1 INFO-VAX	Sat, 09 Aug 2003	Volume 2003 : Issue 437       Contents: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)  Changing group Re: Changing group Re: Changing group Re: Changing group Re: Changing group Re: Changing group Re: HP FUDBusting  Re: HP FUDBusting  Re: HP FUDBusting  Re: HP FUDBusting ( Re: Licence manager and class scheduling Re: mount count  Re: mount count  Re: mount count  Re: mount count  Re: mount count  Re: mount count  Re: Netware is no VMS  Re: Netware is no VMS  Re: Netware is no VMS  Re: Netware is no VMS  Re: Netware is no VMS  Re: Netware is no VMS ' Re: OpenVMS Advanced Technical Bootcamp P Re: preventing DCL delete command from warning that a file does not exist existe Re: TCP/IP feature request Re: TDMS usage question  Re: TDMS usage question  Re: TDMS usage question  TPU Section files... Re: TPU Section files... Re: TPU Section files...4 Re: Using TCPIP$SMTP_SEND_FROM_FILE in a FORTRAN app  F ----------------------------------------------------------------------   Date: 8 Aug 2003 13:23:28 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) J Subject: Re: Can VMS still be improved?  Yes!  (Was Re: Netware is no VMS)3 Message-ID: <Z0ipvpGKl6PP@eisner.encompasserve.org>   Z In article <$Hthi6wS8F1O@kuhub.cc.ukans.edu>, paul@kuhub.cc.ukans.edu (Craig Paul) writes:J > Hmm, I see a mention of cache coherency in distributed file locks, but I/ > see no mention of distributed record locking.   H    How can you lock records in a file system that doens't have anything?  B    I saw two cluster-wide file systems.  I would expect them to doF    something which allows one to access a single file in a coordinatedD    manner, even NFS does that, but not necessarily mandatory locking1    since normal UNIX file access doesn't do that.    ------------------------------   Date: 8 Aug 2003 14:35:59 -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.0308081335.1c238691@posting.google.com>   ] Ken Fairfield <My.Full.Name@intel.com> wrote in message news:<3F33D7DC.3011C531@intel.com>... 0 > Does Shadow_Server use the $xxxCLUEVT services6 > you've listed, or does it function at a lower level?  7 John "AtoZ" Andruszkiewicz kindly provided this answer:   F "The Shadow Server has nothing  directly - to do with the events that surround node failure.  B The SHDriver has a fork level interface to the connection manager,C which notifies the SHDriver when a system departs the cluster.  The E CSID and the CNID of the departed system is presented at that time to D the SHDriver.  Since this notification is GUARANTEED to occur before= any locking operations are resumed, it is used by SHDriver to > determine if the departed system had that shadow set mounted."  < > How early were these added to VMS, i.e., what VMS version?  5 V6.2.  So they're relatively recent additions to VMS.    ------------------------------  # Date: Fri, 08 Aug 2003 18:48:38 GMT # From: "John N." <JNixon@cfl.rr.com>  Subject: Changing group = Message-ID: <agSYa.19808$qg3.1355009@twister.tampabay.rr.com>   K Will it be dangerous to give an in-house application permission to create a D captive account which would  execute a utility to change their GROUPJ membership.  We have several uic groups set up which each have a differentH application version under testing.  We are looking for good ways for theL testers to jump from group to group without having several accounts for each user.    ------------------------------   Date: 8 Aug 2003 14:43:16 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Changing group 3 Message-ID: <pAw$Jlip$fl5@eisner.encompasserve.org>   c In article <agSYa.19808$qg3.1355009@twister.tampabay.rr.com>, "John N." <JNixon@cfl.rr.com> writes: M > Will it be dangerous to give an in-house application permission to create a F > captive account which would  execute a utility to change their GROUP
 > membership.   E It would be particularly dangerous because it would not fully emulate " an actual login to that UIC group.  ? > We have several uic groups set up which each have a different J > application version under testing.  We are looking for good ways for theN > testers to jump from group to group without having several accounts for each > user.   F Accounts are cheap, and the use of separate login directories prevents, the users from tromping all over each other.   ------------------------------   Date: 8 Aug 2003 15:45:31 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: Changing group 3 Message-ID: <CAubBZU17J5H@eisner.encompasserve.org>   c In article <agSYa.19808$qg3.1355009@twister.tampabay.rr.com>, "John N." <JNixon@cfl.rr.com> writes: M > Will it be dangerous to give an in-house application permission to create a F > captive account which would  execute a utility to change their GROUPL > membership.  We have several uic groups set up which each have a differentJ > application version under testing.  We are looking for good ways for theN > testers to jump from group to group without having several accounts for each > user.   6    Don't use groups for this.  Use rights identifiers.   ------------------------------  % Date: Fri, 08 Aug 2003 20:22:11 -0400 - From: Jay Newman <fredthecat@i-love-cats.com>  Subject: Re: Changing group 8 Message-ID: <bvd8jvc4vehke2lvhtcvdlv9ocr5d3idmn@4ax.com>  F On Fri, 08 Aug 2003 18:48:38 GMT, "John N." <JNixon@cfl.rr.com> wrote:  L >Will it be dangerous to give an in-house application permission to create aE >captive account which would  execute a utility to change their GROUP K >membership.  We have several uic groups set up which each have a different I >application version under testing.  We are looking for good ways for the M >testers to jump from group to group without having several accounts for each  >user. >  >   K Is this a production environment? If so, I would strongly recommend against N separating production from testing just through changing group membership. OneP small oversight (i.e. something being defined at system or user level instead ofN just in the group tables,) and you may accidentally hit a production area when you did not intend to.M Even if it is just a group of test environments,  is it worth skimping on the O number of user accounts when you increase the risk of "collisions" between your  various applications?   H Also, my understanding is that enabling users to change group membershipL requires granting CMKRNL privilege. This can be abused by  people who shouldN know better, using freeware tools such as "become" to hide their tracks and/or! pretend to be the SYSTEM account. O If your utility was an image that you installed in memory with cmkrnl privilege N rather than give any users cmkrnl, and it was able to nothing else than changeL group membership, you might avoid that. However, I would still not rest easyD over the increased chance of working with the wrong data & programs.  I Just my humble opinion, based on the headaches I experienced in a similar  situation in the past.       ___________________    Regards,
 Jay Newman Systems Programmer The Pepsi Bottling Group Mississauga, Ontario, Canada   ------------------------------   Date: 8 Aug 2003 22:17:06 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: Changing group 3 Message-ID: <f2iVWofCeK5V@eisner.encompasserve.org>   h In article <bvd8jvc4vehke2lvhtcvdlv9ocr5d3idmn@4ax.com>, Jay Newman <fredthecat@i-love-cats.com> writes:H > On Fri, 08 Aug 2003 18:48:38 GMT, "John N." <JNixon@cfl.rr.com> wrote: > M >>Will it be dangerous to give an in-house application permission to create a F >>captive account which would  execute a utility to change their GROUPL >>membership.  We have several uic groups set up which each have a differentJ >>application version under testing.  We are looking for good ways for theN >>testers to jump from group to group without having several accounts for each >>user.  >> >> > M > Is this a production environment? If so, I would strongly recommend against L > separating production from testing just through changing group membership.  G In serious business settings, production and test should not even be on  the same cluster.   H But even in a properly separate testing environment, one should not relyD on little tricks for testing that are not used in actual production.   ------------------------------  % Date: Fri, 08 Aug 2003 21:36:49 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Changing group ' Message-ID: <3F345E41.349AD611@fsi.net>    Bob Koehler wrote: > e > In article <agSYa.19808$qg3.1355009@twister.tampabay.rr.com>, "John N." <JNixon@cfl.rr.com> writes: O > > Will it be dangerous to give an in-house application permission to create a H > > captive account which would  execute a utility to change their GROUPN > > membership.  We have several uic groups set up which each have a differentL > > application version under testing.  We are looking for good ways for theP > > testers to jump from group to group without having several accounts for each	 > > user.  > 8 >    Don't use groups for this.  Use rights identifiers.  @ It's very likely that application logicals are segregated by theD LNM$GROUP_nnnnnn table. That is, each group-level LNM table has LNMSG that point to a different place (different disk, directory tree, etc.).   E To answer the OP's question, any time you try circumvent the system's G security archtiecture you're going to be introducing the possibility of  trouble.  G That said, one could use SET UIC (no longer documented, dunno if it was F ever supported) along with some clever use of the LNM$FILE_DEV logicalG name to simulate a persona change to some degree. Liberal use of GRPPRV G could be useful in such a scenario, so long as those deploying the app. B understand that production users should have only those privilegesF needed to make the application work as documented, and should not have# the same priv.'s as the developers.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 8 Aug 2003 15:46:42 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: HP FUDBusting3 Message-ID: <jmGUDWuDIO+s@eisner.encompasserve.org>   Y In article <bh1084$1k1n$1@jeeves.eng.abbnm.com>, peter@abbnm.com (Peter da Silva) writes:  > D > Things like non-executable stacks are common to a lot of operatingE > systems, including UNIX variants... even some fairly old UNIX ones.   F    Non-executable stacks require underlying hardware support no matter?    what OS you put on the system.  Few processors provide this.    ------------------------------  * Date: Fri, 8 Aug 2003 20:12:52 +0000 (UTC)& From: peter@abbnm.com (Peter da Silva) Subject: Re: HP FUDBusting0 Message-ID: <bh1084$1k1n$1@jeeves.eng.abbnm.com>  6 In article <20030729151558.15837.qmail@gacracker.org>,6 Doc.Cypher  <Use-Author-Address-Header@[127.1]> wrote:O > My (somewhat limited) understanding was that things like the stack protection G > measures were sufficiently similar to those in VMS that they could be A > described as separate implementations of very similar concepts.    Um.   B Things like non-executable stacks are common to a lot of operatingC systems, including UNIX variants... even some fairly old UNIX ones.    --  O I've seen things you people can't imagine. Chimneysweeps on fire over the roofs O of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All L these things will be lost in time, like chalk-paintings in the rain.   `-_-'K Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`    ------------------------------  * Date: Fri, 8 Aug 2003 21:56:34 +0000 (UTC)& From: peter@abbnm.com (Peter da Silva) Subject: Re: HP FUDBusting0 Message-ID: <bh16ai$1pjj$4@jeeves.eng.abbnm.com>  3 In article <jmGUDWuDIO+s@eisner.encompasserve.org>, : Bob Koehler <koehler@eisner.aspm.encompasserve.org> wrote:I > In article <bh1084$1k1n$1@jeeves.eng.abbnm.com>, peter@abbnm.com (Peter  > da Silva) writes: F > > Things like non-executable stacks are common to a lot of operatingG > > systems, including UNIX variants... even some fairly old UNIX ones.   H >    Non-executable stacks require underlying hardware support no matterA >    what OS you put on the system.  Few processors provide this.   D Few modern RISC processors, maybe, but is that a cause or an effect?  J After all, the PDP-11 happens to be one that does. Split I&D and pure textD code in V7 and 2BSD had non-executable stacks by default... but 4BSDG didn't, and running 4BSD is what most RISCs seem to have been optimised  for.   --  O I've seen things you people can't imagine. Chimneysweeps on fire over the roofs O of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All L these things will be lost in time, like chalk-paintings in the rain.   `-_-'K Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`    ------------------------------   Date: 8 Aug 2003 22:18:02 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: HP FUDBusting3 Message-ID: <lJmU0a8eydra@eisner.encompasserve.org>   Y In article <bh1084$1k1n$1@jeeves.eng.abbnm.com>, peter@abbnm.com (Peter da Silva) writes: 8 > In article <20030729151558.15837.qmail@gacracker.org>,8 > Doc.Cypher  <Use-Author-Address-Header@[127.1]> wrote:P >> My (somewhat limited) understanding was that things like the stack protectionH >> measures were sufficiently similar to those in VMS that they could beB >> described as separate implementations of very similar concepts. >  > Um.  > D > Things like non-executable stacks are common to a lot of operatingE > systems, including UNIX variants... even some fairly old UNIX ones.   I And on Solaris one must specifically disable it for certain applications.    ------------------------------  # Date: Sat, 09 Aug 2003 02:26:35 GMT " From: bugs@pu.net (Mark Hittinger)1 Subject: Re: Licence manager and class scheduling . Message-ID: <vZYYa.102289$YN5.70882@sccrnsc01>  + jbrankin@ntlworld.com (Jim Brankin) writes: B >I have a suggestion for an enhancement to VMS. How about coupling1 >class based scheduling with the licence manager.   J While class based schedulers can divvy up cpu time for a user process theyD don't do a very good job on the system overhead caused by a process.  I A process using file locks, performing interprocess communication, doing  J hard page faults, or queuing large block i/o's will tend to defeat a class; scheduler as system call overhead isn't throttled by class.   K Larger and more expensive software that could use such a license model, say L oracle, could easily consume the other half of the cpu in overhead.  SmallerI software that isn't resource intensive would be a better target but would * probably be far cheaper to license anyway.  J Class schedulers tend to suck and license managers tend to suck.  It couldA be a marriage made in DEC heaven :-)  but it could never last :-)    Later    Mark Hittinger bugs@pu.net    ------------------------------  # Date: Fri, 08 Aug 2003 17:41:44 GMT > From: Michael Austin <maustin@no-more-spam.firstdbasource.com> Subject: Re: mount count6 Message-ID: <shRYa.20$zb.5@newssvr23.news.prodigy.com>   amit sawhney wrote:    > Hi8 > WHen I issue a SHOW DEVICE DSA?? I notice that the the8 > mount count is 2. What does this mean? How can I check& > the details? Could this be an error?  G exactly what it says... it is mounted twice. If you do a "sh cluster",  G odds are that you have a 2 node cluster and it is mounted once on each   node...      >  > Thanks >  > =====  > Sincerely  > Amit S > J > ________________________________________________________________________G > Want to chat instantly with your online friends?  Get the FREE Yahoo! * > Messenger http://uk.messenger.yahoo.com/   ------------------------------   Date: 8 Aug 2003 13:32:19 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: mount count3 Message-ID: <KONwsd5PG+hV@eisner.encompasserve.org>   w In article <shRYa.20$zb.5@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes:  > amit sawhney wrote:  >  >> Hi 9 >> WHen I issue a SHOW DEVICE DSA?? I notice that the the 9 >> mount count is 2. What does this mean? How can I check ' >> the details? Could this be an error?  > I > exactly what it says... it is mounted twice. If you do a "sh cluster",  I > odds are that you have a 2 node cluster and it is mounted once on each  	 > node...   E    This doesn't mak sense, DSA are LAD protocol software devices not  G    servable via MSCP.  Each node creates its own connection to the LAD  &    server and sees its own DSA device.   ------------------------------   Date: 8 Aug 2003 13:45:43 -0500 4 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt) Subject: Re: mount count3 Message-ID: <2ygOWpOFxEF6@eisner.encompasserve.org>   q In article <KONwsd5PG+hV@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: y > In article <shRYa.20$zb.5@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes:  >> amit sawhney wrote: >>   >>> Hi: >>> WHen I issue a SHOW DEVICE DSA?? I notice that the the: >>> mount count is 2. What does this mean? How can I check( >>> the details? Could this be an error? >>  J >> exactly what it says... it is mounted twice. If you do a "sh cluster", J >> odds are that you have a 2 node cluster and it is mounted once on each 
 >> node... > G >    This doesn't mak sense, DSA are LAD protocol software devices not  I >    servable via MSCP.  Each node creates its own connection to the LAD  ( >    server and sees its own DSA device. >   I On my machines DSA devices are the ones created when I make a shadow set. I In that instance the mount count can be up to a maximum of the number of   nodes in the cluster.    ------------------------------   Date: 8 Aug 2003 15:44:38 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: mount count3 Message-ID: <Ts2wuaDK3QAA@eisner.encompasserve.org>   j In article <2ygOWpOFxEF6@eisner.encompasserve.org>, kuhrt@nospammy.encompasserve.org (Marty Kuhrt) writes: > K > On my machines DSA devices are the ones created when I make a shadow set. K > In that instance the mount count can be up to a maximum of the number of   > nodes in the cluster.      Name space pollution?      ------------------------------  * Date: Fri, 8 Aug 2003 21:23:06 +0000 (UTC)7 From: moroney@world.std.spaamtrap.com (Michael Moroney)  Subject: Re: mount count( Message-ID: <bh14bq$nsg$1@pcls4.std.com>  = koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:   x >In article <shRYa.20$zb.5@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes: >> amit sawhney wrote: >>  J >> exactly what it says... it is mounted twice. If you do a "sh cluster", J >> odds are that you have a 2 node cluster and it is mounted once on each 
 >> node...  F >   This doesn't mak sense, DSA are LAD protocol software devices not H >   servable via MSCP.  Each node creates its own connection to the LAD ' >   server and sees its own DSA device.   F Regarding shadowsets.  While it is true that each node creates its ownF DSA device and the members, not the DSA devices, are served, each nodeD creates the _exact same_ DSA device, with everything coordinated viaE the lock manager to ensure that this is so, so it is exactly the same = as if there was one physical DSA device mounted on each node.   : I'm not familiar with LAD devices so cannot comment there.  C Besides, at the lowest level, each node has its own independent UCB ; for each disk, anyway.  The 'mount count' is really a count E of nodes accessing particular locks, whose resource names are derived G from the device name and the volume label name.  Being mounted (on VMS) F is a particular pattern of locks in particular modes for the most partD (some stuff written on the disk is different for a mounted drive andB a properly dismounted drive, but even that isn't too important. IfD VMS starts mounting an unmounted disk with an 'already mounted' diskE appearance, you'll get the %MOUNT-I-REBUILD message and it will start  cleaning things up.) --   -Mike    ------------------------------  # Date: Fri, 08 Aug 2003 22:31:51 GMT + From: Ryan Moore <rmoore@rmoore.dyndns.org>  Subject: Re: mount count6 Message-ID: <Pine.LNX.4.53.0308081529370.10170@jaipur>  & On Fri, 8 Aug 2003, Bob Koehler wrote:y > In article <shRYa.20$zb.5@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes:  > > amit sawhney wrote:  > >> Hi ; > >> WHen I issue a SHOW DEVICE DSA?? I notice that the the ; > >> mount count is 2. What does this mean? How can I check ) > >> the details? Could this be an error?  > > J > > exactly what it says... it is mounted twice. If you do a "sh cluster",J > > odds are that you have a 2 node cluster and it is mounted once on each > > node...  > F >    This doesn't mak sense, DSA are LAD protocol software devices notH >    servable via MSCP.  Each node creates its own connection to the LAD( >    server and sees its own DSA device.  I Sorry, you got confused there.  DSA are host-based shadowed devices.  LAD G protocol disks show up as DADx: and LAD protocol tapes show up as MADx: : At least they did on the machines I used with Infoservers.  H If you do an MCR LADCP and issue the connect command, I believe you will get a DADx: device.    ------------------------------  # Date: Fri, 08 Aug 2003 18:32:11 GMT # From: "John Smith" <a@nonymous.com>  Subject: Re: Netware is no VMSH Message-ID: <L0SYa.89156$4UE.21661@news01.bloor.is.net.cable.rogers.com>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:KkMXA$PuHud9@eisner.encompasserve.org...      >  [snip]F >....But the point is it isn't uncommon to keep something that is very important underwraps.   D Just curious...I guess that applies to VMS advertising and marketing too?  F In 'government speak', VMS advertising would be classified as a 'blackD project'. You could be locked up in Marion, Ill. if you even thought< about it, much less mentioned it or hinted at its existence.   ------------------------------  # Date: Fri, 08 Aug 2003 18:35:18 GMT # From: "John Smith" <a@nonymous.com>  Subject: Re: Netware is no VMSH Message-ID: <G3SYa.89159$4UE.66367@news01.bloor.is.net.cable.rogers.com>  , <VAXman- @SendSpamHere.ORG> wrote in message* news:00A23F89.A986455D@SendSpamHere.ORG... > F > Respectfully, Brian who is trying to figure out how to buy that brew > pub he's always want to run.    C Perhaps you and David can get together and run a 'brew-tique'.  :-)    ------------------------------  % Date: Fri, 08 Aug 2003 16:43:28 -0400 - From: Jonathan Boswell <jsb@ost.cdrh.fda.gov>  Subject: Re: Netware is no VMS0 Message-ID: <3F340B70.EF43F383@ost.cdrh.fda.gov>   jlsue wrote:I > What new features do you believe are needed above what's already in the 
 > roadmap?  N HPTC stuff such as MPI.  This was ported to T64 including message passing overG shared memory on multiprocessor Alphas.  Nothing ever surfaced for VMS, N resulting in zero HP sales from my VMS-only shop this year.  Previously, I hadM offered researchers impressive horsepower on ES40s running VMS, and we had 21 P VMScluster nodes total.  But I'm the only one who has ever used the cluster as aO compute engine, and I did that by spitting out as many separate batch processes N as I have Alpha processors.  Our younger researchers all expect MPI/PVM/OpenMPP or *something* that will speed up their jobs beyond that of a single processor. M They are used to Beowulf cluster performance in academia that completes their K jobs in minutes.  The same jobs takes all day on a single Alpha processor.  F Needless to say, they are agitating for Linux on commodity processors.    - JB    ------------------------------  % Date: Fri, 08 Aug 2003 08:29:14 -0700 + From: "Barry Treahy, Jr." <Treahy@MMaz.com>  Subject: Re: Netware is no VMS' Message-ID: <3F33C1CA.8030408@MMaz.com>    John Travell wrote:   7 >"Barry Treahy, Jr." <Treahy@MMaz.com> wrote in message " >news:3F327E6D.5080402@MMaz.com... >    >  >>jlsue wrote: >> >>     >>L >>>On Wed, 06 Aug 2003 10:33:10 -0700, "Barry Treahy, Jr." <Treahy@MMaz.com>	 >>>wrote:  >>>  >>> 	 >>>        >>> H >>>>A very interesting Linux project, which is underway, is a clusteringI >>>>project that creates a unified process space between cluster nodes so F >>>>that true load balancing and failover can occur when a member nodeK >>>>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 aL >>>>logical conclusion; Process failover/recovery and true load-balancing... >>>> >>>> >>>> >>>>         >>>>L >>>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...  >>G >>Think about it, when is the last time you had a rogue VMS application  >>take down a VMS system?  >>     >> > L >While not a crash as such, even though a forced crash was needed to recover. >from the situation, I have seen TPU hang VMS.M >Someone with an inappropriately large pagefile quota (about half of the only H >pagefile) and moderate working set tried to use TPU to edit a couple of >files at the same time.   >    > @ That is neither VMS, application, or hardware error, but rather < administrator error in either undersizing the page file, or 3 over-granting virtual address space to the user(s).    Barry    --    > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                            ------------------------------   Date: 9 Aug 03 07:24:45 +0200 ) From: p_sture@elias.decus.ch (Paul Sture)  Subject: Re: Netware is no VMS) Message-ID: <iXS+gTxvHgQb@elias.decus.ch>   U In article <3F327E6D.5080402@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  > jlsue wrote: > K >>On Wed, 06 Aug 2003 10:33:10 -0700, "Barry Treahy, Jr." <Treahy@MMaz.com>  >>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  F >>>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...  >>>  >>>      >>>  >>K >>This is really cool, until you realize that the process that just crashed ? >>system1 got failed over to system2 (then system3, system4...)  >>I > Bah, non-sense!  You must be talking about Windows, because this isn't  0 > an issue with VMS nor most flavors of Unix...  >   4 Sorry, Barry, but you are getting carried away here.   H > Think about it, when is the last time you had a rogue VMS application  > take down a VMS system?     % Both Tuesday and Wednesday this week.   - > With the exception of a faulty driver that  H > couldn't handle unexcepted device failures, I have not had a software H > based VMS failure for over a decade and that was Kernel code!  As for K > user-mode code crashing on VMS, I can't think of much except perhaps EDT  A > on VMS 1.0 but that was all RSX based stuff...  In the case of  J > Supervisor or Executive code, if the code faults, it kills the process,  > not the system...  > C Then you have obviously not pushed your systems the way I have seen  many do.   ------------------------------   Date: 9 Aug 03 07:51:34 +0200 ) From: p_sture@elias.decus.ch (Paul Sture)  Subject: Re: Netware is no VMS) Message-ID: <ScgQlDoq$q$x@elias.decus.ch>   J In article <bgu58o$icb$1@news.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:W > In article <3F327E6D.5080402@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  >>jlsue wrote: >>L >>>On Wed, 06 Aug 2003 10:33:10 -0700, "Barry Treahy, Jr." <Treahy@MMaz.com>	 >>>wrote:  >>>    >>> I >>>>A very interesting Linux project, which is underway, is a clustering  J >>>>project that creates a unified process space between cluster nodes so G >>>>that true load balancing and failover can occur when a member node  L >>>>either slows or dies...  There is no question that VMSclusters has lead I >>>>the way in clustering, but it was never completed to what would be a  L >>>>logical conclusion; Process failover/recovery and true load-balancing... >>>> >>>>     >>>> >>> L >>>This is really cool, until you realize that the process that just crashed@ >>>system1 got failed over to system2 (then system3, system4...) >>> I >>Bah, non-sense!  You must be talking about Windows, because this isn't  0 >>an issue with VMS nor most flavors of Unix...  >>H >>Think about it, when is the last time you had a rogue VMS application F >>take down a VMS system?  With the exception of a faulty driver that H >>couldn't handle unexcepted device failures, I have not had a software H >>based VMS failure for over a decade and that was Kernel code!  As for K >>user-mode code crashing on VMS, I can't think of much except perhaps EDT  A >>on VMS 1.0 but that was all RSX based stuff...  In the case of  J >>Supervisor or Executive code, if the code faults, it kills the process,  >>not the system...  >> >>Barry  >> > D > So it was in Kernel mode when it crashed. Are you saying that only= > processes which NEVER go into Kernel mode can be migrated.   > O > Process runs application in user mode. Application does something which means D > it executes some code in Kernel mode and then crashes the system. > > Process is migrated to second system at its last checkpoint.Q > Process continues. Application executes its code in Kernel mode again - crashes  > second system. >   M Well said David. We also have to consider the case of a DoS here. A few years K ago we had some job which DECScheduler would restart immediately if the job J failed. Oh, we hit version 32767 on the logfile, and until I raised my own? priority above 16, it was hard to get a look in on the problem.   N > Just take a look at the ECOs for DEC TCPIP services. There are usually a fewM > fixes for inappropriate usage of NFS, FTP or other parts of the TCPIP suite 9 > which in certain rare circumstances can crash a system.  >  > eg > ' > ECO B   30-APR-2002     Alpha and VAX  >  >         Problem: > H >         INVEXCEPTN crash in TCPIP$INTERNET_SERVICES due to inadvertentG >         freeing of a BG device for an active TN device kernel client.  >  >         Deliverables:  > : >         TCPIP$INTERNET_SERVICES.EXE             V5.3-18B >  >         Reference: > 7 >         PTR 70-5-1928 / CFS.88690 / Req Id: STLQC0001  >         TCPIP_BUGS Note 25727 >         PTR 70-5-1947 / CFS.89075 / Req Id: HPAQC22TZ  >   D Exactly. Our tour of Wednesday's crash dump took me down the road ofE reading exactly the same text. Then we hit the DSN database and found F tons of examples of similar, but not quite the same, crash situations.  C A maze of twisty passages, all alike, but somehow subtly different.    ------------------------------  * Date: Fri, 8 Aug 2003 20:32:19 +0000 (UTC)3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 0 Subject: Re: OpenVMS Advanced Technical Bootcamp0 Message-ID: <bh11ci$ngq$1@sparta.btinternet.com>   Scholarship please?   < Sue Skonetski <susan_skonetski@hotmail.com> wrote in message6 news:857e9e41.0308080555.73c1c81@posting.google.com.../ > Warren has the agenda posted on the web site.  >  > warm regards,  > sue  >  >  >  -----Original Message-----  > From: Skonetski, Susan* > Sent: Wednesday, August 06, 2003 2:09 PM > To: Skonetski, SusanH > Subject: Official announcement - Registration now open for the OpenVMSF > Advanced Technical Boot camp (previously known as OpenVMS Symposium) > Importance: High >  > Dear OpenVMS Technologist, > G > On behalf of the entire boot camp team, it is my pleasure to announce F > the next OpenVMS Advanced technical boot camp Nov 11-14, 2003 at theC > Sheraton Hotel in Nashua NH. This four day event is geared to theiD > intermediate to advanced OpenVMS technologist and your instructorsD > will be expert members of the HP OpenVMS Team. The agenda has been > enclosed for your use. >eD > Please note that registration is very limited and will be based onE > first come first serve basis.  Please do not hesitate to contact men > with any questions.n >  > Agenda >n >  > B > Please read the following notes as they will answer most of your > questions. >  > Warm Regards,  > Sue Skonetskie > Program Managere >p >M > Notes: >c) > Please make sure you visit the web site B > http://www.hp.com/go/openvms/bootcamp on a regular basis for any > additional information.a< > You are responsible for your own travel and lodging needs.' > Registration is limited to 200 people  > Cost is $1,195C > Breakfast, Breaks, Lunch's and dinner on Tuesday and Thursday arep > included.-G > If you need a letter invitation for a visa please let me know as soonu > as possible.H > You will need to bring a lap top, wireless access is preferred.  We doH > not distribute presentations on paper (its bad for the environment and@ > folks leave the binders in their hotel rooms when they leave.)@ > Dress is casual (blue jeans are fine), please wear comfortable# > clothing these will be long days.o@ > The weather in NH in November is cold, do bring a winter coat.D > The best airport to fly into for domestic travel in Manchester, NH> > The best airport to fly into for international is Boston, MA@ > We will be having an engineering lounge for you to talk to the( > engineers which will be open all week.D > Engineering Roundhouse on Tuesday - the engineers will bring their9 > machines from ZKO to show you what they are working on.t: > Contact for the Engineering roundhouse is Catherine Ward > (Catherine.Ward@hp.com)rC > Partners Roundhouse on Thursday -  a number of different partnersmA > showing you what they do and what applications they have (not a3' > tradeshow). This is very informative. 7 > Contact for the Partners Roundhouse is Signe Maximouse > (Signe.Maxious@hp.com)? > The Advanced IA64 block has a prerequisite, you must take theo > intermediate class first.i, > Hotel information will be on the web site.A > Room rate at the Sheraton is $92 per night please note that theg* > registration rate will stop on Oct 26th)C > There will be a free RDB Forum on Monday and Tuesday please visiteG > http://www.oracle.com/rdb/tech_forums/ to register.  Location will behE > the Sheraton Hotel in Nashua, NH. You must register for both events - > the RDB forum is free the Boot Camp is not.i >a >n > ( > To register please note the following:
 > Course: 650s > Class ID: 518704( > Location: Nashua, NH - Sheraton Nashua. > Dates: November 11, 2003 - November 14, 2003 > To Register:' > US 1-800-732-5741, Option 4, Option 22 > Canada 1-800-392-7024r< > Outside US/Canada - Email HPRegistrar3@registration.hp.comG > When Registering via 800# or Email, please be prepared to provide the $ > following information Student Name > Mailing Address1	 > Phone #t > Fax #a > Email Addressl > Payment Informatione@ > Tuition: $1,195 (All Registrants Must Pay ... this includes HP/ > Employees, End Users/Customers, and Partners)O? > Methods of Payment: Credit Card, Check, PO#, or Cost Center #e > (Employees ONLY).t > C > For Purchase orders only you will need the following information:  > Vendor Hewlett Packard" > Address 20555 SH249, CCM0302-052  > Phone # 800-732-5741, Option 4 > Fax # 281-514-2113 > Contact name:Kim Springer  >m > Hotel Informatione >LK > Go to: <http://www.starwood.com/sheraton/meetings/attend_enter_code.html>nE > In addition to booking their rooms, your attendees can view a slidea2 > show of the hotel and general hotel information.B > Guests can sign up for our award winning frequent guest program,* > Starwood Preferred Guest on the web site > Enter Nashua, New Hampshiret > Your Meeting Code is 6920 " > Select The Sheraton Nashua Hotel9 > Enter your Arrival and Departure Dates, Select Continue < > Confirmation appears at the end and we also offer an email > confirmationF > For reservations questions, attendees may call us at 1-888-627-7183.   ------------------------------  % Date: Fri, 08 Aug 2003 21:27:35 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>sY Subject: Re: preventing DCL delete command from warning that a file does not exist existe-' Message-ID: <3F345C17.9D25F15F@fsi.net>u   Charlie Hammond wrote: > n > In article <e7fbc28d.0308080716.fa0a2e4@posting.google.com>, joseph_kanney@yahoo.com (Joseph Kanney) writes: > >hi! > >aH > >I'd like to make the DCL delete command behave like the unix "rm -f",L > >i.e., I don't want to see warnings if a file is not found. If a file doesE > >not exist, I want it to do nothing, but silently. Any hints on howo > >to do this? Thanks! > >n > >-- Joe Kanney >  > Here is an idea. >  > $ CREATE RM.COMw0 > IF F$SEARCH("''P1'") .NES. "" THEN DELETE 'P1' > [EXIT] <-- PRESS CTRL/Zi > $ RM = "@RM.COM" > % > Now use the sumbol RM as a command.   H Ooohhh, careful there! I'm already seeing places where lines of DCL code7 not preceded by a "$" are being rejected as image data.-     --   David J. Dachtera  dba DJE Systems0 http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/c   ------------------------------   Date: 8 Aug 2003 13:28:46 -0500G; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)r# Subject: Re: TCP/IP feature requeste3 Message-ID: <g0rBqIOpXFd3@eisner.encompasserve.org>x  [ In article <r2i0hb.1kb.ln@news.hus-soft.de>, Albrecht Schlosser <ajs856@tiscali.de> writes:a > Jakob Erber wrote: >> iK >> I just would like to tell you, that we would also be very glad, if theres< >> would be finally a FTP API provided for HP TCPIP for VMS. > J > And so would I. "Callable FTP" (like callable mail) is something I wouldH > really appreciate. I thought about writing my own, but that would be a > bigger project :-(  D    First things first.  Did they do STRU VMS yet?  (Yes, I know they)    have a companion FDL file capability).s  J    I have a callable FTP package, but it's in Java and I couldn't justify J    the JNI hooks to put in code that could do STRU VMS since the work was I    really targeting several other platforms and I was just using VMS for t    its development.p   ------------------------------  % Date: Fri, 08 Aug 2003 21:24:56 -0500u1 From: "David J. Dachtera" <djesys.nospam@fsi.net>h  Subject: Re: TDMS usage question' Message-ID: <3F345B78.9D909C4D@fsi.net>u  * Alan Winston - SSRL Admin Cmptg Mgr wrote: > 	 > Gang --  > P > If anybody still uses TDMS and can perhaps find their manuals (which I can't), > maybe you can tell me: > F > How can I get a signed longword integer to display with a (literally > meaningless) decimal point?  > N > That is, my field is   999999.99  and if the value of what I'm displaying isP > 123456 I want it to display 1234.56, not 123456.00.  I would have thought thatN > a scale factor of 0 would do the trick, but that doesn't seem to be allowed,O > and if I blank out the scale factor in the field definition in the TDMS Formso% > editor, the editor puts it back in.-  H Seems to me (from the old BASIC Plus-2 / VAX BASIC days) that a scale of 2 should do what you want, no?   --   David J. DachteraW dba DJE Systemsc http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/-   ------------------------------  # Date: Sat, 09 Aug 2003 02:25:13 GMTiL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")  Subject: Re: TDMS usage question6 Message-ID: <00A2416B.40B4C56D@SSRL.SLAC.STANFORD.EDU>  [ In article <3F345B78.9D909C4D@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:l+ >Alan Winston - SSRL Admin Cmptg Mgr wrote:g >> g
 >> Gang -- >>  Q >> If anybody still uses TDMS and can perhaps find their manuals (which I can't),s >> maybe you can tell me:  >> dG >> How can I get a signed longword integer to display with a (literallyi >> meaningless) decimal point? >> gO >> That is, my field is   999999.99  and if the value of what I'm displaying is Q >> 123456 I want it to display 1234.56, not 123456.00.  I would have thought thatnO >> a scale factor of 0 would do the trick, but that doesn't seem to be allowed,dP >> and if I blank out the scale factor in the field definition in the TDMS Forms& >> editor, the editor puts it back in. > I >Seems to me (from the old BASIC Plus-2 / VAX BASIC days) that a scale of4 >2 should do what you want, no?e   I solved it.  L The problem was with the interaction of TDMS and the CDD.   The screen fieldO gets a scale of -2, but my Fortran-based program was using an unscaled longwordrG defined in an Rdb relation for the field.   I made up a scaled longwordtN variable for screen communication in a different CDD record, and it works fine now.   -- Alane   --  O ===============================================================================t0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056gM  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025yO ===============================================================================p   ------------------------------  # Date: Sat, 09 Aug 2003 02:31:24 GMTaL From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")  Subject: Re: TDMS usage question6 Message-ID: <00A2416C.1E322610@SSRL.SLAC.STANFORD.EDU>   In article <00A2416B.40B4C56D@SSRL.SLAC.STANFORD.EDU>, winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") writes:t\ >In article <3F345B78.9D909C4D@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:, >>Alan Winston - SSRL Admin Cmptg Mgr wrote: >>>  >>> Gang --r >>> R >>> If anybody still uses TDMS and can perhaps find their manuals (which I can't), >>> maybe you can tell me: >>> H >>> How can I get a signed longword integer to display with a (literally >>> meaningless) decimal point?  >>> P >>> That is, my field is   999999.99  and if the value of what I'm displaying isR >>> 123456 I want it to display 1234.56, not 123456.00.  I would have thought thatP >>> a scale factor of 0 would do the trick, but that doesn't seem to be allowed,Q >>> and if I blank out the scale factor in the field definition in the TDMS Formsm' >>> editor, the editor puts it back in.s >>J >>Seems to me (from the old BASIC Plus-2 / VAX BASIC days) that a scale of  >>2 should do what you want, no? >v
 >I solved it.o >UM >The problem was with the interaction of TDMS and the CDD.   The screen fieldaP >gets a scale of -2, but my Fortran-based program was using an unscaled longwordH >defined in an Rdb relation for the field.   I made up a scaled longwordO >variable for screen communication in a different CDD record, and it works fine8 >now.1  L Further clarification, since I realize now that I was unclear: This was TDMSN working with the CDD the way it was supposed to, not any kind of a bug. I justN had trouble remembering all the bits required to make it work correctly, sinceK it's been a very long time since since I last invented a new TDMS form with  a scaled longword on it.  ( (Those pesky apps just keep on working.)   -- Alans   -- rO ===============================================================================r0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056cM  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025 O ===============================================================================    ------------------------------  # Date: Fri, 08 Aug 2003 17:51:03 GMT-> From: Michael Austin <maustin@no-more-spam.firstdbasource.com> Subject: TPU Section files...i6 Message-ID: <bqRYa.21$En.9@newssvr23.news.prodigy.com>  I After using TPU for <mumble> years, I finally have a problem I am having rF difficulty in solving.... Writing a simple TPU Section file that does   the equivilant of the following:   $edit <file> replacem ~e crlf allS exit  	 Thanks...     F Note... return address is munged, but I am guessing anyone wishing to 2 respond via email will be able to figure it out...   Michael Austin.    ------------------------------   Date: 8 Aug 2003 13:34:57 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)h! Subject: Re: TPU Section files... 3 Message-ID: <Ob3OL8nrtQyN@eisner.encompasserve.org>,  w In article <bqRYa.21$En.9@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes:.K > After using TPU for <mumble> years, I finally have a problem I am having iH > difficulty in solving.... Writing a simple TPU Section file that does " > the equivilant of the following: >  > $edit <file>	 > replaceg > ~r > crlf > allH > exit >  > Thanks...s  G    Just what is that supposed to do?  Break all lines having ~ into twoiD    lines?  Stop thinking in terms of stream files and just break the    lines in two.   ------------------------------  # Date: Fri, 08 Aug 2003 20:10:46 GMT@3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)s! Subject: Re: TPU Section files...n1 Message-ID: <atTYa.1895$3v6.806@news.cpqcorp.net>@  w In article <bqRYa.21$En.9@newssvr23.news.prodigy.com>, Michael Austin <maustin@no-more-spam.firstdbasource.com> writes: J >After using TPU for <mumble> years, I finally have a problem I am having G >difficulty in solving.... Writing a simple TPU Section file that does e! >the equivilant of the following:e > 
 >$edit <file>o >replace >~ >crlfa >all >exitp  0 I assume you can get the file into a tpu buffer.J I don't have the detailed syntax at hand, but it goes something like this.   Position to beginning of buffer  Loop     search for ~     exitif not found     delete the character     split_line endloopp   -- oJ       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Fri, 08 Aug 2003 16:24:14 -0500m( From: brandon@dalsemi.com (John Brandon)= Subject: Re: Using TCPIP$SMTP_SEND_FROM_FILE in a FORTRAN appc1 Message-ID: <03080816241443@dscis6-0.dalsemi.com>h  N Earlier today I posted a request for a snippet using tcpip$smtp_send_from_fileE in a FORTRAN app.  It seems that the tcpip$smtp_send_from_file is a Ca) application.  So... I have a work-around:s  I -------------------------------------------------------------------------e FORTRAN:   c  c send message using SMTP SFFa ct  9         iostat = sendpage_c(%ref(infile_name), %val(len))y  I -------------------------------------------------------------------------  C:  / unsigned int tcpip$smtp_send_from_file(char *);   + sendpage_c (char infile_name[256], int len)l   {c     int status=0;c     len = len +1;e   infile_name[len] = '\0';  2   status = tcpip$smtp_send_from_file(infile_name);   }s  I --------------------------------------------------------------------------     This resolves the problem.       J*o*h*n B*r*a*n*d*o*nu VMS Systems Administratorf* firstname.lastname.spam.me.not@dalsemi.com   ------------------------------   End of INFO-VAX 2003.437 ************************> WHen I issue a SHOW DEVICE DSA?? I notice that the the8 > mount count is 2. What does this mean? How can I check& > the details? Could this be an error?  G exactly what it says... it is mounted twice. If you do a "0 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