1 INFO-VAX	Fri, 08 Aug 2003	Volume 2003 : Issue 435       Contents: Re: 128 Bit EncryptionA Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) A Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS) ) Re: Computer Systems & Automation Manager ! Re: DS-35x-FA Dip switch settings  Re: DSSI problem4 Re: How many errors are too many when making CD-R's?4 Re: How many errors are too many when making CD-R's? Re: HP FUDBusting + Re: Idiot (I know that's redundant) spammer  Re: Netware is no VMS  Re: Netware is no VMS  Re: WASD and RDB?   F ----------------------------------------------------------------------  # Date: Thu, 07 Aug 2003 23:06:55 GMT , From: Wayne Morrison <Wayne.Morrison@hp.com> Subject: Re: 128 Bit Encryption & Message-ID: <3F32DB87.9EF8695A@hp.com>   Larry Kilgallen wrote: > p > In article <3f2e36e5$0$45998$65c69314@mercury.nildram.net>, "Gary Sherwood" <gary@protsys.demon.co.uk> writes:O > > Does anyone know if  there any standard encryption routines (preferably 128  > > bit) in VMS? > + > Check "CDSA" on the latest Alpha release.  > F > But you should be careful about what you want if your description isE > "128 bit".  There is a whole lot more to good cryptography than key 	 > length.   I Thanks for the plug, Larry.  I see you've been keeping up with the latest 	 features.   N Yes, CDSA will provide a standard way to use encryption of varying key lengthsM on OpenVMS, as well as other platforms that support CDSA (The Intel reference H code is available on SourceForge).  It also allows you to write your ownK encryption plug-in module - called a Cryptographic Service Provider in CDSA H terms - for new or special methods that may not be currently supported. N Out-of-the-box, CDSA supports DES, 3DES, and a number of other encryption/hashA types.  AES is coming in at some point in the future (post V2.0).   & Check out the OpenVMS documentation atH http://h71000.www7.hp.com/doc/os731_index.html and look for "Open SourceN Security for OpenVMS Alpha Volume 1: Common Data Security Architecture".  It'sO the V1.0 documentation, and we're about to release V2.0, but it will give you a & good feel for the basic functionality.  N CDSA ships as part of OpenVMS V7.3-1 and up, but the kit will install & run on V7.2-2 as well.    	Wayne Morrison  	CDSA & Kerberos Project Leader  	OpenVMS Engineering   ------------------------------  % Date: Thu, 07 Aug 2003 11:03:02 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com> J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)' Message-ID: <3F329456.8040302@MMaz.com>    Rob Young wrote:  V >In article <3F326FEE.9060709@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes: >    > - >>For those that do not take this particular  C >>rag because of its Unix roots, the link for more reading on this   >>specific article is:  : >>http://www.samag.com/documents/s=8817/sam0313b/0313b.htm >>F >>Bottom line, the Linux Open Source folks are attempting to tackle a K >>clustering issue that should have been handled long, long ago in a world   >>called Digital...  >> >>     >> > = >	What on that punch list of 9 items can't or isn't performed 9 >	by VMS?  I don't see anything there that is not part of  >	VMS other than:  > L >9. Application/Resource Monitoring and Failover/Restart -- A subsystem thatP >takes certain user-defined actions when a node fails, or an application stops.  >  >	 > G Sorry, the prior article on openMosix which is based on OpenSSI at URL  8 http://www.samag.com/documents/s=8817/sam0313a/0313a.htm  H openMosix addresses teh Single System Image issue and the unified user,  process, and file systems...   Barry    --    > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                            ------------------------------   Date: 7 Aug 2003 13:32:39 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <Gfj017WkGiLe@eisner.encompasserve.org>   a In article <VIb6DP4xrsHW@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: W > In article <3F326FEE.9060709@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  >> Bob Ceculski wrote: >>   > . >> For those that do not take this particular D >> rag because of its Unix roots, the link for more reading on this  >> specific article is: ; >> http://www.samag.com/documents/s=8817/sam0313b/0313b.htm  >>  G >> Bottom line, the Linux Open Source folks are attempting to tackle a  L >> clustering issue that should have been handled long, long ago in a world  >> called Digital... >>   > > > 	What on that punch list of 9 items can't or isn't performed: > 	by VMS?  I don't see anything there that is not part of > 	VMS other than: > M > 9. Application/Resource Monitoring and Failover/Restart -- A subsystem that Q > takes certain user-defined actions when a node fails, or an application stops.   >  > 	respawn?  >   > 	cron:2:respawn:/usr/sbin/cron >  > 	So VMS lacks respawn?  B Without further explanation I don't see how restartable batch jobs9 together with generic queues don't handle that just fine.   A > 	AIX with HACMP (failover clustering) does just that, start the B > 	application up on the fallover node.  And boy can it by painfulA > 	(not work) if things aren't synchronized properly (filesystems : > 	for example).  So maybe the Linux crowd improves HACMP.   ------------------------------   Date: 7 Aug 2003 13:46:15 -0500 + From: young_r@encompasserve.org (Rob Young) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <PoS0CsLU0vzB@eisner.encompasserve.org>   U In article <3F329456.8040302@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  > Rob Young wrote:  I > Sorry, the prior article on openMosix which is based on OpenSSI at URL  : > http://www.samag.com/documents/s=8817/sam0313a/0313a.htm > J > openMosix addresses teh Single System Image issue and the unified user,  > process, and file systems... >    	Okay.  8 http://www.samag.com/documents/s=8817/sam0313a/0313a.htm  N In deciding whether to install an openMosix cluster at your installation, it'sJ important to realize that the smallest manageable unit in openMosix is theK process. An application that runs as a single process will not benefit from G running under openMosix. If the same application is broken into several F processes that can run in parallel, these processes can be distributedE throughout an openMosix cluster to enhance performance (see Robbins). L Ultimately, if you're looking for a transparent, dynamic method of balancingF workload across a Linux cluster, and your workload can run as parallel2 processes, openMosix may be the solution for you.     F 	So a modern multi-threaded application (Oracle 9i multi-thread mode)  	won't benefit.   A 	When a node in that case gets overloaded, you would dynamically  8 	migrate CPUs to the load (Galaxy) solving that problem.  ? 	As more applications become threaded (natural evolution), move < 	away from that awful Unix paradigm mentioned above "if yourG 	application is broken into several processes" ("Gee - let's just fork  = 	another process!!!").  You run with fewer processes and gain D 	the niceities that go with multi-threading (increased scalability).  B 	Guess those Mosix folks want to get cranking on migrating threads 	to CPUs on other nodes  ;-).   F 	But boy, process startup on other nodes sure sounds sexy, doesn't it?   				Rob    ------------------------------  # Date: Thu, 07 Aug 2003 20:38:49 GMT & From: jlsue <jefflsxxxz@sbcglobal.net>J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)8 Message-ID: <u6e5jvospa8dl03n4k973nkhcsmue7q573@4ax.com>  E On 7 Aug 2003 13:32:39 -0500, Kilgallen@SpamCop.net (Larry Kilgallen)  wrote:     > C >Without further explanation I don't see how restartable batch jobs : >together with generic queues don't handle that just fine. >   A Yeah, I was wondering the same thing... you beat me to the punch.   9 We've had job fail-over/restart for years in VMSclusters. 0 Checkpoint-restart logic is much older than VMS.   ------------------------------   Date: 7 Aug 2003 15:44:53 -0500 + From: young_r@encompasserve.org (Rob Young) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <rLpbtrW7WCdx@eisner.encompasserve.org>   c In article <Gfj017WkGiLe@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: c > In article <VIb6DP4xrsHW@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:    >>  N >> 9. Application/Resource Monitoring and Failover/Restart -- A subsystem thatR >> takes certain user-defined actions when a node fails, or an application stops.  >>   >> 	respawn? >>  ! >> 	cron:2:respawn:/usr/sbin/cron  >>   >> 	So VMS lacks respawn?  > D > Without further explanation I don't see how restartable batch jobs; > together with generic queues don't handle that just fine.  >    	Not the same.   $ show entry   3084   Entry  Jobname         Username     Blocks  Status4   -----  -------         --------     ------  ------7     308  HP_T4_V32       SYSTEM               Executing 1          On available batch queue NODE1_SYS$BATCH   < $ delete/entry=308  ! Gone for good, never resubmitted, etc.  A 	# kill -9 `ps -ef | grep cron | grep -v grep | awk '{print $2}'`   9 	and cron is whacked but inittab restarts it immediately.    				Rob    ------------------------------  $ Date: Thu, 7 Aug 2003 18:19:44 -0400* From: "Bill Todd" <billtodd@metrocast.net>J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)2 Message-ID: <NhOdnfmHdOAPTa-iXTWJig@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:PoS0CsLU0vzB@eisner.encompasserve.org...    ...   F > So a modern multi-threaded application (Oracle 9i multi-thread mode) > won't benefit.  J Of course it will:  individual multi-threaded instances can be distributed around the cluster.    > A > When a node in that case gets overloaded, you would dynamically 9 > migrate CPUs to the load (Galaxy) solving that problem.   K Don't confuse partitioned MP solutions (within which Galaxy features apply) H with clustering solutions - until such time as the hardware MP platforms? which support that kind of dynamic migration provide comparable B failure-isolation to clustering (including disaster-tolerance...).   > @ > As more applications become threaded (natural evolution), move= > away from that awful Unix paradigm mentioned above "if your G > application is broken into several processes" ("Gee - let's just fork  > another process!!!").   G You're close to a decade out of date, Rob.  Unix systems have supported L multiple process threads for a long time now, and people have learned how toH take advantage of them.  The Mosix comment applies more to being able toD spread the load over multiple nodes when one is no longer capable of handling the load adequately.   '   You run with fewer processes and gain E > the niceities that go with multi-threading (increased scalability).   K Fewer processes <> one process, Rob:  the two facilities are complementary. H One allows flexible use of UP and MP nodes, the other extends use across multiple such nodes.   > C > Guess those Mosix folks want to get cranking on migrating threads  > to CPUs on other nodes  ;-).  F I hope the smiley is an indication that you realize how silly that is.   - bill   ------------------------------  % Date: Thu, 07 Aug 2003 15:56:19 -0700 , From: Ken Fairfield <My.Full.Name@intel.com>J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)) Message-ID: <3F32D913.4FBCF297@intel.com>    Rob Young wrote: > W > In article <3F326FEE.9060709@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  > > Bob Ceculski wrote:  > >  > . > > For those that do not take this particularD > > rag because of its Unix roots, the link for more reading on this > > specific article is:< > > http://www.samag.com/documents/s=8817/sam0313b/0313b.htm > > G > > Bottom line, the Linux Open Source folks are attempting to tackle a L > > clustering issue that should have been handled long, long ago in a world > > called Digital...  > >  > E >         What on that punch list of 9 items can't or isn't performed A >         by VMS?  I don't see anything there that is not part of  >         VMS other than:  > M > 9. Application/Resource Monitoring and Failover/Restart -- A subsystem that P > takes certain user-defined actions when a node fails, or an application stops. >  >         respawn? > ' >         cron:2:respawn:/usr/sbin/cron  >  >         So VMS lacks respawn?  > H >         AIX with HACMP (failover clustering) does just that, start theI >         application up on the fallover node.  And boy can it by painful H >         (not work) if things aren't synchronized properly (filesystemsA >         for example).  So maybe the Linux crowd improves HACMP.   ? 	In another follow-up, Larry K. mentioned (checkpointed?) batch G jobs in generic queues.  I thought I'd mention an alternative technique H that a colleague dreamt up to monitor and (quickly) notify when any nodeD leaves the cluster (or hangs, etc.).  It could easily be extended toE (re)start applications/processes on another cluster node when a given  node goes down.   ; 	Assuming an existing cluster (bootstrapping is no problem  ? though), when node "NODEA" boots, as part of startup it runs a  E detached process that takes out an exclusive lock on a resource named H "NODEA_CLUSTER_MEMBER".  That process then uses SYSMAN to start detached@ processes on all other cluster members which each try to get theB exclusive lock on that resource...and "hang" waiting for an AST toE complete when the lock is granted (which is isn't as long as NODEA is @ a cluster member).  The detached process on NODEA, after settingD everything up, simply suspends itself, so minimal resources consumedE and lock remains held on NODEA until or unless it leaves the cluster, C killing the detached process and releasing the lock in the process.   < 	One of the "watchers" of NODEA on the other cluster membersA will be granted the lock.  Which member is arbritrary but doesn't C really matter: the process getting the lock then notifies "us" that  NODEA has left the cluster.   @ 	My point in relating this is that there are conceptually simpleB mechanisms already present in VMSclusters that allow for automatic8 "application failover" like that found in lesser cluster implementations.B It seems to me it wouldn't take a lot of engineering to put such aG facility into VMS itself: mostly packaging a toolset to allow customers  easy configuration.   8 	Is there anything in Rob's quoted text above that this  couldn't do?   	-Ken  --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfield " D1C Automation VMS System Support ) kenneth[dt]h[dt]fairfield[ta]intel[dt]com    ------------------------------  $ Date: Thu, 7 Aug 2003 19:35:13 -0400% From: "John Vottero" <John@mvpsi.com> J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)/ Message-ID: <vj5oh45drus5f6@news.supernews.com>   8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:rLpbtrW7WCdx@eisner.encompasserve.org... K > In article <Gfj017WkGiLe@eisner.encompasserve.org>, Kilgallen@SpamCop.net  (Larry Kilgallen) writes: 7 > > In article <VIb6DP4xrsHW@eisner.encompasserve.org>, - young_r@encompasserve.org (Rob Young) writes:  >  > >>K > >> 9. Application/Resource Monitoring and Failover/Restart -- A subsystem  thatL > >> takes certain user-defined actions when a node fails, or an application stops. > >>
 > >> respawn?  > >>" > >> cron:2:respawn:/usr/sbin/cron > >> > >> So VMS lacks respawn? > > F > > Without further explanation I don't see how restartable batch jobs= > > together with generic queues don't handle that just fine.  > >  >  > Not the same.  >  > $ show entry   3086 >   Entry  Jobname         Username     Blocks  Status6 >   -----  -------         --------     ------  ------9 >     308  HP_T4_V32       SYSTEM               Executing 3 >          On available batch queue NODE1_SYS$BATCH  > > > $ delete/entry=308  ! Gone for good, never resubmitted, etc. > B > # kill -9 `ps -ef | grep cron | grep -v grep | awk '{print $2}'` > : > and cron is whacked but inittab restarts it immediately. >   L That's absurd.  Entry 308 is gone because a suitably privileged user told itH to go away.  If that user had wanted it to abort and restart, they would
 have done:   $ STOP/QUEUE/REQUEUE/ENTRY=308   ------------------------------  % Date: Thu, 07 Aug 2003 19:21:23 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)' Message-ID: <3F32ED03.EFAD6040@fsi.net>    John Vottero wrote:  > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:rLpbtrW7WCdx@eisner.encompasserve.org... M > > In article <Gfj017WkGiLe@eisner.encompasserve.org>, Kilgallen@SpamCop.net  > (Larry Kilgallen) writes: 9 > > > In article <VIb6DP4xrsHW@eisner.encompasserve.org>, / > young_r@encompasserve.org (Rob Young) writes:  > >  > > >>M > > >> 9. Application/Resource Monitoring and Failover/Restart -- A subsystem  > thatN > > >> takes certain user-defined actions when a node fails, or an application > stops. > > >> > > >> respawn?  > > >>$ > > >> cron:2:respawn:/usr/sbin/cron > > >> > > >> So VMS lacks respawn? > > > H > > > Without further explanation I don't see how restartable batch jobs? > > > together with generic queues don't handle that just fine.  > > >  > >  > > Not the same.  > >  > > $ show entry   3088 > >   Entry  Jobname         Username     Blocks  Status8 > >   -----  -------         --------     ------  ------; > >     308  HP_T4_V32       SYSTEM               Executing 5 > >          On available batch queue NODE1_SYS$BATCH  > > @ > > $ delete/entry=308  ! Gone for good, never resubmitted, etc. > > D > > # kill -9 `ps -ef | grep cron | grep -v grep | awk '{print $2}'` > > < > > and cron is whacked but inittab restarts it immediately. > >  > N > That's absurd.  Entry 308 is gone because a suitably privileged user told itJ > to go away.  If that user had wanted it to abort and restart, they would > have done: >   > $ STOP/QUEUE/REQUEUE/ENTRY=308  @ I think a closer parallel would be to STOP/ID the queue manager.  F Actually, that would be even more disruptive because all the batch and5 print jobs would die as would the symbiont processes.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 7 Aug 2003 19:13:41 -0700 1 From: keithparris_NOSPAM@yahoo.com (Keith Parris) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)= Message-ID: <cf15391e.0308071813.63c7d354@posting.google.com>   ] Ken Fairfield <My.Full.Name@intel.com> wrote in message news:<3F32D913.4FBCF297@intel.com>...  > an alternative techniqueJ > that a colleague dreamt up to monitor and (quickly) notify when any node& > leaves the cluster (or hangs, etc.).  B (I don't think your colleague's technique works when a node hangs,8 unless it leaves the cluster as a result of the hang...)  ; Has your colleague heard of the system services $SETCLUEVT, A $TSTCLUEVT, and $CLRCLUEVT?  They allow one to ask that an AST be 2 delivered when a node leaves or joins the cluster.  F These system services were added at the request of the Rdb folks.  TheB idea was that the AST from this system service was to be deliveredD before any completion ASTs for pending lock requests.  That way, theD database software would know, for example, that a node had died, andD that a lock request about to be granted was likely granted because aC process on another node had gone away with the node (and might have A left the shared resource protected by the lock in a state needing B recovery or cleanup work), rather than just being granted the lock@ because someone else was done with it and had left it in a clean state.   ------------------------------   Date: 7 Aug 2003 22:08:45 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <Ca8WaXIdXWG$@eisner.encompasserve.org>   a In article <rLpbtrW7WCdx@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: e > In article <Gfj017WkGiLe@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: d >> In article <VIb6DP4xrsHW@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: >  >>> O >>> 9. Application/Resource Monitoring and Failover/Restart -- A subsystem that S >>> takes certain user-defined actions when a node fails, or an application stops.   >>> 
 >>> 	respawn?  >>> " >>> 	cron:2:respawn:/usr/sbin/cron >>>  >>> 	So VMS lacks respawn? >>  E >> Without further explanation I don't see how restartable batch jobs < >> together with generic queues don't handle that just fine. >>   >  > 	Not the same. >  > $ show entry   3086 >   Entry  Jobname         Username     Blocks  Status6 >   -----  -------         --------     ------  ------9 >     308  HP_T4_V32       SYSTEM               Executing 3 >          On available batch queue NODE1_SYS$BATCH  > > > $ delete/entry=308  ! Gone for good, never resubmitted, etc. > C > 	# kill -9 `ps -ef | grep cron | grep -v grep | awk '{print $2}'`  > ; > 	and cron is whacked but inittab restarts it immediately.   9 Certainly there is _some_ unix action to stop it forever, 4 just as there is some VMS action to stop it forever.9 The point is that a system crash will not prevent the job 4 from restarting (even though an operator can do so).   ------------------------------   Date: 7 Aug 2003 22:48:17 -0500 + From: young_r@encompasserve.org (Rob Young) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <4JLU8B7w$6KU@eisner.encompasserve.org>   _ In article <NhOdnfmHdOAPTa-iXTWJig@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:  > : > "Rob Young" <young_r@encompasserve.org> wrote in message/ > news:PoS0CsLU0vzB@eisner.encompasserve.org...  >  > ...  > G >> So a modern multi-threaded application (Oracle 9i multi-thread mode)  >> won't benefit.  > L > Of course it will:  individual multi-threaded instances can be distributed > around the cluster.   ? 	Where, which project?  Not the openMosix project I referenced. ; 	Not trying to pull an Andrew and talk about something else 	 	are you?    >                          >>B >> When a node in that case gets overloaded, you would dynamically: >> migrate CPUs to the load (Galaxy) solving that problem. > M > Don't confuse partitioned MP solutions (within which Galaxy features apply) J > with clustering solutions - until such time as the hardware MP platformsA > which support that kind of dynamic migration provide comparable D > failure-isolation to clustering (including disaster-tolerance...). >   > 	You like to bandy about that word confused.  Read on, I think* 	you will quickly realize who is confused.   >>A >> As more applications become threaded (natural evolution), move > >> away from that awful Unix paradigm mentioned above "if yourH >> application is broken into several processes" ("Gee - let's just fork >> another process!!!"). > , > You're close to a decade out of date, Rob.   	No.   >  Unix systems have supportedN > multiple process threads for a long time now, and people have learned how to > take advantage of them.   E 	Yes.  But the Unix paradigm is to fork a process.  Done a ps -ef on	 7 	a moderately loaded Unix server and notice how many of @ 	the processes are the same?  It is a great deal about processes 	starting processes.  1 > The Mosix comment applies more to being able to F > spread the load over multiple nodes when one is no longer capable of > handling the load adequately.    @ 	Right.  But again, it is about processes.  Go back and RE-READ	 	THE PARAGRAPH I QUOTED!!!!   = 	The openMosix project is about ---processes--- being started ; 	on another node when the original node is overloaded.  See 3 	the reference and the paragraph I quoted from that B 	one.  In fact, there is another project/paper/research effort to  	overcome that limitation:  - http://jazzalbums.net/maask/Migshm_Report.pdf   J The aim of the project is to enable migration of processes that explicitlyN share memory and migration of multi-threaded applications, without any loss ofO performance and normal operation ensured, on an openMosix cluster. openMosix is O an extension to the Linux kernel which provides a cluster computing environment M with automatic load balancing among nodes of the cluster. Migration of shared K memory processes is not handled in the current versions of openMosix. Hence M applications like web servers, database servers, certain rendering tools that H use shared memory cannot benefit from the openMosix cluster. Our project' overcomes this limitation of openMosix.     ? 	Hey, did you catch the reference to overcoming the limitations ' 	regarding multi-threaded applications?                     >  >>D >> Guess those Mosix folks want to get cranking on migrating threads >> to CPUs on other nodes  ;-).  > H > I hope the smiley is an indication that you realize how silly that is. >   ? 	Yes!  I did and I do know why it is silly.  But MAYBE now YOU   	know why it is silly?   				Rob    ------------------------------   Date: 7 Aug 2003 23:49:13 -0500 + From: young_r@encompasserve.org (Rob Young) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <$6dbACgwnJC1@eisner.encompasserve.org>   c In article <Ca8WaXIdXWG$@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: c > In article <rLpbtrW7WCdx@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: f >> In article <Gfj017WkGiLe@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:e >>> In article <VIb6DP4xrsHW@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:  >>   >>>>  P >>>> 9. Application/Resource Monitoring and Failover/Restart -- A subsystem thatT >>>> takes certain user-defined actions when a node fails, or an application stops.  >>>>   >>>> 	respawn? >>>>  # >>>> 	cron:2:respawn:/usr/sbin/cron  >>>>   >>>> 	So VMS lacks respawn?  >>> F >>> Without further explanation I don't see how restartable batch jobs= >>> together with generic queues don't handle that just fine.  >>>  >>   >> 	Not the same.  >>   >> $ show entry   308 7 >>   Entry  Jobname         Username     Blocks  Status 7 >>   -----  -------         --------     ------  ------ : >>     308  HP_T4_V32       SYSTEM               Executing4 >>          On available batch queue NODE1_SYS$BATCH >>  ? >> $ delete/entry=308  ! Gone for good, never resubmitted, etc.  >>  D >> 	# kill -9 `ps -ef | grep cron | grep -v grep | awk '{print $2}'` >>  < >> 	and cron is whacked but inittab restarts it immediately. > ; > Certainly there is _some_ unix action to stop it forever, 6 > just as there is some VMS action to stop it forever.; > The point is that a system crash will not prevent the job 6 > from restarting (even though an operator can do so).  7 	Depends.  Swapper can't be stopped.  But the point not 9 	to be overlooked by David Dachtera is someone might stop : 	the queue /RESET and Poof!  Several of us have made typos 	on that command.   = 	The respawn mechanism is pretty slick in that if the process 6 	falls off the system it automatically gets restarted.   				Rob    ------------------------------   Date: 7 Aug 2003 17:24:04 -0700 # From: dooleys@snowy.net.au (dooley) 2 Subject: Re: Computer Systems & Automation Manager= Message-ID: <1ca82fc6.0308071624.70fca908@posting.google.com>   a Paul Repacholi <prep@prep.synonet.com> wrote in message news:<87k79qcwte.fsf@prep.synonet.com>... ? > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:  > g > > In article <1ca82fc6.0308050008.292ed469@posting.google.com>, dooleys@snowy.net.au (dooley) writes: 	 > >> Phil F > >> http://jobs.careerone.com.au/search/dsp_show_job.cfm?AD_ID=229537 >   * > >   So how is Port Kembla in the spring? > F > Spring? Kembla does not have spring... Standard is rain/drizle/storm > wit coal dust.  ; All things are relative, have you ever been to Port Talbot? 4 > At least they are closing down the smelter, so the1 > copper content of everything should drop some.   >  > The steel works is it?# Yes, one of the few vms sites left. ? I think at one time Woolongong used to have it's own decus lug. ; Kembla Coal & Coke, Illawarra Electricity, and a lot of the ( "heavy industry" were big dec customers. Phil   ------------------------------   Date: 7 Aug 2003 16:59:53 -0500 4 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)* Subject: Re: DS-35x-FA Dip switch settings3 Message-ID: <g0WkVHC3lCmG@eisner.encompasserve.org>   _ In article <CIEJLCMNHNNDLLOOGNJIEEMOHLAA.tom@kednos.com>, "Tom Linden" <tom@kednos.com> writes: 9 > This is the personality module for a BA356, but I can't > > locate any manual that will give the settings.  Anybody have* > a URL ref or have the settings?  tia Tom  C The dip switches are used for addressing.  If all of them are off,  D then the disks in the shelf show up as ID0 through ID6.  If you haveB one, two and three ON, then the disks show up as ID8 through ID14.  C One other thing to note is that if the SCSI cable that connects the F shelf to the system is attached to the front connector it will disable? the terminator in the shelf.  If you want the shelf to provide  B termination, hook up your cable to the top connector (horizontal).   ------------------------------   Date: 07 Aug 2003 17:58:09 GMT2 From: Thierry Dussuet <thierry@squeeeez.no-ip.com> Subject: Re: DSSI problem 1 Message-ID: <slrnbj54pi.q8g.thierry@VENUS.Family>   C In article <87el04kmmz.fsf@prep.synonet.com>, Paul Repacholi wrote: ; >>> You didn't knock either of the terminators off did you? 7 >> Do you know where the terminators are on a VAX 4000?  > B > On the silver panel for the CPU, there are 2 `Honda' connectors,? > one at the bottom left corner, another a bit above and right. @ > These need aither a cable to a storage box or other system, orC > a terminator. DSSI terminators have a largish window with a green  > LED under them.   F Oh... I got the box with only one terminator on it.  IIRC, it was in aG DSSI cluster before, so there must have been a cable and another VAX... @ But it worked like that for quite a long time, why did it changeC suddenly?  Is there some way to let the VAX believe that there is a A terminator on there, too? (the 'Honda' is on the upper right one)   F > There should also be one on the Honda at the extreme left end of the+ > system, past the Q-bus slots. Or a cable.   E Nothing here, too... Just a red plastic thing on the centronics port,   but the upper one has nothing...  H > If you have a 400RX(??) box and only one CPU, the terminator will liveH > on the `other' connector on the storage box. They will be micro-telcos > though, not Hondas.   E No, there is only the CPU box, no storage box.  The 3 DSSI hard disks ) and the TK70 are on the upper part of it.    Thierry    ------------------------------  % Date: Thu, 07 Aug 2003 21:02:57 +0100 + From: John Laird <john@laird-towers.org.uk> = Subject: Re: How many errors are too many when making CD-R's? 8 Message-ID: <uvb5jvgimgrkbquf9mkde139nn5l55r6oj@4ax.com>  F On 7 Aug 2003 10:04:44 -0700, spamsink2001@yahoo.com (Alan E. Feldman) wrote:  l >John Laird <john@laird-towers.org.uk> wrote in message news:<cpv3jvgh47m0nsoiai2en47nnqh9k661sa@4ax.com>... >>M >> I think I would certainly experiment with a few different brands if I weree: >> you.  None are what you could call "expensive" anyway ! >lE >Well, I was told by my dept. that I would only get the stack versione
 >due to cost.!  5 Writing at 1x, how many can you be getting through ?!o  L If the difference is 50 cents against a dollar, your department must place a poor value on the data content.    --   John   ------------------------------  # Date: Thu, 07 Aug 2003 20:36:37 GMTp& From: jlsue <jefflsxxxz@sbcglobal.net>= Subject: Re: How many errors are too many when making CD-R's?o8 Message-ID: <m2e5jv0davn5g67k2l7irl78j3tvmoqhbv@4ax.com>  F On 7 Aug 2003 10:04:44 -0700, spamsink2001@yahoo.com (Alan E. Feldman) wrote:     > E >I will compare disk reading with both sets of CD-R's and report backpD >here. But this new large error rate is a relatively new phenomenon.C >Either the new disks are bad or the drive has just recently gotten' >dirty.o  J Just out of curiosity... are you labelling these disks with anything prior to comparing them?   ------------------------------  # Date: Fri, 08 Aug 2003 01:01:32 GMTa' From: nospam <x@wedontwantyourspam.com>- Subject: Re: HP FUDBusting5 Message-ID: <BB5933E2.17DD5%x@wedontwantyourspam.com>r  @ in article AHUSa.503$4Q.409@news.cpqcorp.net, Fred Kleinsorge at: my-last-name@stardotzko.dec.com wrote on 22/07/2003 02:40:  E > Also add a very large grain of salt to anything your read by a verya > disgruntled ex-DEC employee. > I > My bias is upfront and clear:  I am a VMS developer employed by HP.  MyiJ > opinion and advice for VMS is that it has a long term future here in HP.K > From the bottom up, I have experienced nothing but positive comments fromyF I've spoken with people with that attitude before. Sitting high in theL foliage of there tree with a nice clear view of the future only too see themL start hit the ground with an unexpected chain saw. The history of vendors isL not a pretty picture and I trust nothing much any more. Though I hope things do continue as you see them.   ------------------------------  $ Date: Thu, 7 Aug 2003 16:25:01 -04000 From: "Brian Tillman" <Tillman@sparkingwire.com>4 Subject: Re: Idiot (I know that's redundant) spammer$ Message-ID: <3f32b5a0$1@news.si.com>   >My wife showed meC >how to copy a group of files.  What a fucking pain in the ass thatrC >is.  She said you click on one and then use the shift key to clicke >on others you want.  OK.   H If this is true, then your wife showed you incorrectly.  You don't clickI with a shift key; you don't click with ANY key.  You can hold a key whilea6 clicking a mouse button, but that's not what you said.  J To select a group of contiguous file names, click on the first one.  Then,H while holding down the shift key, click on the last one.  They'll all beK selected, just like in DECwindows.  To select a discontiguous set of files,bK click the first one.  Then, whie holding down the Control key, click on the A ones you want.  They'll all be selected, just like in DECwindows.   F You can also click near the first obkect you wish to select andm whileK holding down the mouse button, drag the pointer to the last object you wishrH to select.  While you drag, the intervening files will be selected. just like in DECwindows.   I Surely your experience with DECwindows isn't that hard to carry over to as PC.  -- ,I Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot comw5 Smiths Aerospace  Addresses modified to prevent SPAM.oD 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991h8        This opinion doesn't represent that of my company   ------------------------------  * Date: Thu, 7 Aug 2003 18:20:08 +0000 (UTC) From: david20@alpha1.mdx.ac.uk Subject: Re: Netware is no VMS) Message-ID: <bgu58o$icb$1@news.mdx.ac.uk>c  U In article <3F327E6D.5080402@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:e
 >jlsue wrote:h >aK >>On Wed, 06 Aug 2003 10:33:10 -0700, "Barry Treahy, Jr." <Treahy@MMaz.com>o >>wrote: >>   >>H >>>A very interesting Linux project, which is underway, is a clustering I >>>project that creates a unified process space between cluster nodes so fF >>>that true load balancing and failover can occur when a member node K >>>either slows or dies...  There is no question that VMSclusters has lead  H >>>the way in clustering, but it was never completed to what would be a K >>>logical conclusion; Process failover/recovery and true load-balancing...- >>>- >>>    a >>>  >>K >>This is really cool, until you realize that the process that just crashed:? >>system1 got failed over to system2 (then system3, system4...)) >>H >Bah, non-sense!  You must be talking about Windows, because this isn't / >an issue with VMS nor most flavors of Unix... e >.G >Think about it, when is the last time you had a rogue VMS application yE >take down a VMS system?  With the exception of a faulty driver that  G >couldn't handle unexcepted device failures, I have not had a software  G >based VMS failure for over a decade and that was Kernel code!  As for  J >user-mode code crashing on VMS, I can't think of much except perhaps EDT @ >on VMS 1.0 but that was all RSX based stuff...  In the case of I >Supervisor or Executive code, if the code faults, it kills the process,   >not the system... >  >Barry >   B So it was in Kernel mode when it crashed. Are you saying that only; processes which NEVER go into Kernel mode can be migrated. h  M Process runs application in user mode. Application does something which meansyB it executes some code in Kernel mode and then crashes the system. < Process is migrated to second system at its last checkpoint.O Process continues. Application executes its code in Kernel mode again - crashes  second system.  L Just take a look at the ECOs for DEC TCPIP services. There are usually a fewK fixes for inappropriate usage of NFS, FTP or other parts of the TCPIP suitea7 which in certain rare circumstances can crash a system.i   eg  % ECO B   30-APR-2002     Alpha and VAXe           Problem:  F         INVEXCEPTN crash in TCPIP$INTERNET_SERVICES due to inadvertentE         freeing of a BG device for an active TN device kernel client.n           Deliverables:b  8         TCPIP$INTERNET_SERVICES.EXE             V5.3-18B           Reference:  5         PTR 70-5-1928 / CFS.88690 / Req Id: STLQC0001          TCPIP_BUGS Note 25725         PTR 70-5-1947 / CFS.89075 / Req Id: HPAQC22TZs            
 David Webb VMS and Unix team leader CCSS Middlesex University d   ------------------------------  $ Date: Thu, 7 Aug 2003 17:43:01 -0400* From: "Bill Todd" <billtodd@metrocast.net> Subject: Re: Netware is no VMS2 Message-ID: <mHadnXH7xthnWq-iXTWJkw@metrocast.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:KkMXA$PuHud9@eisner.encompasserve.org...o@ > In article <nY6cndhuz9ESnq-iXTWJiQ@metrocast.net>, "Bill Todd"  <billtodd@metrocast.net> writes: > >i >s > > L http://groups.google.com/groups?selm=rdeininger-2202030718590001%40user-uinj4 > > 4lr.dialup.mindspring.com&oe=UTF-8&output=gplain > >>C > >> In article <G8ednVIlc4atTsujXTWcpQ@metrocast.net>, "Bill Todd" $ > >> <billtodd@metrocast.net> wrote: > >>I > >> >> I doubt enhancements to TPU will get higher priority than, say, ao newr > > file > >> >> system.f > >> >I > >> >And since plans for a new file system seem to have pretty much been I > >> >languishing for over three years now, with no obvious progress (nott even > > acE > >> >mention on the roadmap through early 2005), well, draw your owni > > conclusions. > >>E > >> What makes you think the file system has been languishing?  Justg becauseb > >> nothing has shipped?n > >oJ > > Don't be a dolt, Rob:  not only has nothing shipped, but there's not a damnL > > thing about it on the pitiful VMS roadmap you've been touting as 'proof'' > > that development is alive and well.o > >  >h@ > Careful - follow the links.  The Google reference holds a key: >n > selm=rdeiningeri >y= > Those are his comments.  Someone on the inside looking out.g  % Obviously with the usual blinders on.      So= > I'm not a dolt - you just aren't being very careful in your? > criticism.  H Or one could suggest that you're not sufficiently careful to distinguishD your own words from what you choose to cut-and-paste from elsewhere.  L If the question did not come from you, then it appears that you responded toI the statement of mine with absolutely nothing save irrelevant quotes from E elsewhere.  To refresh your memory, the interchange to which you weret (theoretically) responding was   <quote>   I > What new features do you believe are needed above what's already in the 
 > roadmap?  J How about that new file system that VMS developers have been talking about for 3 - 4 years now?   </quote>  G My response above seems pretty straight-forward:  it mentions a feature K which VMS development itself has said it needs and which does not appear injH the roadmap.  The two items you cited in your response in no way address that.D  C Back in February I responded to Deininger's assertion about VMS noteG scheduling work that far ahead with the observation that a year earlier L VMS's roadmap had stretched a year farther into the future (i.e., that afterI a year no specific extensions had been made to the year-earlier roadmap - I and the same continues to be true today, now after 18 months).  He didn't  have an answer for that one.   >n > >> > >>L http://groups.google.com/groups?selm=v5kru8khbin516%40corp.supernews.com&oe= > > UTF-8&output=gplainh > >> > >> re: new file system > >>J > >> At our last LUG meeting here in Chicago, we heard that a new VMS file
 > > systemJ > >> is being worked on.  There were lessons learned the last time around. > > J > > As I noted above, it 'was being worked on' back in late 1999 (not evenL > > starting from scratch, but leveraging an already-complete code base from anF > > earlier effort).  That gives you a pretty good idea of the kind of priority > > it's been given. > >g >  > How do you know?  J Because after nearly four years it not only hasn't shipped but hasn't evenG appeared on the roadmap, idiot.  Not even in the explicitly uncommitted0C 'investigations' slated for 2005 - six years after conversion of an:< *existing product* was supposedly already being 'worked on'.  1   How do you know there isn't considerable designe; > effort underway right now?  Should they make that public?e  I What the hell use is their 'roadmap' supposed to be if they *don't*?  ItstJ not as if anyone outside the existing customer base were going to react inE any competitive way to the information:  it's only of interest to thed existing base.     WhatD > did the public know about Galaxy prior to November 1997?  I'm sureG > over its history there were dozens of secret VMS development efforts.f  D Perhaps where disclosure of truly novel architecture was threatened.I Preannouncing a planned ship date for a new file system hardly qualifies: F hiding it would be just plain stupid - and certainly inconsistent with traditional roadmap policies.d   >eB > In this case, it may be very necessary to be totally quiet aboutE > the development.  Especially if like Galaxy a number of key patentsoE > would be applied for.  Are patents going to be applied for?  Maybe,eB > maybe not.  But the point is it isn't uncommon to keep something$ > that is very important underwraps.  K In your dreams, Rob.  VMS isn't a leader in this area any more, nor does HPeK show any evidence whatsoever of wanting to change that situation.  If a newlK file system eventually does appear, there's every reason to believe that itaI will indeed be a shoestring effort aimed at patching over inadequacies in K the existing one rather than something that breaks any new ground (at least D any ground new to the industry as a whole, rather than just to VMS).   - bill   ------------------------------   Date: 8 Aug 2003 02:00:38 GMTu From: healyzh@aracnet.com  Subject: Re: WASD and RDB?, Message-ID: <bgv08602erq@enews1.newsguy.com>  M "Alan Winston - SSRL Admin Cmptg Mgr" <winston@ssrl.slac.stanford.edu> wrote:o3 >>I'll have to try and find Rdb modules for Perl.  -  P > You want DBD::Rdb, but for some reason I've only been able to download it when$ > I found it by author than by name.  ) > http://search.cpan.org/author/ASTILLER/M  	 Thanks!  D  K BTW, while looking at the WASD website I was reminded that I have a copy ofEK your book, so I broke it out.  I must say, it's been one of the most usefule  books on OpenVMS I've purchased!   		Zane   ------------------------------   End of INFO-VAX 2003.435 ************************ on this > > specific article is:< > > http://www.samag.com/documents/s=8817/sam0313b/0313b.htm > > G > > Bottom line, the Linux Open Source folks are attempting to tackle a L > > clustering issue that should have been handled long, long ago in a world > > called Digital...  > >  > E >         What on that punch list of 9 items can't or isn't performed A >         by V2.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 accepted.  <<< PORT 147,162,156,46,200,346 >>> 200 Port 200.34 at Host 147.162.156.46 acc