| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 61.1 | Early exposure to internal mechanisms! Good! | NITTY::THORNE | Department of Redundancy Department | Wed Apr 24 1991 10:58 | 16 |
| Having only seen the course specification's list of topics, I was still
glad to see the section titled 'Additional VMS Concepts'. If this is
written well, it provides, I believe, a necessary early exposure to the
fundamental structures used to assess and manipulate system
performance, etc.
On the other hand, I see earlier in the module what appears to be a
transplant of sections straight from the Network Mgr I course of the
first module. Students' eyes tend to roll back in their heads when
this module is covered -- mainly because most of it is pretty obvious.
While I agree that the components of network management should be
lightly covered here, I think the students' perception of material that
is covered too lightly is that it's pablum. I hope this section has
been rewritten to provide a bit more substance.
Mark Thorne
|
| 61.2 | SYSNET I Chapter 10 Views | DLO10::TARLING | | Fri May 10 1991 10:52 | 31 |
| Howdy;
1. Page 1-5a, typo/Document problem, "willbe".
2. Page 1-9, What is a "DSSI bus"? A definition would be helpful.
3. Page 1-13, Artwork jamed into title at top of page.
4. Page 1-16, Satellites 2nd bullet, Perhaps this could read:
"May not have local system disks"?
5. Page 1-18a, 1st bullet, "Two or more processors" could be confusing.
Perhaps we could use "Two or more nodes".?
6. Page 1-18, 1st bullet, again I prefer nodes to processors.
7. Page 1-20, This may be the "pablum" that Mark refers to in .1. I
vote to leave material of this type (that whic is intuitively
obvious to the most casual of observers) out.
8. Page 1-23 to 1-29, 1-37 to 1-39, There is a real opportunity to
develope a sound conceptual understanding of basic VMS principles.
I would like to see some of the materials from the "Concepts"
course, even if it means sliding some material to SYSNET N.
9. Page 1-39, 1st bullet, It is entirely possible to have several
hundred processes without "any" swapping, and "very little"
paging.
Arnold.
|
| 61.3 | View from the UK | UKEDU::SHONE | Keith Shone @RKA 830-4074 | Fri May 10 1991 13:12 | 138 |
| Module: A Closer Look at the VMS Environment
My views - not necessarily those of my colleagues.
Some of these comments I've made elsewhere - I think.
I liked the content and depth of this module. Sure there's plenty of
it but as a non-cistern mangler I appreciated the broad cover.
Page Comment
1-5a Line 3: typo willbe -> will be
Line 3: typo out -> put
Line 3: typo inperspective -> in perspective
(I know they're not strictly typos but they're easier to
explain).
1-5 Line 2: first word ALthough -> Although
Bullet 5: typo Udate -> Update
This bullet might read better as:
o Update, Install amd Maintain system software
(the printed version has tense changes).
1-7a The SHOW CPU command needs priv (I found CMKRNL works).
1-8 Bullet 6: the word members is an unfortunate choice of
word. I read this differently from that intended!
How about "...on member nodes." ?
1-10a Line 8: "...connected to on system..." ->
"...connected to a system..."
Line 9: "...until it releases the device for reallocation."
might read better as:
"...until released for reallocation."
1-12 Couldn't tie the first bullet notes with the Figure.
The two HSC controllers don't appear to have two dual-ported
disks connected to them (I'm probably mis-reading the
figure...)
1-13 Figure needs repositioning away from its heading
1-16a <Nit> Line 2: Ethernet (capital E)
1-17a Bullet 3: typo intial -> initial
1-23 Line 3: bit of purist English! -
none means not one and is followed by is not are. Thus
"...although none of them is a system task."
1-27 Penultimate bullet: I thought NETSERVER.LOG files
were only written when a login occurred on the remote
node. For example, if I was in Bedford, Mass. and
wished to copy a file from UKEDU I would use a COPY
command thus:
$ COPY /LOG UKEDU"shone youmustbejoking"::LOGIN.COM *
If the directory and file were world readable I could
simply:
$ COPY /LOG LOGIN.COM *
(Actually I could login to UKEDU and MAIL it to my
account at Bedford but...)
1-28a Last line: typo indictaed -> indicated
1-28 Bullet 2: <Nit> Subprocess -> subprocess
1-29 line 2: add word "to" before "create"
1-30 Bullet 2: I know "usually do not" is accurate but
terminal output from a subprocess certainly affects other
processes in the job
Bullet 4: PID - has this been defined yet?
1-31 Flag waving - liked that, can we have more?
1-32a Footnote: typo untill -> until
1-32 Example 1-2: The SHOW SYSTEM command is displaced right
Callout 3: typo subprocces -> subprocess
I've a display with a little more going on if you want to
use it:
VAX/VMS V5.4-2 on node UKEDU 10-MAY-1991 14:33:46.05 Uptime 5 02:59:24
Pid Process Name State Pri I/O CPU Page flts Ph.Mem
00000041 SWAPPER HIB 16 0 0 00:02:22.59 0 0
00001EC2 SHONE_1 CUR 4 65 0 00:00:01.00 298 283 S
000017C4 Jan LEF 9 33501 0 00:04:02.31 7739 495
00000045 CONFIGURE HIBO 8 -- swapped out -- 142
00000047 ERRFMT HIB 8 4705 0 00:00:50.75 380 68
00000048 OPCOM HIB 8 2516 0 00:00:45.40 611 141
00000049 AUDIT_SERVER HIB 10 631 0 00:00:26.35 2975 287
0000004A JOB_CONTROL HIB 10 85006 0 00:15:47.81 4604 418
0000004C DNS$ADVER HIB 4 5986 0 00:01:32.52 450 105
0000004D NETACP HIB 10 14674 0 00:35:32.03 43767 1738
0000004E EVL HIB 5 113 0 00:00:49.31 111386 47 N
0000004F REMACP HIB 9 598 0 00:00:02.45 75 49
00000050 LATACP HIB 14 7 0 01:11:58.87 69143 290
00000051 DFS$COM_ACP HIB 10 25 0 00:00:00.44 109 150
00000052 DFS$00010001_1 HIB 10 26 0 00:00:00.86 451 531
00000053 SYMBIONT_0001 HIB 4 17943 0 00:06:52.57 4060 766
00000056 NMAIL HIB 4 34577 0 00:04:44.05 126808 37
00000057 VPA_DC HIB 15 20679 0 01:51:45.44 65911 1135
00000298 CHAMBERS LEF 6 148 0 00:00:03.93 849 187
00001A59 BBazzoo� LEF 2 7733 0 00:04:42.25 92834 504 B
0000111C SYMBIONT_0009 HIB 6 3645 0 00:01:54.49 2280 761
0000141F Paul on again LEF 5 668 0 00:00:20.29 4684 145
000015E1 TURNER HIB 8 2740 0 00:00:38.99 4527 415
000020E2 Keith Shone HIB 5 370 0 00:00:08.14 1343 255
Page Comment
1-38 Bullet 2: punctuation adrift at the end of the line.
Not sure I can make out what is intended.
Bullet 3: typo oves -> moves ?
Since this sentence refers to portions the final words
should be:
"...working set as they are needed."
1-39a End of para. 4: "...consistent with..." rather than
"...consistent will..."
1-46a Last line: typo fo -> of
1-48 Bullet 1: typo duites -> duties
|
| 61.4 | view of Holland | NWGEDU::WIERSMA | Drive a BENTLEY or walk... | Mon May 20 1991 17:06 | 106 |
| Review chapter
A CLOSER LOOK AT THE VMS ENVIRONMENT
Page 1-5: First bullet second dash
When they can access the system---> When can they access the system
Page 1-7a: SHOW CPU command needed CMKRNL privilege.
Page 1-9: second bullet second dash DSSI (Digital Storage System Interconnect)
Page 1-10a: line 4: perform I/O to any disk.
Only if this disk is MSCP-served.
$MCR SYSGEN show mcsp_serve_all
Page 1-11: Why not changing the picture to this:
| vax c
| +---+ +---+
|-----| |\ | H |
| +---+ \ | S |-----+---------+--
| \ /| C | | |
| \ / +---+ | |
| vax b \ / ++++ ++++
| +---+ \ / ++++ ++++
|-----| |------()\ ++++ ++++
| +---+ / \ +---+ | disk | disk
| / \ | H | | |
| / \| S |-----+---------+--
| vax a / | C |
| +---+ / +---+
|-----| |/
| +---+
\
\
+++++
+++++ disk
+++++
Page 1-12 two-dual ported disks ---> 1 dual ported disk and
1 dual ported tape device
Page 1-13: Picture a little lower.
Page 1-16a: line 2 (LAVc or MI) must be (NI or MI).
|
---------- since 1-sept-1989
line 3 It might vote. Better not. Because the cluster can wait for
a satellite to continue.
Mention on the instructor page
$MCR SYSMAN (oper,sysprv)
SYSMAN> set environment/cluster
For showing cluster information.
Page 1-24 Digital Command Languange (DCL)
first bullet: A languange definition of English like commands ---->
--> A command languange definition of English-like words.
(CLI)
first bullet: a user ---> an user
first bullet: Translate English-like commands from an user -->
--> Translate the commands from an user to the CPU
Page 1-19: fourth bullet RUN/DETACH sometimes DETACH privilege needed.
Page 1-34: point 7 There are 3 processes in this job.
A little bit confusion. Why not REGNELL
REGNELL_1
REGNELL_2 (*)
Page 1-36a: fourth bullet, Requires SETPRV. Correct but as an authorized(!)
privilege.
Page 1-38a: Mention here at the end of the Balance set story these two system
parameters:
$MCR SYSGEN
SYSGEN> SHOW BALSETCNT ------- processes in memory
SYSGEN> SHOW MAXPROCESSCNT --- maximum processe you can make
Page 1-38: working set quotas:
- Working set default (WSdef)
- Working set quota (WSquo)
- Working set extent (WSextent)
Page 1-42: Different layout logical SYS$SYSROOT. Correct one is:
$SHOW LOGICAL SYS$SYSROOT
"SYS$SYSROOT" = "$1$DUA0:[SYS2.]"
= "SYS$COMMON" (without colon!!)
1 "SYS$COMMON" = "$1$DUA0:[SYS2.SYSCOMMON.]"
Page 1-45a: fourth bullet, line 6: Use the comand SHOW PROTECTION to determine
the protection of a file.
SHOW PROTECTION only shows the default protection
DIRECTORY/PROTECTION filespec shows the protection of a file.
Page 1-48: first bullet There is only a dash.....?
So far
Bye bye Arjen.
|
| 61.5 | Processor .NES. Node | NWGEDU::WIERSMA | Drive a BENTLEY or walk... | Mon May 20 1991 17:14 | 10 |
| Re 61.2
Hello Arnold,
About your remarks processors changing in nodes.
This is confusing, because 1 node could have 2-4 processors.
I hope we can find something else, or we keep it this way. All right?
Bye Arjen.
|
| 61.6 | And another list... | GOONS::BAKER | What does "ignorant" mean? | Tue May 21 1991 12:49 | 223 |
| Here are my review comments for Chapter 10 - A Close Look at the VMS
Environment.
They may include comments and typos already mentioned. Please excuse
the "terseness" of some items, but I think they convey the necessary
information.
Overall, I think this module achieves its objectives very well.
p1-4 Nothing under RESOURCES heading.
1-5a willbe = will be
help them out the = help them put the
inperspective = in perspective
1-5 line 1 ALthough = Although
bull 2, dash 1 Should this be "How much disk space..." ?
bull 3 dash 2 Suggest "What files they can read, write or delete on the
system"
bull 5 Suggest "Updating, installing and maintaining system software"
1-6a & 1-6 Not happy about classifying a system as a "Vector Processor"
It is really a scalar with attached vector processor, but that
may be confusing too.!
1-6 List of CPUs Why not make it "complete as at time of publishing"?
Include 8650, 8250, 6000-510, VAX 4000-???.
MicroVAX 3000 implies just one machine. Suggest using either
MicroVAX 3000 series or list them all.
And not forgetting the workstations...
1-7a 4th para. Implies that vector processors are independant of other
processors. Also implies that any multiprocessor config can
have attached vector processors. (Only certain 6000 and 9000
systems)
1-7 2nd bull. Suggest that "Cannot operate independantly" should become
"Cannot have processors operating independantly" (or something
like that)
1-9 1st bull Suggest including VAXstations as well.
3rd bull List the supported interconnects.
CI (Computer Interconnect)
DSSI (Digital Storage Systems Interconnect)
NI (Network Interconnect - Ethernet)
FDDI (soon to be supported V5.4-3?)
1-10a 2nd para ...connected to one system...
1-10 Title Suggest "Members of a VAXcluster."
1-12 1st/2nd bull talks of two disks. Diagram only has one disk and a tape
2nd bull, 3rd line Suggest "..upon the availabilily of any VAX
processor..."
1-13 Diagram a bit too high
2nd bull. Definition of a "satellite" is a node which boots remotely, via
a boot server, regardless of whether it has local disks or not.
Suggest the following text to replace both bullets.
o The local area VAXcluster system configuration shows a number
of VAX systems connected by way of Ethernet local area
network connections. The top VAX system functions as a boot
server for the other systems (called satellites) in the
cluster. A satellite downloads VMS from the boot server, and
consequently does not use a locally attached system disk.
o The top VAX system also functions as a disk server making
disks attached to the node available to the satellites. A
satellite may have local disks. It can make these available
to other cluster members if desired, thus becoming a disk
server also.
1-15, 1st bull, last two words. Replace "VAXcluster system" with "cluster"
2nd bull With 5.4, a node can have connection to more than one CI.
1-16 4th bull VAX 4000s and VAXft 3000s can be satellites as well.
1-17a 3rd bull intial = initial
1-17 Title Suggest "VAXcluster Integrity"
1-21 2nd bull. Unclear about the use of "area". Do you mean node address area
or geographical area?
1-25 4th bull, 3rd dash Suggest should be
"Typically invoked as a consequence of issuing DCL commands"
Diagram Suggest "Object program" becomes "Object File"
1-26 2nd bull I'm not clear on what this is telling me. I think it's saying
that a process has associated quotas and limits which restrict
it's use of system resources. Suggest some rewording.
1-27a Table, subproc entry.
Suggest "$CREPRC" becomes "SYS$CREPRC" (like the others)
Also "Some quotas are pooled with owner..."
1-28 Table entries 1,2,3,4 Detached processes don't have parents. Suggest
using "creating" in place of "parent".
1-29 Last bull Suggest replacing "Shares" with "Uses" or "Has"
1-31 FLAG.ESS
Oh dear! It is a shame not to be able to use such literary talent ;-) but in
England (at least) I fear it will become somewhat of a "talking point" to those
of us with "minds close to the gutter" - which is probably why I'm picking up
on this and Keith didn't :-) :-)
I'm too polite to tell you what "fannies" (in the singular form) is slang for
over here, but suffice to say that you won't find any male bleached-blond
pseudo-wavers with a tan AND one of these!. "Suspenders" are garments more
usually found adorning the area in question, and preventing stockings from
becoming wayward.
To find an old dodger wearing work boots and red suspenders would be quite a
turn-up in the UK!!! :-)
Personally I could quite happily live with this text in the book and make an
English joke or two about the content - but I (sadly) believe that the
occasional student will take offense or raise an eyebrow or two.
1-31 (again). Use of EASYnet nodenames/usernames in "public" documentation.
I know that in some places throughout these chapters, attempts have been
made to remove "real" nodenames. Node WAYOUT, however, exists in the
UK, even though there's apparantly not a user SPEAKE there. Not sure
what the policy is on this. Just wanted to point it out.
1-32a Table. I'm not clear on what the OK, Preferred and Other columns are
saying. An explanation would be useful.
COLP = COLPG
CUR entry, suggest last column should read Executing on a CPU
Mutex/Misc entry should also specify MWAIT as well as RWxxx
Suggest PFW definition should be "Waiting for a disk paging
I/O.."
I question the validity of the footnote. Why are SUSP processes
a problem? (unless someone is maliciously SUSPending them)
1-35, callout 6. The information given does not show usage of physical
memory, but use of a scratch area in a process' P1 space.
I suggest this be clarified in the callout text, otherwise it
is misleading.
1-36a 1st bull. /DUMP causes .DMP to be written when process/image dies with an
error. The current text implies that a dump will be written
when the command is executed.
4th bull. Suggest including text that says you do not need SETPRV to make
authorized privs become current privs.
last bull, last line. Suggest rewording: "...until another process resumes
it with /RESUME or deletes it."
1-36 example Use of node SUPER, prefixing terminal name and devices alloc'd.
Should this be edited to become node LDYLDY?
1-37a Item 1 The sentence "Real time processes of the same priority
participate in round-robin scheduling (within their priority)
if they are simultaneously computable" is not true. As
mentioned in the text which precedes it a realtime process
keeps going until it goes into a wait state. The sentence can
be omitted.
Last para. The text uses the term top and bottom of the queue. I prefer
front and back. Also, processes are returned to the queue
associated with their CURRENT priority, which will change after
they have been scheduled.
These comments are a bit picky for the purposes of the discussion, but
I think it's good to get it right.
1-37 1st sent. Suggest changing to "Since a processor..." which covers the
multiprocessor systems.
1-38a. The comments on system WS adjustment are inaccurate. The parameters
exist in 5.4, but are not used. The adjustment code is commented out.
May be used in a future version though.
1-39a 1st para. Suggest changing the word "translation" to "mapping" to make
the sentence more technically accurate.
2nd para. Suggest changing "The technique..." to "One technique..."
Swapping is actually done as a last resort.
1-39 3rd bull. Suggest adding "process" to end of item.
Suggest changing "The swapping file:" to just "The swap file:"
1-41 [SYSMGR] contents. What are the print forms definitions in [SYSMGR]?
The definitions are part of JBCSYSQUE.DAT, which lives in
[SYSEXE].
1-42 SYS$SYSDEVICE translation. The physical Device = The physical device.
1-44a. Penult. para
The comment about typically using [SYS0] is not true in a
clustered environment. Since clusters have already been
mentioned it might be good to mention this here.
1-44 2nd bull - as above.
1-45a 4th bull 2nd para SHOW PROT is used to determine the default protection
applied to new files. DIR/PROT should be used to view
the protection of an existing file.
1-46 last entry. VWS is a layered product, it is not part of the VMS distrib.
1-48 1st bull duites = duties.
Missing "dash" items
2nd bull, 3rd dash Suggest replace with VAXclusters.
5th bull Suggest replace "...editors, VMS utilities..." with
"...editors, utilities..."
Phew! The end!
|
| 61.7 | For the record... | SUPER::MATTHEWS | | Thu May 23 1991 12:15 | 7 |
| A paraphrase of the policy on node names from the April '91 Corporate
Style Guide: use of real node names is OK (thank goodness, since
it's unavoidable unless we want to call them all VAXA and VAXB!) but we
must not publish user names or directory specs that exist on those
nodes.
Val
|
| 61.8 | many small things | TEACH::CHUCK | | Sat Jun 01 1991 01:51 | 38 |
| Module 10: Page 1-5.a "willbe" is run together.
Page 1-11 and 1-12 should face each other.
Page 1-13 Figure needs to be lowered.
Header runs through top.
Page 1-14 and 1-15 should face each other.
Page 1-25 At very bottom of page I would
add below the diagram under "source program"
"ASCII file". below "object program" add
".obj file" and below "image" add ".exe
file".
Page 1-28 The word "PARENT" should not be
used to refer to Detached processes. Once
detached processes are created they are on
their own (no parent). There is no connection
to another process. How about, as an example,
changing:
"given by the parent process"
to:
"given by the UIC of the creating
process"
The whole detached part of this chart should be
re-written to show reality. All characteristics
come from the UIC of the creating process (which
means the authorize.dat file) and/or any specific
qualifiers on the detach command.
Page 1-29 Line two is not a complete sentence.
Add the word "to" after the word "reasons".
Page 1-34 and 1-35 should face each other. They
do now but subsequent printings could change that.
|
| 61.9 | comments from DC Training Center | TEACH::WENDY | | Mon Jun 03 1991 13:20 | 93 |
|
Module 10
A Closer Look at the VMS Enviornment
1-10 Local to each member can be the following
Memory (It shouldnt be in the list because it is local, this implies
it could be local or shared.
1-12 Third bullet
Reads if one HSC fails, all traffic to the connected disks automatically
switches to the second HSC. *(If you have them set up for failover.)
1-13 Move up heading
last bullet
Reads- The top Vax system also functions as a boot server for the diskless
Satellites (those VAX systems without local disks). *(I think it should
say wihout local system disks, because two of the satellites have local
disks, granted they are for paging and swapping, but they are still local.
1-15 The mixed interconnect VAXcluster system configuration shows a combination
of the other VAXcluster system configurations. *(elaborate on other types)
2nd bullet
Each vax system can have one CI connection and one E-net connection *(I
thought V5.4 supported multiple CI connections. *(5-23 Perf ins guide)
1-16Disk Servers
Can serve
-Locally connected disks
-The HSC disks if connected to an HSC
*(The HSC disks, if the disk server is also connected to the HSC)
-The DSSI disks if connected to a DSSI
*(The DSSI disks if the disk server is connected to the DSSI Bus)
1-17 2nd bullet from the bottom
Quorum management ensures that, if partitioning occurs, the part that
represents lessthan half of the votes will hang. *(The part of the cluster
that represents...)
1-19 .because the two nodes are connected in a network, a user logged into one
of the nodes can:
-communicate with a user who is logged into the other node *(into)
-Access disk files stored on the other node (If you have proxy account)
1-28 Put this page as a summary after you talk about creating these proceses.
Its sort of a summery.
1-30 2nd bullet from the bottom
Processes created with the SUBMIT command read their instructions from
a command procedure.
*(processes created with the SUBMIT command read their instructions from
the command procedure that was submitted.
1-38 Balance Set
.Analogous to the Working Set *(this is unclear to me)
.Limited by a system parameter *(don,t say this if we are not talking
about parameters)
1-39 The swapping file
.Stores process working sets that do not fit into physical memory
*(should elaborate more on this,its not that they don't fit, memory
is tight and certain processes arent active so the idea is to get
them out of memory to let the active processes continue.
1-40
.system volume
.system device
*(to me volume and device have been used interchangably until now)
I think the following pages should be re-numbered.
1-41 should be 1-43, and 1-42 should be 1-44
1-44 should be 1-41 *(show them what concealed rooted directoried are)
1-43 should be 1-42
I just think that you should show them how to set up concealed rooted
directories before you show them the system directory stucture and
logical names that point to each place.
In all should be numbered 1-44,1-43,1-41 then 1-42
1st 2nd 3rd fourth
Wendy Mullenhoff
|
| 61.10 | more comments from dc | TEACH::RINE | | Wed Jun 05 1991 14:53 | 14 |
| The following are comments on Module 10 A closer look at the VMS Environment
Page 1-10. It may be too early to mention Page and Swap Files without
going into detail about what they are and so on.
Page 1-13. We should call this a LAVC not an Ethernet only cluster.
Page 1-38. There is Typo in the 3rd Bullet. oves instead of Moves.
Page 1-38. it could be a little too confusing to say a working set is
analagous to the balance set.
Page 1-39. It may be better to say that process are swapped when the
available memory on the system becomes very critical.
|
| 61.11 | Capitol Ideas | TEACH::LYNN | | Sun Jul 07 1991 00:22 | 36 |
| A Closer Look at the VMS Environment
Chapter 10
Overall: Good list of objectives for this chapter.
1-5a Second line typo - willbe
Second line Second sentence doesn't make sense.
1-5 Typo, next to the last bullet udate should be update
1-13 Before this page discuss the concepts of the cluster and the
different types.
1-16 Maybe move this page before 1-13.
1-17 Since you are bringing up the concept quorum you might as well finish
the job here. Cover all the information about quorum and partitioning
and show some examples.
1-29 Fix the wording on the first line please.
1-35 Bullet 2 Be careful with the wording. I understand the sentence
"It is continuously updated." but the student will maybe think
the screen is updated. Maybe make this a littler clearer.
1-36 It might not hurt to define a job tree. It is used alot in course
and VMS documentation.
1-38 Third bullet typo "oves" should be "moves"
1-45 Take this page out. Since this is SYSNET I remove it. They don't
need to know about this dangerous act of changing protection on
system files and directories.
Lynn White
Washington D.C.
|
| 61.12 | URGENT: Deletion review requested | SUPER::REGNELL | Modularity Maven | Thu Aug 29 1991 19:34 | 29 |
| The following topics from the "Advanced VMS Topics" chapter
of SYSNET I have been put up for deletion.
During the pilot we found that these topics were way beyond most
students and since they are presented in such simplistic terms, are
often useless anyway. 'Content Free' in effect.
Rather than beef them up and add a large chunk of time and lab to an already
packed course, we are suggesting deleting them here. They are already
presented again in SYSNET II and are actually covered in depth in SYSNET III.
Please take a look and if you have concerns or questions, enter a reply
here or send mail to SUPER::REGNELL.
Deadline for discussion: Sept 6, 1991
Thanks
Mel
Scheduling
Priority Levels
Working Set
Paging
Swapping
Balance Set
|
| 61.13 | Yep, this is a "Don't need to know" | DUCK::SHONEK | Keith Shone UK Edu 830-4074 | Fri Aug 30 1991 03:55 | 35 |
| Mel
I haven't looked at the chapter concerned - it must be somewhere in the
15 inches of paper on the new curriculum ;-)
This is one of those "need to know", "nice to know" issues. For me its
a "who cares if they don't know"! Seriously...
My students are encouraged to use HELP, snoop about and try things
out on VMS. They'll find
$ SHOW WORKING_SET
$ SHOW PROCESS
$ SHOW MEMORY
$ SHOW EVERYTHING_ELSE
and ask questions about:
Scheduling
Priority Levels
Working Set
Paging
Swapping
Balance Set
Simple answers on an individual basis will do fine.
Just so long as instructors know where to find "...simplistic terms..."
that's OK with me.
Some of these questions arise from familiarity with other operating
systems. A similar approach may be adopted.
Let's keep it simple and leave this out. (Now why didn't I say that at
the start?)
|
| 61.14 | | NITTY::DIERCKS | None of your business!!!! | Tue Sep 03 1991 14:56 | 10 |
|
But, where will these topics be placed? These are definitely "need to
know" kind-o-things for the person interested in the performance end of
things.
(Greg -- who suggested they be beefed up, not deleted, after sitting in
on the pilot of SYSNET I.)
|
| 61.15 | | SUPER::MATTHEWS | | Wed Sep 04 1991 13:37 | 10 |
| A reason to keep the memory-management discussion is that the system
manager has to know where the paging and swapping file(s) belong
(especially when adding a cluster node in Sys/Net III, but I think the
existence of the paging file crops up before then).
Maybe you don't have to understand what the paging file is for in order
to do that, but aren't 99% of students going to ask what it is if we
don't tell them?
Val
|
| 61.16 | {Menory management section of 10} | TEACH::CHUCK | | Thu Sep 05 1991 01:22 | 31 |
| Mel,
Having looked at all 3 courses, the memory management topics
are covered in each.
I agree in droping them from SYSnet I. The big picture is
a string of courses that, when the student is done, he/she will
have what is needed to be a good system manager. Right now the
plan is to have 4 courses: SYSnet I thru IV. Correct me if I
am wrong (I've been there before), The information for these 4
courses is coming from and will replace the following:
Utilities & Commands I
Utilities & Commands II
System Manager I
System Manager II
Cluster Management
Network Management I
If we are to try to jam 6 weeks into 4 then we do not need to
reteach topics. Let us teach it once then move on. My recom-
mendation for the Memory Management is to teach it all/completely
in SYSnet II, then drop it from SYSnet III. It is out of place
there in the maintaining disks module anyway. Is this too soon?
I do not think so. I teach it now in UC II, SM I and Sys Arch.
These are the second course for many students after uc I. They
in general can handle it very well.
Prediction: We may be seeing the birth of SYSnet V and I'm
not sure that is wrong.
|
| 61.17 | ...conversation tidbit... | SUPER::REGNELL | Modularity Maven | Fri Sep 06 1991 11:02 | 12 |
|
Hi Greg,
Yes, I know you suggested they be beefed up. And you were
right, they were not covered in any way that was sufficient
or even helpful. But in looking at how much we needed to add
to really teach them...and considering they are mentioned/taught
in all three...we opted to suggest removing them instead of
making SYSNET I unteachably long and loaded.
Mel
|