (IBMers, too, btw.) The good news is that this conversation has re-focused the Eye of Sauron once again upon the requirement. (In this case, that's a Good Thing.) If anyone would like to express their support for this requirement, contact the Support Center or your rep/BP and ask to have your name added to the Interested Parties list of requirement MR00025368, which covers both spooling and paging. Alan Altmark Senior Managing z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Just a idea, but what if we put the back link on the DASD page (I don't know much about the structure of the page, but I think there's a header there somewhere), not on central memory? We would only have to read the back link on the page header on DASD and we would know where the page is, migrate it and release the disk. I think that would probably break NSS/DCSS processing, since those can appear in multiple address spaces, and you'd need to fix all those references too. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
-Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Agblad Tore Sent: Tuesday, June 21, 2011 9:37 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space Just a perhaps crazy idea. - Add new page volumes - drain all old page vlumes to stop adding new data there - for each Linux, one at a time change the reserved memory to the same as memsize (for this to work, z/VM now must restore all paged memory into real memory) - wait until this is done (hehe, have no clue here about the time or how to check it out, sorry) - remove that reserved memory (or is it dedicated?) - and all new paging will we on the new page volumes Any chance this can work anyone ? :) Nope. CP does some paging on it's own behalf that isn't associated with a virtual machine, so no amount of fiddling with virtual machines will free up those pages. CP has to be convinced to move them, and there's no external way outside CP to do that at the moment. Alan's suggestion is probably the most useful at this point -- visible pressure on IBM by everybody signing up as interested parties will likely get this fixed sooner than trying to speculate. This is Deep Magic Code; not to be trifled with lightly. Better to let Bill Holder worry about it. 8-) -- db -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Just a idea, but what if we put the back link on the DASD page (I don't know much about the structure of the page, but I think there's a header there somewhere), not on central memory? We would only have to read the back link on the page header on DASD and we would know where the page is, migrate it and release the disk. Mauro That could work, though it would mean redefining the CP owned volume format to contain room for the additional metadata, which would be somewhat disruptive (particularly for spool space and release compatability). Each record would have to be at least 4104 bytes to contain the 4K page plus 8 bytes (likely more) for the pointer. An alternative would be to keep the pointers separate from the 4K records they describe, but then that just adds another mapping to maintain. Either way, I'm not sure it's a better solution overall than scanning, nor am I sure it really would result in a smaller (shorter time to market) project - there's lots of potential for unforeseen disruption. But it's certainly worth keeping in mind. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Just a perhaps crazy idea. - Add new page volumes - drain all old page vlumes to stop adding new data there - for each Linux, one at a time change the reserved memory to the same as m= emsize (for this to work, z/VM now must restore all paged memory into real memor= y) - wait until this is done (hehe, have no clue here about the time or how to= check it out, sorry) - remove that reserved memory (or is it dedicated?) - and all new paging will we on the new page volumes Any chance this can work anyone ? :) Cordialement / Vriendelijke Groeten / Best Regards / Med V=E4nliga H=E4lsni= ngar Tore Agblad Hi - there are (at least) two problems: First, issuing SET RESERVED does not cause z/VM to actively restore (page in) any paged-out pages (make them resident in storage), it merely prevents any more from being stolen. Second, the SET RESERVED protection is not absolute - if the Demand Scan (page replacement) function reaches the 3rd pass emergency scan, the SET RESERVED setting is not honored and the pages will be taken away anyway. z/VM provides no permanent way to dedicate storage. However, this approach might work (for user pages) if the CP LOCK command is used rather than SET RESERVED - LOCK will bring the pages in. However, there are various things the guest program may choose to do to defeat the LOCK command (any page release function such as Diag x'10'), and it also won't help with CP system-owned pages (including vdisk pages, for what it's worth), which cannot be locked via the CP LOCK command. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
I think that would probably break NSS/DCSS processing, since those can appe= ar in multiple address spaces, and you'd need to fix all those references t= oo. That does indeed add complexity, but is solvable - for shared NSS/DCSS pages, the owning entity is the NSS/DCSS itself, that's what we'd point back to (they each have their own set of DAT structures). Luckily, the forward pointer that needs to be found and updated is in the PGMBK, which for shared segments, is shared across all of the users who have it loaded (so we wouldn't need to visit each user). For exclusive pages, since each user gets their own copy (effectively of the PGMBK, not just the individual page) on first reference, we'd point back to the user structures as for any other private user owned page. ... Better to let Bill Holder worry about it. Thanks. I've got plenty of worrying going already, but this is on the list (including all of the interesting ideas generated from this discussion). :-) - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On Tue, Jun 21, 2011 at 10:31 PM, Bill Holder hold...@us.ibm.com wrote: and it also won't help with CP system-owned pages (including vdisk pages, for what it's worth), which cannot be locked via the CP LOCK command. But since Linux does know how to evacuate swap space, you can define a new VDISK for swap and free the old VDISK. Once you detach it, the slots on page space will be freed up. Whether it makes sense to shake up your memory reference patterns like that is an entirely different issue... | Rob -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Thanks, Bill. Good explanation, enough for me. At least, we seeded the idea. Best regards, __ Clovis From: Bill Holder hold...@us.ibm.com To: LINUX-390@vm.marist.edu Date: 17/06/2011 18:39 Subject: Re: z/VM page space Sent by: Linux on 390 Port LINUX-390@vm.marist.edu Hi, People. My cent to this discussion. We are talking about pages on dasd, not pages in memory. So, these pages already was swapped. Except some control pages, I think. To free the dasd, VM will need: Know what pages are on the specific dasd. It can be done by Page out routines, updating one table. Drain the dasd, to prevent new allocation Put on the bit to force a Page In for the pages on that dasd. Returning to normal page process. After some time, no pages are on this dasd. It can be purged from the page tables, removing the control pages. Of course, it is not so simple to implement, but looks like viable. What do you, the experts, think about? __ Clovis Hi Clovis - I'm not sure I completely follow what you're proposing, but I believe I already covered it (though maybe I wasn't very clear). What you describe seems similar to the design approach I suggested. The challenge is this: There is is only one way to do your first step Know what pages are on the specific DASD; that is to scan all pages in all pageable spaces using CP's translation tables to determine whether each page is on that DASD. Once all pages on the DASD have been located, moving them is easy enough. Finding them is the hard part. The problem is that there is no way to get from the structures representing the 4K records on the DASD (the paging slots containing the pages) back to the structures which tell us what page is in each slot. All we have to represent each paging slot is one bit (allocated or not) in a bitmap, there is no room for a back pointer to the description of the page that the slot contains. So we have to find the pages the other way - from the top down, scanning all pages in the system to check if they are on the volume in question. This is not a small effort. You might ask why we don't just add that missing back pointer; it seems easy enough. It would indeed be easy enough to code, but the cost in terms of the additional storage usage to hold all of those pointers would be prohibitive - for a maximum sized paging system using ECKD volumes pushed to today's supported limits (about 11TB of paging space), it would mean at least an additional 22GB of non-pageable central storage usage just to hold those back pointers - for a function that is only very rarely used. Using a more real-world example: if you have a 32GB system with a 4:1 overcommit ratio of total virtual to real, with paging configured according to recommendations, those back pointers would take up at least 512MB of non-pageable central storage. It just seems excessively wasteful to spend that much central storage resource representing things that aren't even in storage, especially when it's only needed on a very rare basis. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On Monday, 06/20/2011 at 08:56 EDT, Clovis Pereira gclo...@br.ibm.com wrote: At least, we seeded the idea. The requirement to provide this function has been on the books and on VM Development's collective conscience since 1990. :-( *Everyone* runs into a dasd migration/reclamation project at some point that forces an IPL. (IBMers, too, btw.) The good news is that this conversation has re-focused the Eye of Sauron once again upon the requirement. (In this case, that's a Good Thing.) If anyone would like to express their support for this requirement, contact the Support Center or your rep/BP and ask to have your name added to the Interested Parties list of requirement MR00025368, which covers both spooling and paging. Alan Altmark Senior Managing z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
I doubt CP moves the pages unless it needs to pull the page into memory for actual use. Otherwise that would be page migration. A need to move spool is indeed rare. This idea is a mutant variant of the Do you know where the NSSes are? notion. Come up clean with 1 spool volume, build your segments, then pull other spool volumes online. Of course this goes back to the day when spooling was done on multiple mod3s. Now I just spool wherever - even 27s - as long as I have tested a SYSTEM RESTART invoked dump to insure that there is enough room to hold the dump. Now we just need need to update standalone dumping to go to DASD. David Original Message Subject: Re: z/VM page space From: David Boyes dbo...@sinenomine.net Date: Thu, June 16, 2011 11:59 pm To: LINUX-390@VM.MARIST.EDU On 6/16/11 7:36 PM, David Kreuter dkreu...@vm-resources.com wrote: evil thought. Come up without paging space - paging early life will go to spool. Have the autolog machine bring up the real page volumes. Hmm. You know, that might just work in these days of gigabyte real memory configurations. Still no way to force CP to move the pages once they're out there, but we change our spool volumes even less frequently than we do page, and we probably don't really care so much with a small spillover into spool. It's been a while since I've tried that, and I don't remember whether CP tries hard to get that stuff out of the spool areas when more paging space becomes available. Moving spool a whole 'nuther thing - so who cares? Apparently nobody but me. 8-) Ask Neale -- this happens all the time here. He patiently listens to me think out loud. But, a what if that could avoid a full IPL seems interesting... -- db -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
evil thought. Come up without paging space - paging early life will go to spool. Have the autolog machine bring up the real page volumes. Moving spool a whole 'nuther thing - so who cares? David That would certainly prevent any page allocation on paging volumes. Just be aware that if you do need to depopulate the spool volumes that paging overflowed to, you'll have the same problem - the existing techniques for moving spool (SPXTAPE and such) won't work on paging overflowed to spool. Paging on spool space is still paging. But I should mention that the DRAIN statement in the config file can be used to similar effect, on a per volume and per allocation type basis, Wouldn't that be easier? - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On 6/17/11 10:03 AM, Bill Holder hold...@us.ibm.com wrote: That would certainly prevent any page allocation on paging volumes. Just be aware that if you do need to depopulate the spool volumes that paging overflowed to, you'll have the same problem - the existing techniques for moving spool (SPXTAPE and such) won't work on paging overflowed to spool. Paging on spool space is still paging. OTOH, that idea pretty much guarantees that those seldom-referenced pages that get paged out during startup end up on lesser-performance DASD where they're out of the way. Few of us (if any) want that miscellaneous cruft on the really high-performance paging disk, which we'd probably never use for spool space. The renaissance of SSD in the outboard disk boxes may make the prioritization of paging hierarchy important again. (geez. Shades of HPO. I feel all DMKxxx'y again. 8-)) But I should mention that the DRAIN statement in the config file can be used to similar effect, on a per volume and per allocation type basis, Wouldn't that be easier? Would save the steps of listing paging volumes in Offline_at_IPL, VARY ON and ATTACH x to SYSTEM, anyway, particularly with multiple paths to volumes, etc. We can just use START for the individual volumes. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Hi David - one further comment - although it is indeed typical to see some CP owned pages written immediately after the system IPL, it's really not safe to make the assumption that that is the only time it will happen, In particular, the ISFC and Virtual Free Storage spaces are pageable, and page reference activity, and therefore paging behavior, for them can be driven at times other than system IPL. Bill Holder, Senior Software Engineer I'm curious about why CP would write pages immediately after IPL? At that point, there is likely sufficient memory and paging isn't needed.. so why not wait until it is? I'm sure there's a good explanation - just wondering... Scott Rohling -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Hi, People. My cent to this discussion. We are talking about pages on dasd, not pages in memory. So, these pages already was swapped. Except some control pages, I think. To free the dasd, VM will need: Know what pages are on the specific dasd. It can be done by Page out routines, updating one table. Drain the dasd, to prevent new allocation Put on the bit to force a Page In for the pages on that dasd. Returning to normal page process. After some time, no pages are on this dasd. It can be purged from the page tables, removing the control pages. Of course, it is not so simple to implement, but looks like viable. What do you, the experts, think about? __ Clovis From: Scott Rohling scott.rohl...@gmail.com To: LINUX-390@vm.marist.edu Date: 17/06/2011 12:59 Subject: Re: z/VM page space Sent by: Linux on 390 Port LINUX-390@vm.marist.edu Hi David - one further comment - although it is indeed typical to see some CP owned pages written immediately after the system IPL, it's really not safe to make the assumption that that is the only time it will happen, In particular, the ISFC and Virtual Free Storage spaces are pageable, and page reference activity, and therefore paging behavior, for them can be driven at times other than system IPL. Bill Holder, Senior Software Engineer I'm curious about why CP would write pages immediately after IPL? At that point, there is likely sufficient memory and paging isn't needed.. so why not wait until it is? I'm sure there's a good explanation - just wondering... Scott Rohling -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
I'm curious about why CP would write pages immediately after IPL? At that point, there is likely sufficient memory and paging isn't needed.. so why not wait until it is? I'm sure there's a good explanation - just wondering... Scott Rohling I can't really speak for how good the explanation really is, but I know of several cases where we're basically saving for later some data that will only needed in unexpected / exceptional circumstances and that needn't be kept resident. If we were writing such code anew these days, we might instead choose to put such data in a pageable virtual CP utility space, and leave it sitting there, to be paged out only if/when there is actual need (that's exactly how the Virtual Free Storage space I mentioned works) - but there is plenty of code which long predates such pageable CP utility spaces, that page things out explicitly rather than waiting for demand. And as always, if there's old code that isn't broken or getting in the way of new design changes, we're unlikely to change it, even if newer and better methods now exist - the usual cost / benefit / risk trade offs. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
One page must be the minimum. The memory fairies recently visited here so we have 6 systems that are undercommitted for memory at the moment. Each of them has exactly 1 page in use. Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@vm.marist.edu] On Behalf Of Bill Holder Sent: Friday, June 17, 2011 11:20 AM To: LINUX-390@vm.marist.edu Subject: Re: [LINUX-390] z/VM page space I'm curious about why CP would write pages immediately after IPL? At that point, there is likely sufficient memory and paging isn't needed.. so why not wait until it is? I'm sure there's a good explanation - just wondering... Scott Rohling I can't really speak for how good the explanation really is, but I know of several cases where we're basically saving for later some data that will only needed in unexpected / exceptional circumstances and that needn't be kept resident. If we were writing such code anew these days, we might instead choose to put such data in a pageable virtual CP utility space, and leave it sitting there, to be paged out only if/when there is actual need (that's exactly how the Virtual Free Storage space I mentioned works) - but there is plenty of code which long predates such pageable CP utility spaces, that page things out explicitly rather than waiting for demand. And as always, if there's old code that isn't broken or getting in the way of new design changes, we're unlikely to change it, even if newer and better methods now exist - the usual cost / benefit / risk trade offs. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Hi, People. My cent to this discussion. We are talking about pages on dasd, not pages in memory. So, these pages already was swapped. Except some control pages, I think. To free the dasd, VM will need: Know what pages are on the specific dasd. It can be done by Page out routines, updating one table. Drain the dasd, to prevent new allocation Put on the bit to force a Page In for the pages on that dasd. Returning to normal page process. After some time, no pages are on this dasd. It can be purged from the page tables, removing the control pages. Of course, it is not so simple to implement, but looks like viable. What do you, the experts, think about? __ Clovis Hi Clovis - I'm not sure I completely follow what you're proposing, but I believe I already covered it (though maybe I wasn't very clear). What you describe seems similar to the design approach I suggested. The challenge is this: There is is only one way to do your first step Know what pages are on the specific DASD; that is to scan all pages in all pageable spaces using CP's translation tables to determine whether each page is on that DASD. Once all pages on the DASD have been located, moving them is easy enough. Finding them is the hard part. The problem is that there is no way to get from the structures representing the 4K records on the DASD (the paging slots containing the pages) back to the structures which tell us what page is in each slot. All we have to represent each paging slot is one bit (allocated or not) in a bitmap, there is no room for a back pointer to the description of the page that the slot contains. So we have to find the pages the other way - from the top down, scanning all pages in the system to check if they are on the volume in question. This is not a small effort. You might ask why we don't just add that missing back pointer; it seems easy enough. It would indeed be easy enough to code, but the cost in terms of the additional storage usage to hold all of those pointers would be prohibitive - for a maximum sized paging system using ECKD volumes pushed to today's supported limits (about 11TB of paging space), it would mean at least an additional 22GB of non-pageable central storage usage just to hold those back pointers - for a function that is only very rarely used. Using a more real-world example: if you have a 32GB system with a 4:1 overcommit ratio of total virtual to real, with paging configured according to recommendations, those back pointers would take up at least 512MB of non-pageable central storage. It just seems excessively wasteful to spend that much central storage resource representing things that aren't even in storage, especially when it's only needed on a very rare basis. - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On 6/15/11 10:02 PM, O'Brien, Dennis L dennis.l.o'br...@bankofamerica.com wrote: If you have the luxury of an IPL, this is easy. Just put Drain records in SYSTEM CONFIG for the volumes that you don't want pages on. CP won't put anything on them, not even its own pages. The original question was whether there's a way to move the pages later. Yes, I understand that. If you can/are willing to IPL, great, none of this is a concern. The problem that exhibits is that once you're up, there is the situation that there are internal-to-CP pages out there that no amount of virtual machine fiddling can do anything about -- you end up with a volume that has 1 CP page on it, so you can't do anything with that volume except IPL with DRAIN records. There's no way to ask CP to please move that one page so I can get that volume offline without doing a full IPL. If and when Live Guest Relocation comes out, If. If is good. But anyway. IPL is a very big hammer to fix something like this. It's an annoying thing that bugs me, which causes me to pick at it to see if there is a solution short of wait and see. Just thinking out loud. -- db -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
I'm sorry. That is not what we did. We shut down the z/VM LPAR as well. So our linux guests were down as well as z/VM when we re-initialized our page volumes. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Bill Holder Sent: Wednesday, June 15, 2011 4:44 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space We ran into the same issue. Our page space got up to 70% - 80% and we saw performance go down the toilet. Our page volumes were on MOD-9's but only a third of each volume was usable because the page space had been defined as a MOD-3. We tried the DRAIN and that stopped using that page volume but didn't clear it. So we shut down all of our virtual machines and re-initialized the page space on the volumes we were using and now are sitting at 19% - 20% page utilization. Because they were production guests, we had to wait for our Sunday morning window to re-init our page volumes. I was just wondering if there was another way to do what we did. We didn't want to add more volumes, we just needed to fix the ones we were using. Thanks for all of your help and ideas. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 Gene, so are you saying you reformatted live paging volumes while they were still attached to the system? That is not really very safe, even if all of the guests have been logged off - CP itself has some pageable structures that could well be lost, potentially causing a latent abend that would occur at some later time. I suppose if you verified that usage of the volumes in question had really gone completely to zero before re-initializing, that might be safe enough, but it still seems rather dicey - we make no promises about when or why CP might choose to page something out (yes, the overwhelming majority of cases are when storage is full and the system starts paging, which isn't going to happen with all of the guests logged off, but there are a few other oddball cases that could happen when the system is not otherwise paging). - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended recipient(s) or their authorized agents is strictly prohibited. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
evil thought. Come up without paging space - paging early life will go to spool. Have the autolog machine bring up the real page volumes. Moving spool a whole 'nuther thing - so who cares? David Original Message Subject: Re: z/VM page space From: David Boyes dbo...@sinenomine.net Date: Thu, June 16, 2011 9:37 am To: LINUX-390@VM.MARIST.EDU On 6/15/11 10:02 PM, O'Brien, Dennis L dennis.l.o'br...@bankofamerica.com wrote: If you have the luxury of an IPL, this is easy. Just put Drain records in SYSTEM CONFIG for the volumes that you don't want pages on. CP won't put anything on them, not even its own pages. The original question was whether there's a way to move the pages later. Yes, I understand that. If you can/are willing to IPL, great, none of this is a concern. The problem that exhibits is that once you're up, there is the situation that there are internal-to-CP pages out there that no amount of virtual machine fiddling can do anything about -- you end up with a volume that has 1 CP page on it, so you can't do anything with that volume except IPL with DRAIN records. There's no way to ask CP to please move that one page so I can get that volume offline without doing a full IPL. If and when Live Guest Relocation comes out, If. If is good. But anyway. IPL is a very big hammer to fix something like this. It's an annoying thing that bugs me, which causes me to pick at it to see if there is a solution short of wait and see. Just thinking out loud. -- db -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On 6/16/11 7:36 PM, David Kreuter dkreu...@vm-resources.com wrote: evil thought. Come up without paging space - paging early life will go to spool. Have the autolog machine bring up the real page volumes. Hmm. You know, that might just work in these days of gigabyte real memory configurations. Still no way to force CP to move the pages once they're out there, but we change our spool volumes even less frequently than we do page, and we probably don't really care so much with a small spillover into spool. It's been a while since I've tried that, and I don't remember whether CP tries hard to get that stuff out of the spool areas when more paging space becomes available. Moving spool a whole 'nuther thing - so who cares? Apparently nobody but me. 8-) Ask Neale -- this happens all the time here. He patiently listens to me think out loud. But, a what if that could avoid a full IPL seems interesting... -- db -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
We ran into the same issue. Our page space got up to 70% - 80% and we saw performance go down the toilet. Our page volumes were on MOD-9's but only a third of each volume was usable because the page space had been defined as a MOD-3. We tried the DRAIN and that stopped using that page volume but didn't clear it. So we shut down all of our virtual machines and re-initialized the page space on the volumes we were using and now are sitting at 19% - 20% page utilization. Because they were production guests, we had to wait for our Sunday morning window to re-init our page volumes. I was just wondering if there was another way to do what we did. We didn't want to add more volumes, we just needed to fix the ones we were using. Thanks for all of your help and ideas. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 Gene, so are you saying you reformatted live paging volumes while they were still attached to the system? That is not really very safe, even if all of the guests have been logged off - CP itself has some pageable structures that could well be lost, potentially causing a latent abend that would occur at some later time. I suppose if you verified that usage of the volumes in question had really gone completely to zero before re-initializing, that might be safe enough, but it still seems rather dicey - we make no promises about when or why CP might choose to page something out (yes, the overwhelming majority of cases are when storage is full and the system starts paging, which isn't going to happen with all of the guests logged off, but there are a few other oddball cases that could happen when the system is not otherwise paging). - Bill Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On Tue, 14 Jun 2011 11:03:49, David Boyes dbo...@sinenomine.net wrote: Problem is still going to be dealing with pages written during the first few seconds after an IPL by CP itself. Hi David - one further comment - although it is indeed typical to see some CP owned pages written immediately after the system IPL, it's really not safe to make the assumption that that is the only time it will happen, In particular, the ISFC and Virtual Free Storage spaces are pageable, and page reference activity, and therefore paging behavior, for them can be driven at times other than system IPL. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On 6/15/11 5:55 PM, Bill Holder hold...@us.ibm.com wrote: On Tue, 14 Jun 2011 11:03:49, David Boyes dbo...@sinenomine.net wrote: Problem is still going to be dealing with pages written during the first few seconds after an IPL by CP itself. Hi David - one further comment - although it is indeed typical to see some CP owned pages written immediately after the system IPL, it's really not safe to make the assumption that that is the only time it will happen, In particular, the ISFC and Virtual Free Storage spaces are pageable, and page reference activity, and therefore paging behavior, for them can be driven at times other than system IPL. Yeah, I figured as much, although if you could get the system up and drain the volume, then those pages would be ineligible for the drained pack too, right? The just after IPL case is IMHO the hardest case because there's no way to gain control before that can happen (it frequently happens even before OPERATOR is created), so there's really no chance to tinker before you've spattered at least a few pages onto the defined page volumes that pretty much rarely ever get looked at again for the lifetime of that IPL. That's why I think we'd have to identify the pages somehow, and then force CP to do the page-in/page-out sequences internally to actually clear the pages from the pack. Running multiple passes of the various storage page chains could probably be done in userspace by a class B user to get a least a starter set of pages to move, but AFAICT, CP is the only entity that knows where a particular page is being used and could possibly do the deed safely without hitting some weird edge cases. I may look at SNAPDUMP more closely to see if there might be some more ideas there just for the heck of it. In any case, if SSI works, I doubt anyone will ever get anywhere with this, unless I feel like poking around a lot of control blocks for fun, which doesn't seem very likely. It just bugs me in that it doesn't seem like it would be that hard to do, given what's already been done. You're almost there, but, yeah. No business case. Too bad. It's kind of an interesting problem. -- db Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Yeah, I figured as much, although if you could get the system up and drain the volume, then those pages would be ineligible for the drained pack too, right? The just after IPL case is IMHO the hardest case because there's no way to gain control before that can happen (it frequently happens even before OPERATOR is created), so there's really no chance to tinker before you've spattered at least a few pages onto the defined page volumes that pretty much rarely ever get looked at again for the lifetime of that IPL. If you have the luxury of an IPL, this is easy. Just put Drain records in SYSTEM CONFIG for the volumes that you don't want pages on. CP won't put anything on them, not even its own pages. The original question was whether there's a way to move the pages later. If there aren't any CP pages on the volume, you can drain it and start shutting down virtual machines until the page count goes to zero. At that point, it's safe to detach the volume from SYSTEM and format it or whatever. If and when Live Guest Relocation comes out, you could move eligible guests to other systems instead of shutting them down. You don't even have to move all of the guests at once. You can move one guest at a time to another system and then move it back. When it comes back, it won't have any pages on the drained volume. Dennis O'Brien If man does find the solution for world peace it will be the most revolutionary reversal of his record we have ever known. -- General George C. Marshall -- This message w/attachments (message) is intended solely for the use of the intended recipient(s) and may contain information that is privileged, confidential or proprietary. If you are not an intended recipient, please notify the sender, and then please delete and destroy all copies and attachments, and be advised that any review or dissemination of, or the taking of any action in reliance on, the information contained in or attached to this message is prohibited. Unless specifically indicated, this message is not an offer to sell or a solicitation of any investment products or other financial product or service, an official confirmation of any transaction, or an official statement of Sender. Subject to applicable law, Sender may intercept, monitor, review and retain e-communications (EC) traveling through its networks/systems and may produce any such EC to regulators, law enforcement, in litigation and as required by law. The laws of the country of each sender/recipient may impact the handling of EC, and EC may be archived, supervised and produced in countries other than the country in which you are located. This message cannot be guaranteed to be secure or free of errors or viruses. References to Sender are references to any subsidiary of Bank of America Corporation. Securities and Insurance Products: * Are Not FDIC Insured * Are Not Bank Guaranteed * May Lose Value * Are Not a Bank Deposit * Are Not a Condition to Any Banking Service or Activity * Are Not Insured by Any Federal Government Agency. Attachments that are part of this EC may have additional important disclosures and disclaimers, which you should read. This message is subject to terms available at the following link: http://www.bankofamerica.com/emaildisclaimer. By messaging with Sender you consent to the foregoing. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended recipient(s) or their authorized agents is strictly prohibited. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Performance Toolkit can show you the paging load for users (menu option 22 - FCX113).. how much in XSTORE, how much on DASD ... but it doesn't identify the paging volume(s) the pages are stored on, if that's what you want to do. Scott Rohling On Tue, Jun 14, 2011 at 7:03 AM, Nelson, Gene C. gcnel...@fedins.comwrote: In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended recipient(s) or their authorized agents is strictly prohibited. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? I am unaware of any way to tell which users (or other pageable entities) have their pages on a given paging volume in z/VM. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Friends, Please, take a look at SPOOLCHN http://www.vm.ibm.com/download/packages/ Follow the announce: (c) Copyright International Business Machines Corporation 1993,2005 All Rights Reserved. *-* *Name: SPOOLCHN, version 5.C * * * * Description: Extensions to spool Query commands * * * * Author: Richard Ross (rich...@us.ibm.com) * * * *Date: August, 2005 * *-* SPOOLCHN is a VM system programmer utility (class C or E) which will display files in the spool system. SPOOLCHN has the following advantages over the standard spool Query commands: - can show spool usage (blocks of spool data) - shows more information than the spool Query commands - output can be directed to terminal, stack, disk, or variables in REXX - output can include an exec for manipulating the spool files - more search criterea than spool Query, such as number of records, age of file, etc. - wildcard searches allowed --- SPOOLCHN has been tested on all current releases of VM. --- SPOOLCHN is used by a privileged user to query files in the spool system. Unlike QUERY RDR, QUERY PRT, or QUERY PUN, SPOOLCHN will show spool files that are open (as these files do take up space in the spool system). It will also not tie up system resources the way that a QUERY RDR ALL will. SPOOLCHN requires class C or E to display the spool file blocks in real storage (DIAGNOSE 4). It also requires class D if the DIAGNOSE D8 option is used. --- SPOOLCHN is available from VMTOOLS and http://www.vm.ibm.com Richard Ross _ __ Clovis From: Scott Rohling scott.rohl...@gmail.com To: LINUX-390@vm.marist.edu Date: 14/06/2011 10:32 Subject: Re: z/VM page space Sent by: Linux on 390 Port LINUX-390@vm.marist.edu Performance Toolkit can show you the paging load for users (menu option 22 - FCX113).. how much in XSTORE, how much on DASD ... but it doesn't identify the paging volume(s) the pages are stored on, if that's what you want to do. Scott Rohling On Tue, Jun 14, 2011 at 7:03 AM, Nelson, Gene C. gcnel...@fedins.comwrote: In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution
Re: z/VM page space
Alan, Do you think IBM will think about adding a 'page delete' feature so you can migrate page volumes to another disk subsystem so you don't have to shutdown all of the guest and z/VM? It sure would be handy. I was hoping to move all of the paging to the new devices so I would have less to do on the weekend when I move the z/VM res/spool/page/guest volumes. Sam Bass -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
The question is why do you need to know? If you are running out of space, add more. If you can't, you'll have to shut virtual machines down. Start with the largest ones. BTW, don't let your paging space get more than 50% full or performance goes down the toilet. Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@vm.marist.edu] On Behalf Of Bill Holder Sent: Tuesday, June 14, 2011 7:44 AM To: LINUX-390@vm.marist.edu Subject: Re: [LINUX-390] z/VM page space In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? I am unaware of any way to tell which users (or other pageable entities) have their pages on a given paging volume in z/VM. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
This is for SPOOL -- not PAGE! Completely different allocation types, etc.. Scott Rohling On Tue, Jun 14, 2011 at 9:19 AM, Clovis Pereira gclo...@br.ibm.com wrote: Friends, Please, take a look at SPOOLCHN http://www.vm.ibm.com/download/packages/ Follow the announce: (c) Copyright International Business Machines Corporation 1993,2005 All Rights Reserved. *-* *Name: SPOOLCHN, version 5.C * * * * Description: Extensions to spool Query commands * * * * Author: Richard Ross (rich...@us.ibm.com) * * * *Date: August, 2005 * *-* SPOOLCHN is a VM system programmer utility (class C or E) which will display files in the spool system. SPOOLCHN has the following advantages over the standard spool Query commands: - can show spool usage (blocks of spool data) - shows more information than the spool Query commands - output can be directed to terminal, stack, disk, or variables in REXX - output can include an exec for manipulating the spool files - more search criterea than spool Query, such as number of records, age of file, etc. - wildcard searches allowed --- SPOOLCHN has been tested on all current releases of VM. --- SPOOLCHN is used by a privileged user to query files in the spool system. Unlike QUERY RDR, QUERY PRT, or QUERY PUN, SPOOLCHN will show spool files that are open (as these files do take up space in the spool system). It will also not tie up system resources the way that a QUERY RDR ALL will. SPOOLCHN requires class C or E to display the spool file blocks in real storage (DIAGNOSE 4). It also requires class D if the DIAGNOSE D8 option is used. --- SPOOLCHN is available from VMTOOLS and http://www.vm.ibm.com Richard Ross _ __ Clovis From: Scott Rohling scott.rohl...@gmail.com To: LINUX-390@vm.marist.edu Date: 14/06/2011 10:32 Subject: Re: z/VM page space Sent by: Linux on 390 Port LINUX-390@vm.marist.edu Performance Toolkit can show you the paging load for users (menu option 22 - FCX113).. how much in XSTORE, how much on DASD ... but it doesn't identify the paging volume(s) the pages are stored on, if that's what you want to do. Scott Rohling On Tue, Jun 14, 2011 at 7:03 AM, Nelson, Gene C. gcnel...@fedins.comwrote: In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential
Re: z/VM page space
We ran into the same issue. Our page space got up to 70% - 80% and we saw performance go down the toilet. Our page volumes were on MOD-9's but only a third of each volume was usable because the page space had been defined as a MOD-3. We tried the DRAIN and that stopped using that page volume but didn't clear it. So we shut down all of our virtual machines and re-initialized the page space on the volumes we were using and now are sitting at 19% - 20% page utilization. Because they were production guests, we had to wait for our Sunday morning window to re-init our page volumes. I was just wondering if there was another way to do what we did. We didn't want to add more volumes, we just needed to fix the ones we were using. Thanks for all of your help and ideas. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Marcy Cortes Sent: Tuesday, June 14, 2011 9:54 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space The question is why do you need to know? If you are running out of space, add more. If you can't, you'll have to shut virtual machines down. Start with the largest ones. BTW, don't let your paging space get more than 50% full or performance goes down the toilet. Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@vm.marist.edu] On Behalf Of Bill Holder Sent: Tuesday, June 14, 2011 7:44 AM To: LINUX-390@vm.marist.edu Subject: Re: [LINUX-390] z/VM page space In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? I am unaware of any way to tell which users (or other pageable entities) have their pages on a given paging volume in z/VM. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended recipient(s) or their authorized agents is strictly prohibited. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Maybe I'm not understanding - but if you will be moving res/spool/page/guest -- then what's on page doesn't really matter -- it's cleared between IPL's and you will be IPLing off a new res pack .. no? PAGE and TEMP you normally don't worry about.. just make sure the new volumes are formatted/allocated correctly. Scott Rohling On Tue, Jun 14, 2011 at 8:48 AM, Sam Bass sam.b...@mclaneco.com wrote: Alan, Do you think IBM will think about adding a 'page delete' feature so you can migrate page volumes to another disk subsystem so you don't have to shutdown all of the guest and z/VM? It sure would be handy. I was hoping to move all of the paging to the new devices so I would have less to do on the weekend when I move the z/VM res/spool/page/guest volumes. Sam Bass -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Alan Altmark Sent: Monday, June 13, 2011 6:02 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On 6/14/11 10:54 AM, Marcy Cortes marcy.d.cor...@wellsfargo.com wrote: The question is why do you need to know? Short version: to know which guests to recycle to clear an old paging volume without having to take a CP IPL just to get old pages off a paging volume. Problem is still going to be dealing with pages written during the first few seconds after an IPL by CP itself. I suspect that it would be possible to tweak the page device list in real storage to remove the volume in question, run the CP page list in real core and force a page in/page out sequence for pages on the volume in question, but There Lie Big Nasty Dragons. There's a lot of interlocking pieces and CP integrity is at stake, so it would not be simple code. Still, CP DRAIN cuu PAGE MIGRATE sounds like a requirement to me. Its probably going to be necessary to get the clustering stuff to behave well, although I suspect breaking one more thing in that rewrite is probably going to cause riots and much wailing in Endicott. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Nope - issuing a DRAIN is really the only option. In my experience - it almost always requires an IPL to completely get it clear. DRAIN does not actively move pages -- it simply prevents any CP from writing any new ones on it. It's pretty standard to have to wait until the next maintenance window to reclaim drained volumes. Scott Rohling On Tue, Jun 14, 2011 at 9:56 AM, Nelson, Gene C. gcnel...@fedins.comwrote: We ran into the same issue. Our page space got up to 70% - 80% and we saw performance go down the toilet. Our page volumes were on MOD-9's but only a third of each volume was usable because the page space had been defined as a MOD-3. We tried the DRAIN and that stopped using that page volume but didn't clear it. So we shut down all of our virtual machines and re-initialized the page space on the volumes we were using and now are sitting at 19% - 20% page utilization. Because they were production guests, we had to wait for our Sunday morning window to re-init our page volumes. I was just wondering if there was another way to do what we did. We didn't want to add more volumes, we just needed to fix the ones we were using. Thanks for all of your help and ideas. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Marcy Cortes Sent: Tuesday, June 14, 2011 9:54 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space The question is why do you need to know? If you are running out of space, add more. If you can't, you'll have to shut virtual machines down. Start with the largest ones. BTW, don't let your paging space get more than 50% full or performance goes down the toilet. Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@vm.marist.edu] On Behalf Of Bill Holder Sent: Tuesday, June 14, 2011 7:44 AM To: LINUX-390@vm.marist.edu Subject: Re: [LINUX-390] z/VM page space In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? I am unaware of any way to tell which users (or other pageable entities) have their pages on a given paging volume in z/VM. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended recipient(s) or their authorized agents is strictly prohibited. Thank you. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Dear all, I would like to support the idea of having a command that actively empties a used page volume. Well, I don't really need to see which guest is using which page volume but in case I am draining a volume there should be an option to move the allocated pages to an active volume. Last year, we were replacing some of our DS8000s with a DS8700s and so I had to move nearly all VM-volumes. Moving the guest and user volumes and dynamically change disk addresses etc. was not a big deal. For our clusters I didn't had any downtime since i could recycle one node after the other. Very dynamic process. However I could not move some of the page volumes because there were pages in use. And that required then an IPL. For empty page volumes the drain is good but for active volumes does not help really. Kind regards, Florian On Tue, Jun 14, 2011 at 6:19 PM, Scott Rohling scott.rohl...@gmail.comwrote: Nope - issuing a DRAIN is really the only option. In my experience - it almost always requires an IPL to completely get it clear. DRAIN does not actively move pages -- it simply prevents any CP from writing any new ones on it. It's pretty standard to have to wait until the next maintenance window to reclaim drained volumes. Scott Rohling On Tue, Jun 14, 2011 at 9:56 AM, Nelson, Gene C. gcnel...@fedins.com wrote: We ran into the same issue. Our page space got up to 70% - 80% and we saw performance go down the toilet. Our page volumes were on MOD-9's but only a third of each volume was usable because the page space had been defined as a MOD-3. We tried the DRAIN and that stopped using that page volume but didn't clear it. So we shut down all of our virtual machines and re-initialized the page space on the volumes we were using and now are sitting at 19% - 20% page utilization. Because they were production guests, we had to wait for our Sunday morning window to re-init our page volumes. I was just wondering if there was another way to do what we did. We didn't want to add more volumes, we just needed to fix the ones we were using. Thanks for all of your help and ideas. Gene Nelson Federated Insurance Company 121 E. Park Square Owatonna, MN 55060 (507) 455-5200 ext. 4555706 -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Marcy Cortes Sent: Tuesday, June 14, 2011 9:54 AM To: LINUX-390@VM.MARIST.EDU Subject: Re: z/VM page space The question is why do you need to know? If you are running out of space, add more. If you can't, you'll have to shut virtual machines down. Start with the largest ones. BTW, don't let your paging space get more than 50% full or performance goes down the toilet. Marcy -Original Message- From: Linux on 390 Port [mailto:LINUX-390@vm.marist.edu] On Behalf Of Bill Holder Sent: Tuesday, June 14, 2011 7:44 AM To: LINUX-390@vm.marist.edu Subject: Re: [LINUX-390] z/VM page space In z/VM is there a way to tell who or what is using that page space? Coming from the z/OS world I know there was a way to display show was the biggest user of the page slots. Is there a way to do that in z/VM? I am unaware of any way to tell which users (or other pageable entities) have their pages on a given paging volume in z/VM. Bill Holder, Senior Software Engineer IBM z/VM Development, Memory Management, Endicott, NY Phone: 607-429-3640IBM TieLine: 620-3640 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ The information contained in this e-mail message is intended only for the personal and confidential use of the designated recipient(s) named above. This message may be an attorney-client or work product communication which is privileged and confidential. It may also contain protected health information that is protected by federal law. If you have received this communication in error, please notify us immediately by telephone and destroy (shred) the original message and all attachments. Any review, dissemination, distribution or copying of this message by any person other than the intended
Re: z/VM page space
On Tuesday, 06/14/2011 at 11:36 EDT, Sam Bass sam.b...@mclaneco.com wrote: Do you think IBM will think about adding a 'page delete' feature so you can migrate page volumes to another disk subsystem so you don't have to shutdown all of the guest and z/VM? It sure would be handy. I was hoping to move all of the paging to the new devices so I would have less to do on the weekend when I move the z/VM res/spool/page/guest volumes. Not unless folks begin to get louder in their requests for such a function. I can tell you than handiness is not a metric IBM uses to decide such things. Looking at the Single System Image environment, you would relocate the (important) guests to another LPAR, re-IPL this LPAR, and then relocate them back. So I suppose you'd have to have a compelling argument as to why SSI wouldn't be acceptable. Same problem with spool, btw, though at least you can use SPOOLCHN and SPXTAPE to fix it. BTW, I have my own closet full of various T-shirts, and I have one labeled Had to re-IPL to reclaim DASD. :-( Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Thought - and probably did this years ago - I will page it back in - come up with one page volume and necessary spool at IPL. Then have AUTOLOG1/2 bring up the other page volumes before the massive guest startup. Quasi dynamic. Just keep your spool slots intact! ha ha unless you like CLEAN start. It is not worth the bother but there it is, a thought. David Original Message Subject: Re: z/VM page space From: David Boyes dbo...@sinenomine.net Date: Tue, June 14, 2011 12:03 pm To: LINUX-390@VM.MARIST.EDU On 6/14/11 10:54 AM, Marcy Cortes marcy.d.cor...@wellsfargo.com wrote: The question is why do you need to know? Short version: to know which guests to recycle to clear an old paging volume without having to take a CP IPL just to get old pages off a paging volume. Problem is still going to be dealing with pages written during the first few seconds after an IPL by CP itself. I suspect that it would be possible to tweak the page device list in real storage to remove the volume in question, run the CP page list in real core and force a page in/page out sequence for pages on the volume in question, but There Lie Big Nasty Dragons. There's a lot of interlocking pieces and CP integrity is at stake, so it would not be simple code. Still, CP DRAIN cuu PAGE MIGRATE sounds like a requirement to me. Its probably going to be necessary to get the clustering stuff to behave well, although I suspect breaking one more thing in that rewrite is probably going to cause riots and much wailing in Endicott. -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
Sam, You can take a look at the CP DRAIN command for DISK. You can tell CP to stop using a device this way. Aria -Original Message- From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Sam Bass Sent: Monday, June 13, 2011 3:47 PM To: LINUX-390@VM.MARIST.EDU Subject: z/VM page space I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Sam Bass -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/ -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/
Re: z/VM page space
On Monday, 06/13/2011 at 03:48 EDT, Sam Bass sam.b...@mclaneco.com wrote: I am migrating from one disk subsystem to another. I know that you can add PAGE volumes via DEF CPOWNED after you have formatted a volume and ATT *unit* SYSTEM. Is there a way to do a 'page delete' like you can in z/OS so you can move off of the old disk volumes? Unfortunately, no. You can tell CP to stop using (adding new data to) a paging or spooling volume, but you cannot make CP migrate data off of the drained volume onto an active one. Alan Altmark z/VM and Linux on System z 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 -- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 -- For more information on Linux on System z, visit http://wiki.linuxvm.org/