DCSS as SWAP disk for z/Linux

[James G. Stracka]

I am sure that most people SWAP to V-DISK because it is easier, not
because it is better, fast, cheaper, but they are familiar with disk
and it is less complex than setting up swapping to DCSS.

I think it is intellectually foolish to not investigate using DCSSes
for SWAP.

That brings up another question,
"Is it feasible to load the z/Linux kernel into an NSS?"

Believe me, I do not speak penguin. Years ago, at a company far far
away, we loaded our only VSE system into an NSS. This was to run VTAM
for VM (was that VCNA pre-HPO 5?).

Can running from an NSS be done for a few penguins? Is it then
practical to do it for a penguin farm?

Just asking,
Jim

[Adam Thornton]

Post by James G. Stracka
Can running from an NSS be done for a few penguins? Is it then
practical to do it for a penguin farm?

Yes. You save a few megabytes per image.

It makes maintenance more difficult, though.

Adam

[Anne & Lynn Wheeler]

Post by Adam Thornton
Yes. You save a few megabytes per image.
It makes maintenance more difficult, though.

in the original virtual memory mapped paged-map implementation (that
predated DCSS, vm/370 release 3 just picked up a small subset), the
source of the system was on standard virtual machine accessed disk
that was accessed using page mapped semantics. infrastructure layered
on top could be standard filesystem semantics ... and the object
loaded could be managed with standard filesystem operation (the load
command specified both the mapping between paged mapped disk and
virtual address space as well as any sharing options ... with various
kinds of integrity rules applied).

the issue for cms then easily becomes having multiply managed kernel
image ... you create a new kernel image at a new location and update
the boot information to point to the latest kernel image location.
people booting cms after the boot pointer change get the new image,
people that had the earlier kernel image have the previous kernel
image.
http://www.garlic.com/~lynn/subtopic.html#mmap

in theory, linux kernel images and maint. could be handled in a
similar way ... just create a new boot image in a new location.

translated to conventional dcss methodology ... this requires
multiple, similarly named (say linux01-linux-05), systems. maint
process cycles thru the different names.

virtual machines then are modified to load a named system ... say
linuxredirect ... which is just a trivial piece of redirection code
which does the actual loading of named systems. the redirection code
is updated after some maint. process and the appropriate images have
been saved to available place.

[Rick Troth]

Post by Anne & Lynn Wheeler
in the original virtual memory mapped paged-map implementation (that
predated DCSS, vm/370 release 3 just picked up a small subset), the
source of the system was on standard virtual machine accessed disk
that was accessed using page mapped semantics. infrastructure layered
on top could be standard filesystem semantics ...

Sounds a lot like XIP, which is based on EXT2.
Define a partition of the same size as the intended DCSS
(or slightly less), put your stuff there, unmount it,
use 'dd' or something like it to grab a snap-shot, and store.

Post by Anne & Lynn Wheeler

Post by James G. Stracka
Can running from an NSS be done for a few penguins?
Is it then practical to do it for a penguin farm?

Yes. You save a few megabytes per image.
It makes maintenance more difficult, though.

I don't see it as any more difficult than re-saving CMS.
Kernel into an NSS. Possibly one or more filesystems into a DCSS.
Maybe swap to DCSS too!

-- R;

[Anne & Lynn Wheeler]

Post by Rick Troth
Sounds a lot like XIP, which is based on EXT2.
Define a partition of the same size as the intended DCSS
(or slightly less), put your stuff there, unmount it,
use 'dd' or something like it to grab a snap-shot, and store.

there was science center
http://www.garlic.com/~lynn/subtopic.html#545tech

report done on the initial port of vmm from cp67 to vm370

L. Wheeler, CSC VM/370 Extended II: Virtual Memory Management, IBM
Science Center Report ZZ20-6002, July 1974, 19 pp., shared segments,
migration.

the dcss subset of vmm
http://www.garlic.com/~lynn/subtopic.html#mmap

was then done for vm/370 release 3, 30 years ago.

vmm also included a lot of work on address independent
code:
http://www.garlic.com/~lynn/subtopic.html#adcon

mmap support allowed the use of the paging kernel support to directly
read/write any standard disk area. it eliminated a lot of the overhead
of simulating virtual i/o (among other things).

a little drift, i included some part of "migration" in the resource
manager.
http://www.garlic.com/~lynn/2001e.html#45 VM/370 Resource Manager

from above (taken from the "blue" letter):

Enhancements are made to VM/370 in these areas:

Scheduling ALgorithm. A fair share approach to distribute the
resources of the system equally among the users with improved
interactive performance on trivial commands.

Page Migration. Designation of preferred paging areas on DASD with
migration to other devices is based on how long the pages are unused.

Swaptable Migration. Seldom used segment tables are swapped to DASD
thereby freeing up main storage.

Reset Pages and Time Stamp Segments. The working set algorithm
improves page selection, while time stamping facilitates page
migration and swaptable migration.

Working Set Estimate. Dynamically adjusted multiprogramming levels are
achieved by periodic evaluation of total system performance based on
feedback control.

Fast Redispatch Extension. The number of cases where fast redispatch
implementation is used after privileged instruction simulation and I/O
interruptions are increased.

Enable Window. It increases the extent to which VM/370 runs enabled and
thus can accept I/O and external interruptions

Set Favored Extension. The specification of multiple users with the
set favored percent option is provided.

Indicate Command Extension. Additional performance status data is made
available to the systems performance evaluation routine.

Selective Path Length Reductions.

[Anne & Lynn Wheeler]

The following is from a presentation made on 86.10.07 at the SEAS
meeting on Jersey. List of SEAS meetings:
http://www.daube.ch/share/seas01.html#places

It describes a series of test on production VM/370 release 6 system
comparing PAM/CDF, PAM/EDF, and 4k/EDF. Lots of past PAM references
http://www.garlic.com/~lynn/subtopic.html#pam

Comparison of PAM, CDF and EDF

pam/cdf pam/edf 4k/edf
runl 2.89/39l0/33. 3.28/4103/4l. 3 72/1978/95.
run2 2.90/3768/49. 3.32/4ll5/54. 3.65/1965/79.
run3 3.01/3749/47. 3.35/4l23/8l. 3.66/1932/77.
runq 2.94/4033/34. 3.46/3975/46. 3.66/1920/63.
run5 3.0l/3472/36. 3.62/3884/68. 3.75/1948/85.
run6 3.08/3776/70. 3.46/3784/60. 3.79/1986/84.
run7 3.l3/3740/57. 3.37/3979/46. 3.75/1977/84
run8 2.99/3868/87. 3.4l/3836/52.. 3.77/1958/88.

system CPU I/O elapsed
avg. avg. avg.
4k/EDF 3.72 1958 81.
PAM/EDF 3.4l 3836 56.
PAM/CDF 3.02 3790 52.

min, min. min
4k/EDF 3.66 1920 63.
PAM/EDF 3.28 3784 41.
PAM/CDF 2.89 3472 33.

4k/EDF is the EDF CMS filesystem introduced in VM/370 release 6 used
with 4k block option.

PAM/CDF is the original CMS filesystem modified to use 4K records
rather than 800 and page-map interface.

PAM/EDF is the EDF CMS filesystem modified to use the page-map interface.

CPU is total (combination of virtual and supervisor).

The test was run on standard production system with other activity. A
run consisted of doing the same exact operations involving pam/cdf
filesystem, then pam/edf filesystem, then 4k/edf filesystem. This was
repeated eight times and the avg. values and the minimum values used.

There is not a direct comparison between 4k/edf i/o and pam i/o. For
4k/edf, I/O is reported as the number of virtual i/o operations
(independent of number of records transferred). There is slight
variability in 4k/edf i/o based on state of the filesystem and the
number of contiguous records that might be involved for operation.
For pam i/o, reported is the number of 4k page transfers (which might
also include other paging operations for the virtual machine during
the period and doesn't directly indicate the number of physical i/o
operations).

Both PAM test performed better than 4k/edf because of 1) reduced CP
overhead for supporting the operation, 2) better CP logic for chaining
multiple 4k transfers in single I/O,

The minimum values are going to be closest to base operation excluding
interferance from other workload on the system. pam/cdf is nearly
twice as fast and approx. 1/4 less cpu for the workload compared to
the same workload running on 4k/edf filesystem.

When the science center
http://www.garlic.com/~lynn/subtopic.html#545tech

eventually replaced its 360/67 with 370/155, there was a port of a lot
of local modifications from cp67 to vm370. another page from the same
SEAS presentation

Release 2.15 CSC/VM

* Relocatable (floating) Shared Segments
* Paging Access Method (PAM)
* CP/67 Feedback controls
* CP/67 Working Set controls
* CP/67 GLobal LRU Page Replacement
* CP/67 Fastpath
* Restructured Page Supervisor
* Page and Swaptable Migration
* Q3
* Scheduling based on resource objectives
* Resource objectives either fixed or fairshare

note that much of this, I subsequently released in the resource
manager ... blue letter reference in recent posting
http://www.garlic.com/~lynn/2006.html#19 DCSS as SWAP disk for z/Linux

a small subset of the cms & cp relocatable (floating) shared segments
http://www.garlic.com/~lynn/subtopic.html#adcon

(w/o paged-map support) had been released as DCSS (prior to the
resource manager release).

a couple other posts in this thread:
http://www.garlic.com/~lynn/2006.html#17 {SPAM?} DCSS as SWAP disk for z/Linux
http://www.garlic.com/~lynn/2006.html#18 DCSS as SWAP disk for z/Linux

[Anne & Lynn Wheeler]

Post by Anne & Lynn Wheeler
The following is from a presentation made on 86.10.07 at the SEAS
http://www.daube.ch/share/seas01.html#places

re:
http://www.garlic.com/~lynn/2006.html#26 DCSS as SWAP disk for z/Linux

i couldn't find softcopy for my presentation ... so last night, i
scanned the hardcopy. the pdf isn't bad, but the ocr processing was
only about 80-90percent correct, i had to manually correct 10-20
percent of it.

I did find softcopy of the vm project agenda.

my talk was suppose to be on the history of vm performance ... dating
back to spring of '68 when i was an undergraduate (38 years now). I
had an hour and 60+ foils. i managed to only get thru the first couple
foils ... so we scheduled a bof during scids that night ... and i
talked from 6pm to midnight (room off scids, with interruptions for
refreshments).

SEAS OCT 6-10 1986, JERSEY, VM PROJECT.
_______________________________________

*----------------------------------------------------------*
| Monday 6th |1400-1445 | Introduction |
| October | and | to the VM project |
| |1450-1535 | Steve Bouch BT |
| | | (project chairman) |
| | 1.4V | Show and Tell |
| | and | VM-related announcements since |
| | 1.5V | SM86 |
| | | Adrian Walmsley, IBM UK |
| | | VM/SP Entry User experiences |
| | | Vlady Priplata |
| |----------+----------------------------------|
| |1600-1645 | XA I/O Architecture Overview |
| | and | functions used by VM and MVS |
| |1650-1735 | IOCP and sysgen |
| | | Adrian Walmsley, IBM UK |
| | 1.6V | |
| | and | VM XA SF preferred guest planning|
| | 1.7V | configuration, performance tools|
| | | recovery |
| | | Gisela Lotringer, |
| | | IBM South Africa |
| | | (joint with MVS project) |
*----------------------------------------------------------*

*----------------------------------------------------------*
| Tuesday 7th|1400-1445 | IBM announcements |
| October | and | Bob Knaus IBM Endicott |
| |1450-1535 | CMS Update |
| |2.4V, 2.5V| Dave Getson IBM Endicott |
| |----------+----------------------------------|
| |1600-1645 | CMS Free Storage Management |
| | 2.6V | Stuart McRae |
| | | Systems and Telecomms, UK |
| |----------+----------------------------------|
| |1650-1735 | Using 7171s |
| | 2.7V | nn, User |
*----------------------------------------------------------*

SEAS Oct 6-10 1986, Jersey, VM Project. 1

*----------------------------------------------------------*
| Weds 8th |1400-1445 | Electronic Mail |
| October | | Stuart McRae |
| | 3.4V | Systems and Telecomms |
| | | Dave Barker |
| | | Rutherford Lab (?) |
| | | (joint with OA project) |
| |----------+----------------------------------|
| |1450-1535 | VM SECURE User experiences |
| | | Graham Taplin, |
| | | Watson Calculating |
| | 3.5V | (joint with IM project) |
| |----------+----------------------------------|
| |1660-1645 | Page mode printing on VM |
| | | Mike Kay |
| | 3.6V | IBM Almaden Research |
| |----------+----------------------------------|
| |1650-1735 | Business Session |
| | 3.7V | including IBM responses to |
| | | resolutions |
| | | Bob Knaus IBM Endicott |
*----------------------------------------------------------*

*----------------------------------------------------------*
| Thurs 9th |1400-1445 | VM VTAM User Experiences |
| October | 4.4V | John Wilshaw, Metier |
| |----------+----------------------------------|
| |1450-1535 | Automating Operations in VM |
| | 4.5V | nn, User (BT ?) |
| |----------+----------------------------------|
| |1600-1645 | Tape staging |
| | 4.6V | nn, CERN, Geneva |
| |----------+----------------------------------|
| |1650-1735 | Highly available VM systems |
| | 4.7V | Al Griefer, IBM Almaden Res. |
*----------------------------------------------------------*

SEAS Oct 6-10 1986, Jersey, VM Project. 2

*----------------------------------------------------------*
| VM New Users sessions |
|----------------------------------------------------------|
| | | |
| Monday 6th |1600-1645 | Introduction |
| October | and | to REXX |
| |1650-1735 | Jeff Silcock |
| |1.6Y,1.7Y | Scottish College of Textiles|
|------------+----------+----------------------------------|
| Tuesday 7th|1600-1645 | Faster VM with no mods |
| October | 2.6Y | nn, User |
| |----------+----------------------------------|
| |1650-1735 | Introduction to using VMAP |
| | | Per Kristensen, |
| | 2.7Y | IBM UK |
*----------------------------------------------------------*

*----------------------------------------------------------*
| General Sessions proposed by VM project |
|----------------------------------------------------------|
| | | TOOLS - a Disk and Conference |
| Tuesday | 2.22 | Manager |
| 7th October| | Mike Cowlishaw, |
| | | IBM UK Scientific Centre |
| | | Winchester |
|------------+----------+----------------------------------|
| Tuesday | 2.32 | History of VM performance |
| 7th October| | Lynn Wheeler, |
| | | IBM Almaden Research |
*----------------------------------------------------------*

SEAS Oct 6-10 1986, Jersey, VM Project. 3

--
Anne & Lynn Wheeler | http://www.garlic.com/~lynn/