f



SYS$SPECIFIC and SYS$COMMON not enough?

Basically, the idea is that stuff specific to a given node will be in 
its SYS$SPECIFIC directory and stuff common to all nodes (satellites or 
not) booting from the same system disk will be in SYS$COMMON.  A good 
idea as far as it goes, but does it go far enough?

In particular, there are two additional categories, which I'll give the
obvious names SYS$CLUSTER and SYS$WORLD.  SYS$CLUSTER would include
stuff common to all nodes in a cluster, i.e. would differ from
SYS$COMMON if there is more than one system disk in the cluster.  The 
obvious place for this stuff is somewhere which is not on ANY system 
disk.  SYS$MANAGER:SYLOGICALS.TEMPLATE provides (at least in newer 
versions of VMS, starting with 7.2 I believe) a list of at least the 
most common candidates among VMS system files.  Add to this procedures 
to mount disks in a cluster, SYLOGIN.COM etc.  (Obviously, I'm thinking 
of a cluster where as much as possible is shared.)  SYS$WORLD, on the 
other hand, would include stuff which is common to ALL VMS system disks 
in the world, e.g. the executables used by VMS itself.  Obviously, this 
should be on EVERY system disk.

Put another way, the current SYS$COMMON is a mixture of stuff common to 
all nodes which boot from that disk, stuff common to all VMS systems in 
the world and, if the stuff is not located somewhere else (pointed to by 
logicals in the case of VMS system files), stuff common to all nodes in 
the cluster (perhaps duplicated on more than one system disk).

SYS$CLUSTER is currently supported somewhat, though of course this name 
is not used: one can use logicals to point to files used by VMS 
utilities (such as SYSUAF.DAT used by AUTHORIZE) and it is easy to tell 
SYS$COMMON:[SYSMGR]SYSTARTUP_VMS.COM to execute a command file located 
off the system disk to (directly or indirectly) do stuff like mount 
disks within the cluster.  (This DOES have to be maintained, though, for 
each system disk.)

What worries me more is that what is left in SYS$COMMON is a mixture of 
stuff common to all nodes booting from that disk and stuff common to all 
VMS systems in the world.

In a shared-everything cluster, multiple system disks are there just for
redundancy, and one can put the SYS$CLUSTER stuff somewhere else.
Obviously, for the redundancy, one wants multiple copies of SYS$WORLD,
and the present SYS$COMMON seems like a good place to have it.  Still,
one can think of realistic examples of stuff which is shared among all
nodes booting from a common system disk, but not among all nodes in the
cluster, much less all nodes in the world.  (For example, one might have 
one boot server and its satellites in one room, another boot server and 
its satellites in another room, and want to define SYS$PRINT to point to 
(queues which eventually resolve to a queue that prints on) a printer in 
the room where the machines are.  Ugly workarounds are to define the 
logical individually on each node, or use conditional code in a command 
procedure located in the SYS$CLUSTER area.)

Questions:

What are some real-world examples of stuff which is common to all nodes 
which boot from a common system disk, but not common to all nodes in a 
cluster, even in a share-as-much-as-possible cluster?

Is there any chance of extending definitions like SYS$SYSROOT to include 
SYS$CLUSTER while retaining backward compatibility?

One area where the current confusing situation manifests itself is the
following strategy: maintain ONE "master system disk", i.e. upgrade this
disk, install layered products on it etc then replace other system disks
with (modified) copies of this disk (mainly, modifications are stuff
related to changing the node name).  In other words, it would be nice if 
SYS$COMMON (SYS$WORLD in this context) contained ONLY stuff common to 
all VMS system disks in the world.  (Another worry is how to deal with 
SYS$SPECIFIC; it's not just a matter of saving the contents before the 
upgrade and putting them back later, since the upgrade itself might add 
or modify files in SYS$SPECIFIC.)

0
HELBPHI (10)
7/7/2003 5:48:27 PM
comp.os.vms 21904 articles. 1 followers. Post Follow

2 Replies
2078 Views

Similar Articles

[PageSpeed] 29

In article <01KXZKP04MZUAM7Y4A@sysdev.deutsche-boerse.com>, Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes:
:What are some real-world examples of stuff which is common to all nodes 
:which boot from a common system disk, but not common to all nodes in a 
:cluster, even in a share-as-much-as-possible cluster?

  You need to answer this question yourself, and specifically for
  your local environment.  One example would be SYSUAF.DAT quota
  settings for various hosts, if running parallel SYSUAF files. 
  (You need to be VERY careful here, as you must keep all UICs and
  all usernames and all identifiers involved scrupulously matched.)

  I am aware of nothing here, however, that can't also be handled
  "directly" with per-file logical name redirections.  SYSUAF can
  certainly be redirected, of course.

:Is there any chance of extending definitions like SYS$SYSROOT to include 
:SYS$CLUSTER while retaining backward compatibility?

  We have something similar here in OpenVMS Engineering, and we do 
  create and use a CLU$COMMON root within SYS$SYSROOT:

$ sho log sys$sysroot
   "SYS$SYSROOT" = "ddcu:[root.]" (LNM$SYSTEM_TABLE)
        = "SYS$COMMON:"
1  "SYS$COMMON" = "ddcu:[root.SYSCOMMON.]" (LNM$SYSTEM_TABLE)
        = "CLU$COMMON:"
2  "CLU$COMMON" = "ddcuother:[root.]" (LNM$SYSTEM_TABLE)

  though we do know that various of the LP installation procedures can
  and regularly do become somewhat confused with this structure.  Some
  of the engineers have informally discussed implementing and supporting
  this configuration, and (obviously) fixing and extending the various
  OpenVMS installation procedures and tools -- but another and far larger
  project has, um, intruded on this particular OpenVMS enhancement.

  We walk the current list of SYS$SYSROOT translations within some DCL 
  within SYLOGICALS.COM, and we append the CLU$COMMON translation onto
  the searchlist.

 ---------------------------- #include <rtfaq.h> -----------------------------
    For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq
 --------------------------- pure personal opinion ---------------------------
        Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com

0
hoff (611)
7/7/2003 6:52:31 PM
Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> writes in article <01KXZKP04MZUAM7Y4A@sysdev.deutsche-boerse.com> dated Mon, 07 Jul 2003 19:48:27 +0200 (MET):
>What are some real-world examples of stuff which is common to all nodes 
>which boot from a common system disk, but not common to all nodes in a 
>cluster, even in a share-as-much-as-possible cluster?

One use of a cluster is testing new system SW.  You can make a copy of your
main system disk and upgrade the software there (say, a VMS patch).  Then
boot one system using that disk and see how it works, while the rest of the
cluster does your production work using the old system without the upgrade. 
In that case, any files affected by the upgrade would/should be different
for the test node.  

The audit journal file is another example.  It's in each
SYS$COMMON:[SYSMGR].  Why does it need to be on the system disk?  The system
might need to access it at any time, and the system disk is the only disk
which is (or needs to be) guaranteed available.

>Is there any chance of extending definitions like SYS$SYSROOT to include 
>SYS$CLUSTER while retaining backward compatibility?

You can rig all kinds of stuff.  I'm sure there have been discussions here
about having SYS$COMMON and SYS$SPECIFIC point to different devices.  The
"VMSCluster Systems for OpenVMS" manual has a list of ~10 data files which
it recommends be migrated off the system disk if you're using multiple
system disks in your cluster.  

--Keith Lewis              klewis$mitre.org
The above may not (yet) represent the opinions of my employer.
0
lewis16 (87)
7/8/2003 12:44:59 AM
Reply:

Similar Artilces:

sys
Is it OT here to talk about IPC and pipes? Or are all aspects of kernle programming ok here? Bill Bill Cunningham <nospam@nspam.com> wrote: > Is it OT here to talk about IPC and pipes? Or are all aspects of kernle > programming ok here? If you want to discuss how IPC and pipes are implemented in the kernel it's completely ok. But if you're interested in how to use IPC or pipes from a userland program then posting to e.g. comp.unix.programmer or comp.os.linux.development.apps will probably make a lot more sense. R...

sys
computer common oriented http://www.freewebs.com/buildem/ http://amigos.com/go/g904313-pmem ...

SYSMAN: No SYS$SCRATCH/SYS$LOGIN ?
$ mc sysman SYSMAN> set env/node=velo %SYSMAN-I-ENV, current command environment: Individual nodes: VELO Username JFMEZEI will be used on nonlocal nodes SYSMAN> do show log sys$scratch %SYSMAN-I-OUTPUT, command execution on node VELO %SHOW-S-NOTRAN, no translation for logical name SYS$SCRATCH SYSMAN> There is also no SYS$LOGIN defined. (VAX 7,3 and Alpha 8.3) What is the reason behind the lack of those logicals ? JF Mezei wrote: > $ mc sysman > SYSMAN> set env/node=velo > %SYSMAN-I-ENV, current command environment: > Individual nodes:...

com.sys & vcom.sys
It's been a while since I last installed 1.1 and when I did it yesterday I forgot to de-select Ray Gwinn's com drivers. After installation I removed the entries from config.sys and un-remmed com.sys and vcom.sys but they won't load. I checked and it's in c:/os2/boot. What's the cure? :-) Wayne -- While the e-mail address in this post is genuine it is only used as a spamtrap. Any mail sent to me at this address will be considered to be spam and deleted without being read. Wayne wrote: > It's been a while since I last installed 1.1 and when I > did it yesterday I forgot to de-select Ray Gwinn's com > drivers. After installation I removed the entries from > config.sys and un-remmed com.sys and vcom.sys but they > won't load. I checked and it's in c:/os2/boot. What's > the cure? :-) They must come after mouse/vmouse/pointdd device drivers in the config.sys file. Stu/2 On Tuesday 05 Oct 2004 17:06 Stewart Buckingham wrote: >> It's been a while since I last installed 1.1 and when I >> did it yesterday I forgot to de-select Ray Gwinn's com >> drivers. After installation I removed the entries from >> config.sys and un-remmed com.sys and vcom.sys but they >> won't load. I checked and it's in c:/os2/boot. What's >> the cure? :-) > > They must come after mouse/vmouse/pointdd device drivers in the > config.sys file. I believe they are, I checked while I was...

com.sys and usbser.sys question
I got a question, for those knowing more about that matter than myself, which shouldn't be hard ;-) Temporarily having 2 phone lines at my disposal, I wanted to put my old modem back to use as a phone number dialer, using the BonusPack PIM Phonebook application. But I had quite some trouble to get my COM1 port to work at all with any comm application. It finally turned out, that COM.SYS has to be loaded before USBSER.SYS, since otherwise USBSER.SYS is hijacking COM1 unnoticed. Now, in the order shown below, it assigns itself to COM6 (if not remed out), but COM1 still d...

problem with reload(sys) (doing reload on the sys module)
Python 2.4.2 (#67, Sep 28 2005, 12:41:11) [MSC v.1310 32 bit (Intel)] on win32 Type "copyright", "credits" or "license()" for more information. **************************************************************** Personal firewall software may warn about the connection IDLE makes to its subprocess using this computer's internal loopback interface. This connection is not visible on any external interface and no data is sent to or received from the Internet. **************************************************************** IDLE 1.1.2 >&g...

sys.stdout / sys.stderr, subprocess, redirection
Hello, I have a piece of code that gets run in a script that has its stdout closed: import sys sys.stdout = sys.stderr c = subprocess.Popen (..., stdin = subprocess.PIPE, stdout = subprocess.PIPE, stderr = subprocess.STDOUT) and this is what I get: Sending SVNR/permissions Transmitting file data .svn: Commit failed (details follow): svn: 'pre-commit' hook failed with error output: Traceback (most recent call last): (...) File ".../__init__.py", line 40, in run stderr...

OS/2 gurus! Help with OS/2 config.sys Fwd: Trouble with a Personal System/2
Folks, I suck at OS/2, and this guy has a problem. Config.sys has been changed. Any way to boot in recovery mode? No further info yet, unknown if this was a result of a FixPack, adapter install, or a program update. Challenge, this may be a production unit, and it may be doing involved things. -------- Original Message -------- Subject: Trouble with a Personal System/2 Date: Thu, 30 Sep 2010 20:40:20 -0400 From: <brian@centralindianapatternandmold.com> To: <ohland@charter.net> please help . config.sys got changed os2 ibm indurstial computer-will not boot? In article <npapo.23064$qU5.8268@newsfe20.iad>, Louis Ohland <ohland@charter.net> wrote: > Folks, I suck at OS/2, and this guy has a problem. > > Config.sys has been changed. > > Any way to boot in recovery mode? No further info yet, unknown if this > was a result of a FixPack, adapter install, or a program update. > > Challenge, this may be a production unit, and it may be doing involved > things. 1. Boot (C-A-D). 2. Wait for "boot blob" at top left corner of the screen (a white box, and "OS/2"). 3. Press Alt-F1. 4. Select "Boot to command prompt (use original CONFIG.SYS)" option. 5. Fix the CONFIG.SYS Alternatively, boot to command prompt from the OS/2 installation diskettes. -- Don Hills (dmhills at attglobaldotnet) Wellington, New Zealand "New interface ...

sys/processor.h sys/loadavg.h in linux?
I have a program I wrote for Solaris. I am trying to port the program to linux. One of the problems I am having is attaining hardware information on the machine. Using sys/processor.h on solaris, I am able to num cpu's and clock speed. Using sys/loadavg.h, I am able to get the load avg u/s/i. I can still get mem, and free mem from sysconf, that is ok. I would rather not parse through /proc/cpuinfo! TIA, ~S ...

DADCONF.NET.SYS is created by JCONFJOB.NET.SYS
Never mind the question about DADCONF.NET.SYS. I found the following: http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/mpeix/onlinedocs/30 216-90325/30216-90325_top.html&con=/mpeix/onlinedocs/30216-90325/00/00/6 8-con.html&toc=/mpeix/onlinedocs/30216-90325/00/00/68-toc.html&searchter ms=dadconf.net.sys&queryid=20041022-080827 For the network to work correctly, stream JCONFJOB. :STREAM JCONFJOB.NET.SYS The JCONFJOB.NET.SYS stream is required for all systems. It includes information files that describe the supported set of servers and services. These information fi...

thread specific sys.stdout?
This may sound a little crazy. I capture the output of one class by redirecting the sys.stdout. However the is another threading running at the same time and occasionaly it output some messages to the redirected sys.stdout irreleveant to the output I want to capture. Is there a way to redirect output specific to some threads? aurora aurora wrote: > This may sound a little crazy. I capture the output of one class by > redirecting the sys.stdout. However the is another threading running at > the same time and occasionaly it output some messages to the redirected > sys.stdo...

Sys Technology LCD monitor (sys-172L-C-JT)
Greetings, I have 2 of these monitors and the power LED just flashes on and off. It has a switch mode power supply. I found no "smoking gun" so to speak. I checked all of the power supply's electrolytics with an ESR meter. The fact that I have 2 with the same problem tells me that there must be others out there too. Has anyone repaired any of these with this problem? If not I will do the work and post what repairs it. Thanks have you tried them both on a different system with a different vid card? If it still happens, there is probably a problem with the display...

redhat 8
hi, i'm running redhat 8 and cannot find the two OS include files sys/dkio.h and sys/dklabel.h. nor can i find references in internet groups that anyone else has this problem. any ideas why this doesn't not come standard with redhat dist, or why i seem to be the only one with this problem (this seems strange to me). and most importantly, where the best place to download this from? thanks in advance, richard boaz On Thu, 02 Sep 2004 10:52:13 +0200, richard wrote: > i'm running redhat 8 and cannot find the two OS include files sys/dkio.h > and sys/dklabel.h. If I r...

PC, Sys Boards, et OS...
For Sale: A7V333 System Board w/AMD 2100 Processor: The A7V333 incorporates VIA� KT333 chipset to support AMD� socket A processors and high-bandwidth PC2700 DDR (double data rate) memory. Loaded with additional features such as ATA133 RAID, USB 2.0 and 1394 ports, 6-channel hardware audio, as well as innovative ASUS thermal protection technologies, the A7V333 delivers optimum performance and value. Nothing else: $60.00 for both. Good Condition. Will put Thermal Paste and Fan on processor. Pentium 4 with Gigabyte 8TX board, which is a 850 Board which takes RDRAM, ...

os._exit vs. sys.exit
Quick question: Why does os._exit called from a Python Timer kill the whole process while sys.exit does not? On Suse. Bryan Bryan wrote: > Why does os._exit called from a Python Timer kill the whole process while > sys.exit does not? On Suse. os._exit calls the C function _exit() which does an immediate program termination. See for example http://developer.apple.com/documentation/Darwin/Reference/ManPages/man2/_exit.2.html and note the statement "can never return". sys.exit() is identical to "raise SystemExit()". It raises a Python exception which may be caught at a higher level in the program stack. Andrew dalke@dalkescientific.com Andrew Dalke wrote: > sys.exit() is identical to "raise SystemExit()". It raises a Python > exception which may be caught at a higher level in the program stack. And which *is* caught at the highest levels of threading.Thread objects (which Timer is based on). Exceptions raised (and caught or not) in a Thread do not have any effect on the main thread, and thus don't affect the interpreter as a whole. -Peter "Peter Hansen" <peter@engcorp.com> wrote in message news:u-SdnRAzQ9KHhnffRVn-gg@powergate.ca... > Andrew Dalke wrote: >> sys.exit() is identical to "raise SystemExit()". It raises a Python >> exception which may be caught at a higher level in the program stack. > > And which *is* caught at the highest levels of thr...

multiprocessing module and os.close(sys.stdin.fileno())
Why is the multiprocessing module, ie., multiprocessing/process.py, in _bootstrap() doing: os.close(sys.stdin.fileno()) rather than: sys.stdin.close() Technically it is feasible that stdin could have been replaced with something other than a file object, where the replacement doesn't have a fileno() method. In that sort of situation an AttributeError would be raised, which isn't going to be caught as either OSError or ValueError, which is all the code watches out for. Graham On Tue, Feb 17, 2009 at 10:34 PM, Graham Dumpleton <Graham.Dumpleton@gmail.com> wrote: > Why...

How to use sys/socket.h functions in windows OS.
Hi All, The following code snippet is a part of s/w which is downloaded from net. While compiling this code I got the following error. ...\..\snmplib\snmpTCPDomain.c(6) : fatal error C1083: Cannot open include file: 'sys/socket.h': No such file or directory NMAKE : fatal error U1077: 'cl.exe' : return code '0x2' Stop. NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~3\VC98\BIN\NMAKE.EXE' : return code '0x2' Stop. I just looked into the code and I found out that the problem is raised because of sys/socket.h file. So please help me as how to use the sys/s...

How to capture screen motion on OS X or Sys 9
What 3rd party program can be used to capture motions on screen for creating software demo? Thanks, cpliu In article <bg6m7p$l679d$1@ID-46967.news.uni-berlin.de>, cpliu <chanciusliuDeleteThis@yahoo.com> wrote: > What 3rd party program can be used to capture motions on screen for > creating software demo? Snapz Pro X http://www.ambrosia.com -- Larry Fransson Aviation software for Mac OS X! http://www.subcritical.com ...

ProDOS 8 prefix on launching SYS files from GS/OS ?
Hi, I'm currently developing a ProDOS 8 SYS application and wonder about the user experience for GS/OS users. When launching a 8 bit SYS file from GS/OS then the system "boots into" ProDOS 8. From that I'd expect that the ProDOS 8 prefix is set the same way it is set when directly booting ProDOS 8 (which is it is set to the volume directory of the boot drive) - is that correct? Is there a way that a user can manipulate / override that behaviour? I've read that a user can "associate" a "file type" with a certain 8 bit SYS program. Does file type her...

Change sys$specific:[tcpip$smtp] directory to another disk
Hi: Our corporate mail runs on OpenVms, some users on 6.2 and some on 7.3. To avoid sys$sysdevice disk trashing I wanted to move the smtp directory To another disk. On Vms 6.2 this was a piece of cake, I just changed ucx$smtp home's directory and voila. On Vms 7.3 I did then same, changing tcpip$smtp home's directory but nothing happened. Then I found TCPIP$SYSTEM:TCPIP$SMTP_RECV_RUN.COM with some convoluted vms sentences which also included a sd to sys$specific:[tcpip$smtp] which I promptly commented out. To be on the safe side stopped and restarted smtp service but still no cha...

Attn.Alexander Shumilov [Was OS/2: W420041111 kernel and config.sys problem]
Hi Alexander I recently received this email:- Hi! I'm working on the patch for os2ldr that caused the problem ( http://groups.google.com/groups?hl=ru&lr=&threadm=b1alBlW2vn3Q092yn%40daina.bga.com&prev=/groups%3Fhl%3Dru%26lr%3D%26group%3Dcomp.os.os2.bugs ) Can you help me with some information? Send me, please, hardware configuration of your computer (especially BIOS manufacter and version). Also, I'm sending you floppy image. Will be nice if you make bootable floppy, boot from it on the computer that caused the problem and send me information you'll se...

upgrade from VAX/VMS 6.2 to 7.3: missing sys$network_services.exe
I tried to go from VAX/VMS 6.2 to 7.3 using the vmsinstal procedure. I answered NO to the "are you doing a rolling update" question. I asked for DECnet Phase IV and all default options. Then I went until the reboot phase. At the reboot phase, after the welcome banner, I get: %EXECINIT, error loading system file - SYS$NETWORK_SERVICES.EXE R0 = 00000910 ?06 HLT INST PC = 80C9A861 and the reboot session loops. Any idea? D. In article <3f9ff7b8$0$10417$626a54ce@news.free.fr>, Didier Morandi <no@spam.com> writes: > I tried to go from VAX/VMS 6.2 to 7.3 using the v...

US-TX-Houston: System Administrator; Integrate HW/SW/OS, HP Proliant Intel sys; (45329137232)
US-TX-Houston: System Administrator; Integrate HW/SW/OS, HP Proliant Intel sys; (45329137232) ============================================================================================== Position: System Administrator Reference: SMC01771 Location: Houston TX Duration: 1Y Scope: Update client's Infrastructure Management System and integrates various ancillary systems and software packages. Update administration documents (Server Administration, Database Administration) monthly to keep current for the IMS ...

US-TX-Houston: System Administrator; Integrate HW/SW/OS, HP Proliant Intel sys; (45330014404)
US-TX-Houston: System Administrator; Integrate HW/SW/OS, HP Proliant Intel sys; (45330014404) ============================================================================================== Position: System Administrator Reference: SMC01771 Location: Houston TX Duration: 1Y Scope: Update client's Infrastructure Management System and integrates various ancillary systems and software packages. Update administration documents (Server Administration, Database Administration) monthly to keep current for the IMS ...

Web resources about - SYS$SPECIFIC and SYS$COMMON not enough? - comp.os.vms

4.3. Implementations and Platforms
The NTP FAQ and HOWTO : Understanding and using the Network TimeProtocol (A first try on a non-technical Mini-HOWTO and FAQ on NTP ) Prev Next ...

Resources last updated: 3/18/2016 1:49:19 AM