1 INFO-VAX	Wed, 12 Apr 2006	Volume 2006 : Issue 202       Contents:- Re: Anybody got the kit for BASIC and COBOL ? - Re: Anybody got the kit for BASIC and COBOL ?  Re: Anyone know David Cathey? + Re: Cluster , locks and lock granting order + Re: Cluster , locks and lock granting order + Re: Cluster , locks and lock granting order + Re: Cluster , locks and lock granting order + Re: Cluster , locks and lock granting order  Re: DECnet IV routing errors Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media. Re: Getting the remote address for SET DISPLAY Re: GLIB and GTK for OpenVMS ? host-based minimerge and VAX  Re: host-based minimerge and VAX, Re: How does a fix become a published patch?' Re: Internal to External PID conversion ' Re: Internal to External PID conversion ' Re: Internal to External PID conversion ' Re: Internal to External PID conversion ' Re: Internal to External PID conversion A Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). % Re: possible to avoid a shadow merge? % Re: possible to avoid a shadow merge? % Re: possible to avoid a shadow merge? % Re: possible to avoid a shadow merge? % Re: possible to avoid a shadow merge? " Re: Shadowed System disk questions" Re: Shadowed System disk questions. Re: SMTP: stop sending "no such user" messages. Re: SMTP: stop sending "no such user" messages. Re: SMTP: stop sending "no such user" messages Re: Using the Serial Terminal   F ----------------------------------------------------------------------  % Date: Tue, 11 Apr 2006 20:14:09 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>6 Subject: Re: Anybody got the kit for BASIC and COBOL ?6 Message-ID: <443C5461.CE6B73AA@NeOaSrPtAhMlNiOnWk.net>  ! "johnhreinhardt@yahoo.com" wrote:  >  > Chuck wrote: > > Hello Peter, Dave, and Ian,  > > G > > I just updated my Google profile with information about my location  > > within Washington State. > >   > > My mailing/living address is > > 1008 Central, S.E. > > Olympia, Washington 98501  > > J > > I tried the http://vmsone.com link, but the connection timed out the 3( > > separate times I tried to access it. > > A > > I very much to appreciate y'all taking the time to assist me, 	 > > Chuck  > G > vmsone.com is registered to the late, great John Wisniewski.  I don't 5 > know who would be taking care of it now, if anyone.   , David Cathey of Montagar.com I should think.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  # Date: Wed, 12 Apr 2006 04:02:42 GMT   From: John Santos <john@egh.com>6 Subject: Re: Anybody got the kit for BASIC and COBOL ?* Message-ID: <CZ__f.5524$wH1.2935@trnddc03>   JF Mezei wrote:  > Hoff Hoffman wrote:  > F >>   I've not seen an HP official statement on acquiring these productG >>kits; on how hobbyists are intended to gain access to the kits not in 9 >>the hobbyist distro.  (Not that I can recall, that is.)  >  >  > E > Just before Sue took over from Mr  Wisniewski, there had bene a big F > brou-haha here about someone announcing availability of downloadableF > kits and now this may stir things up. Perhaps Sue could comment hereI > about any progress made on this issue of making it possible to download ( > kits over the internet for hobbyists ?  ? I just received email today from DSPP saying that kits for most : compilers are now available for download for DSPP members.  + So maybe the same thing soon for hobbyists?    --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Tue, 11 Apr 2006 20:11:20 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>& Subject: Re: Anyone know David Cathey?6 Message-ID: <443C53B8.F29DC815@NeOaSrPtAhMlNiOnWk.net>   Dave Froble wrote: >  > davidc@montagar.com wrote:C > > I'm working on it.  Had a few problems while I was out of town. G > > Something always strange has to happen while I'm out of town...  It  > > should be up now.  > >  > ! > It's a rule.  It has to happen.  > F > I'd be over 2000 miles from home, and I had to get a call over everyD > little crisis, as if I could do a damn thing from 2000 miles away. > - > It's a rule, go out of town, stuff happens.   F Rather like expecting a phone call. Easiest way to make the phone ring is to walk away from it!   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  # Date: Tue, 11 Apr 2006 18:50:39 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>4 Subject: Re: Cluster , locks and lock granting order2 Message-ID: <3US_f.6173$dJ4.6036@news.cpqcorp.net>   JF Mezei wrote: C > OK, say you have a 4 node cluster, node1 through node4, with each  > running the same application.  > A > I want to distribute processing AND provide automatic failover.   C    I might well use DNS and failsafe for this, or a cluster alias.  C (There are also IP tools for load balancing provided within TCP/IP  
 services.)  @    If there is processing going on within the cluster, the batchD queues can potentially be used, and various commands can be used to I create server processes (or pools of server processes) on specific hosts.   J > "chocolate type"  requests would be handled primarily by the app runningH > on NODE1, fail over to NODE2, fail over to NODE3 and finally, when all. > other nodes are down, NODE4 would handle it.  I    A defined ordering of fail-over makes things rather more complex.  If  G you don't really care where the processing ends up, then the effort of  , dealing with failures becomes rather easier.  G    As the application operates, feedback can be sent to the clients to  H cause them to connect over to another host.  This assumes you have some F control over the protocol, or a way to cause the client to switch its   connection over to another host.  J > With $ENQ, to achieve this, it is correct to state that the applicationsF > would have to each do the $ENQ in the the order NODE1, NODE2, NODE3,0 > NODE4 to achieve the desired failover effect ?  H    I'd not bet on the order of locks in the queue.  You either have the I lock, you don't have the lock, or you are waiting for the lock.  I'd not   read into it any further.   H > Is there a way to prioritise $ENQs so that even if NODE2 registers itsJ > $ENQ late, it would still be the first to be given the lock should NODE1 > fail ?  D    Not that I would depend on.  Again, ordering fail-over gets more D complex, particularly given the need to deal with multiple failures.  H > Also, when the application starts, is there a way for it to get a listJ > of processes that either hold the lock or have a pending request to hold5 > that lock ? (remembering that this is on a cluster)       $getlki[w].  H > If all processes $ENQ with a NULL MODE (LCK$L_NLMODE), does it make itE > easier to get the list of processes with an interest in that lock ?   I    You need to hold the lock or be waiting for the lock to be listed, of  B course.  Holding a null-mode lock also makes it more efficient to G convert, and it's also the only way you can perform a synchronous lock  G dequeue operation.  ($deq itself is not synchronous with the dequeuing  G of the lock, so you must first down-convert to a null lock -- which is  / synchronous -- and you can then $deq the lock.)     J > Also, given a process PID, is there a way to get the cluster node name ?G > I looked at the item codes from $GETJPI and I didn't see an item code A > (or was this added since the last grey wall was printed (5.5) ?   H    You will want to use the manuals appropriate for the version in use, E as OpenVMS V5.5 and its documentation is fifteen or so years old now.   ?    But to answer the question asked, you can use JPI$_NODENAME.   E   There are also events when nodes arrive in the cluster, or depart.  E And there are TDF-change events, and other similar system activities  I that an application can request the associated notifications.  When I am  H working on one of these, I tend to use $getsyi[w] to see what nodes are B present (and what hardware), and I code the application to try to D utilize all of the nodes configured.  A small core database can all I designation of specific nodes that are either required or off-limits, of   course.    ------------------------------  % Date: Tue, 11 Apr 2006 15:35:20 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 4 Subject: Re: Cluster , locks and lock granting order, Message-ID: <443C04F2.FBAD40C5@teksavvy.com>   Hoff Hoffman wrote: D >    I might well use DNS and failsafe for this, or a cluster alias.D > (There are also IP tools for load balancing provided within TCP/IP > services.)  F DNS cannot be used in a NAT environment.  IP failover not available on all VMS platforms.  J >    A defined ordering of fail-over makes things rather more complex.  IfH > you don't really care where the processing ends up, then the effort of. > dealing with failures becomes rather easier.  G I care where it ends up. For instance, if a node is not "full featured" F and doesn't have direct access to yje drives containing the web pages,C then when the other nodes fail, this "node of last resort" would be E serving a backup copy of the web pages with warnings that part of the D web site is unavailable etc. So I don't want that node to kick in if. another full function node is still available.  5 >    I'd not bet on the order of locks in the queue.     Thanks.   E >    Not that I would depend on.  Again, ordering fail-over gets more F > complex, particularly given the need to deal with multiple failures.  F Actually, VMS does all that nasty work for me because the lock managerD takes care of node failures. So I just need to work on a good/simpleH robust algorithm to manage the locks so that each application/node knows, exactly who is up, who is handling what etc.   >    $getlki[w].A >    But to answer the question asked, you can use JPI$_NODENAME.   E So this would have been added after V5.5. How come it wasn't added to E GETJPI much earlier than that ? Clustering ahs existed for quite some  time prior to V5.5 ?  F >   There are also events when nodes arrive in the cluster, or depart.  H My goal is to have that application start during booting and always run.F So if a node boots, when that application starts , the other instancesF of that app on the other nodes have to find out about it (blocking ASTB for instance), and relinquish control over services that the newlyF booted node has priority. (eg: of the booting node is the primary nodeD for web server, the backup node would relinquish that IP and let the primary node take it back).   G is there a way for an application to be notified if a node joins/leaves 
 the cluster ?    ------------------------------  % Date: Tue, 11 Apr 2006 20:43:48 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 4 Subject: Re: Cluster , locks and lock granting order/ Message-ID: <O42dneC9ra8l0aHZRVn-tw@libcom.com>    JF Mezei wrote:   I > I care where it ends up. For instance, if a node is not "full featured" H > and doesn't have direct access to yje drives containing the web pages,E > then when the other nodes fail, this "node of last resort" would be G > serving a backup copy of the web pages with warnings that part of the F > web site is unavailable etc. So I don't want that node to kick in if0 > another full function node is still available.  H Well now you've given enough information for others to make suggestions.  I Some of the following would be redundant, and you can choose which parts  ! would fit your requirements best.   A 1) Have a lock that indicated you are free to be the web server.  I whoever has this lock is the web server.  If you don't get the lock, you  G cannot be the web server.  Note that you'd want an immediate return on  / failure, not to wait until the lock is granted.   F 2) Have each system keep track of the existance in the cluster of the E system immediately above itself in the ordered list.  If that system  E drops out, try for the lock in #1.  A periodic look is one method to   make such a determination.  F 3) If the local system is currently the web server, periodically test E for the system above it in the ordered list.  If it exists, stop web  E activity, and release the lock.  Might have to watch for all systems   higher in the ordered list.   G 4) The system highest in the ordered list, upon booting up, should try  A for the lock and wait (maybe some reasonable time) for the lock,  2 figuring that whoever has it will be releasing it.  F Just some initial thoughts.  Not thought out, so definitely should be G able to improve upon them.  The key concepts, a lock can determine who  C is the web server, and testing for the existance of a system could  I indicate that that system *could* be the web server.  There is still the  6 issue of whether it's actually running the web server.  H As Steve and others have said, you either are granted the lock, or not, G whether waiting for it, or returning without the lock.  Don't count on   anything else.   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  + Date: Wed, 12 Apr 2006 01:23:59 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 4 Subject: Re: Cluster , locks and lock granting order( Message-ID: <e1hkrf$8hb$1@pcls4.std.com>  G You could enforce a priority scheme with multiple locks, where the node D holding a certain lock is the owner of whatever it is, a lock with aI value block with a value of the priority of the resource owner (not sure  K offhand whether this should be the same or a different lock from the first) I and a doorbell lock.  A node tries to get the first lock and if it fails, H it checks the value block for the priority of the owner, and if the nodeJ has a higher priority than the lock holder, it rings the doorbell telling @ the lock holder to release the main lock so the node can get it.  B This is oversimplified and needs fleshing out but you can do that.   ------------------------------  % Date: Tue, 11 Apr 2006 22:28:58 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 4 Subject: Re: Cluster , locks and lock granting order+ Message-ID: <443C65CD.97C4A98@teksavvy.com>    Michael Moroney wrote:D > This is oversimplified and needs fleshing out but you can do that.  ' Yeah I was thinking along those lines.    G Every node capable of running the web server takes up the lock with its E priority such as WEB3, as well as attempts to take a lock on the next 6 higher priority until it reaches a lock it cannot get.  G so,. if node2 is gone, node3 would take out locks for web3 and web2 and 0 then wait for the web1 lock to become available.  A When node2 returns, it attempst to take it own lock, triggering a H blocking AST on node3, node3 releases that lock and any lock it had withE higher priority, and node2 can then take its lock and then attempt to  take locks at a higher level.   F The one who is able to take out the WEB1 lock is the one that gets the
 web requests.   M (I still need to work through this to ensure it is working in 100% of cases).    ------------------------------  # Date: Tue, 11 Apr 2006 19:19:08 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> % Subject: Re: DECnet IV routing errors = Message-ID: <MiT_f.51683$wl.26014@text.news.blueyonder.co.uk>   H Phase IV sets all adapters known to the NCP database to use the same MACI address derived from the DECnet address (area*1024 + number, converted to L hex, byte reversed, used to fill in the xx-xx of AA-00-04-00-xx-xx). Phase VJ will allow you to select which adapters to enable / disable Phase IV styleH addressing on (and thus changing the MAC address on those adapters where) you've enable Phase IV style addressing).   C If you have more than one adapter running with the same MAC address G connected to the same network segment (even intermittently), you have a B duplicate MAC address on the LAN segment. That's not a good thing.  I When DECnet starts the adapter it is of course smart enough to check that H no-one else is using the same MAC address already, however, that doesn't4 always work in these days of "intelligent" switches.  K So, that might be one way in which a node is seeing it's own hello packets. H I can't immediately think of a way in which a node with a single adapterG could see it's own hello packets (or would want to) without the network  doing something slightly odd.    --     Hope this helps, Colin. ) colin DOT butcher AT xdelta DOT co DOT uk E It's not mine, but I like this definition: Legacy = stuff that works. / "AEF" <spamsink2001@yahoo.com> wrote in message = news:1144711479.606574.313130@v46g2000cwv.googlegroups.com...    ------------------------------  # Date: Tue, 11 Apr 2006 17:44:05 GMT " From:   VAXman-  @SendSpamHere.ORG# Subject: Re: DSPP and OpenVMS media 0 Message-ID: <00A540F6.76FBB8A4@SendSpamHere.ORG>  Y In article <-ZCdnZjN7sLKlqbZRVn-vw@libcom.com>, Dave Froble <davef@tsoft-inc.com> writes:  >  >  >Rich Jordan wrote: J >>>>I can't help but think this will put a heavier crimp in the ability ofF >>>>hobbyists to get current VMS and LP kits than already exists, too. >>   >>  O >>>To hell with just the hobbyists, it's also hurting the 3rd party development 	 >>>folks.  >>   >>  G >> Umm...we're 3rd party developers, and the hurt we're feeling was the  >> top part of the message.  >>  G >> I'm also a hobby user, so its a double whammy; no more being able to > >> borrow work's updates to keep my home systems up to date... >>   >> Rich  >>   > D >Nothing new.  Since 1978, there have been occasional moves to make H >things better, seperated by many moves to make things worse.  The lack I >of vision has been in place since day 1.  It's a beancounter mentality,  H >"let's see how much blood we can squeeze from a rock", with no thought @ >of survival of the rock.  Turns out the 'rock' is actually the J >foundation the entire VMS world is based upon.  You name them, I'm tired 
 >of doing so.   0 Take one step forward and three steps back... :(   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Tue, 11 Apr 2006 17:46:53 GMT " From:   VAXman-  @SendSpamHere.ORG# Subject: Re: DSPP and OpenVMS media 0 Message-ID: <00A540F6.DAED17D9@SendSpamHere.ORG>  Y In article <-ZCdnZjN7sLKlqbZRVn-vw@libcom.com>, Dave Froble <davef@tsoft-inc.com> writes:  >  >  >Rich Jordan wrote: J >>>>I can't help but think this will put a heavier crimp in the ability ofF >>>>hobbyists to get current VMS and LP kits than already exists, too. >>   >>  O >>>To hell with just the hobbyists, it's also hurting the 3rd party development 	 >>>folks.  >>   >>  G >> Umm...we're 3rd party developers, and the hurt we're feeling was the  >> top part of the message.  >>  G >> I'm also a hobby user, so its a double whammy; no more being able to > >> borrow work's updates to keep my home systems up to date... >>   >> Rich  >>   > D >Nothing new.  Since 1978, there have been occasional moves to make H >things better, seperated by many moves to make things worse.  The lack I >of vision has been in place since day 1.  It's a beancounter mentality,  H >"let's see how much blood we can squeeze from a rock", with no thought @ >of survival of the rock.  Turns out the 'rock' is actually the J >foundation the entire VMS world is based upon.  You name them, I'm tired 
 >of doing so.   L This mentality is one of the reasons I can't see spending my money this yearL to go to the OpenVMS Bootcamp.  There is no reason for anyone to continue to; load cargo on a sinking ship with hopes that it will float.    --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  % Date: Tue, 11 Apr 2006 11:23:30 -0700 # From: "Tom Linden" <tom@kednos.com> # Subject: Re: DSPP and OpenVMS media ) Message-ID: <op.s7uv1ge1zgicya@hyrrokkin>   F On Tue, 11 Apr 2006 10:46:53 -0700, VAXman- <@SendSpamHere.ORG> wrote:  K > This mentality is one of the reasons I can't see spending my money this    > yearD > to go to the OpenVMS Bootcamp.  There is no reason for anyone to  
 > continue to = > load cargo on a sinking ship with hopes that it will float. E Ship is sound, just need to pump out the bilges and set a new course.    ------------------------------  # Date: Tue, 11 Apr 2006 18:24:55 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com># Subject: Re: DSPP and OpenVMS media 2 Message-ID: <XvS_f.6169$QG4.2111@news.cpqcorp.net>   Rich Jordan wrote:) > ... Effective with DSPP renewal you can A > no longer order the OpenVMS Alpha media subscription service...   ,    Based on what has been explained to me...  G    HP has not traditionally provided a consolidated media subscription  D service.  OpenVMS I64 follows this model, and does not have a media  subscription offering.  D    Contract customers can order ("pull") upgrade kits, and can also D download kits.  Access to the new versions is tied into the service 1 contracts, or to the purchase of the new version.   E    Both DIGITAL and Compaq traditionally required a Right To Upgrade  I purchase to use the new versions, as well -- a software license purchase  I typically included rights to then-current and older versions, but not to  
 the upgrades.    ------------------------------    Date: 11 Apr 2006 14:55:56 -0700( From: "Rich Jordan" <jordan@ccs4vms.com># Subject: Re: DSPP and OpenVMS media C Message-ID: <1144792556.356656.180250@v46g2000cwv.googlegroups.com>   G >   HP has not traditionally provided a consolidated media subscription E > service.  OpenVMS I64 follows this model, and does not have a media  > subscription offering.   Hoff, C       it is yet another service downgrade/price increase on the VMS = customers, and in this case specifically on the VMS developer C community.  Just like the license transfer fee increase, it hits us F small timers hardest.  Maybe its just a case of HP continuing to prune9 away at higher level services that it inherited from DEC.   F      The loss of the subscription service for Alpha (where all but oneC of our customers currently live; that one is still VAX) is going to F have a real impact on us, both for our ability to keep our developmentC Alpha up to date, and to help our customers, who nearly all do have ; software support/right to new version support, do the same.   D      It was noted at the recent HP/Intel porting conference that I64B also does not provide an upgrade license;  I assume thats what youE meant above; either you maintain support on a product or else you get @ to repurchase from scratch in order to upgrade, as if you hadn'tE owned/licensed the product already; this is similarly painful for our G small-medium business customers; it potentially takes away a lot of the @ advantage that the lower up front hardware/software costs of theA itanium systems would otherwise provide (though I haven't run the B numbers yet to find out how much).  Sometimes you just have to cutC costs for a while, and contract services is a possible means... but F under the new system that short term necessary savings can turn into a? longer term 'VMS is too expensive to update now' reason to look 
 elsewhere.  F      Please understand that I'm writing this from the perspective of a? long time DEC and Compaq reseller, partner, and VAR, though now C relegated to remote and indirect status by HPs demands for 7-figure E sales to qualify to sell Alpha hardware and VMS, even for those of us E once tiered under a distributor.  Every time HP makes a change to VMS G or Alpha services and situations, it seems to be designed to negatively C impact us one way or another, and make it harder to sell VMS to our E existing and prospective customers.  Either prices go up, services go G down, or both.  Even the potential silver linings, like the aggregation C of licenses with VMS in tiers on the Itanium and the lower up front E costs, and inclusion (finally!) of the unlimited user licenses, comes E with that black cloud of upgrade qualification via continuous support  contract only.  5 Thanks for taking the time to respond.  It matters :)   
 Mark Schafer, F      thank you also for responding.  Any word or chance on getting theE other layered products posted for download via DSPP?  Or even the VMS G installation disks (as image files perhaps)?  It would be nice if since G DSPP provides licenses for it there could be some method of getting the E kits, even piecemeal, short of blowing the budget on a very expensive 
 media kit.   Thanks   Rich   ------------------------------  % Date: Tue, 11 Apr 2006 20:47:50 -0400 ' From: Dave Froble <davef@tsoft-inc.com> # Subject: Re: DSPP and OpenVMS media / Message-ID: <O42dneO9ra810KHZRVn-tw@libcom.com>    Tom Linden wrote: H > On Tue, 11 Apr 2006 10:46:53 -0700, VAXman- <@SendSpamHere.ORG> wrote: > F >> This mentality is one of the reasons I can't see spending my money 
 >> this  year E >> to go to the OpenVMS Bootcamp.  There is no reason for anyone to    >> continue to> >> load cargo on a sinking ship with hopes that it will float. > G > Ship is sound, just need to pump out the bilges and set a new course.   I Ship may be sound, but nobody is manning the pumps, and nobody is on the  E bridge.  Ship is running out of fuel, and nobody is looking for more.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 11 Apr 2006 13:27:27 -0700  From: Z <Z@ids.net> 7 Subject: Re: Getting the remote address for SET DISPLAY ( Message-ID: <MiU_f.1251$X44.80@fe02.lga>   This is fixed.  E It's weird, but SYS$REM_NODE goes away once the LOGIN.COM exits. And   that was confusing me.  I Examining the logical from within the acct's login.com procedure I could  I see that it was formatted as a.b.c.d:: with TCPIP V5.4. But with UCX 4.1  I and 4.2, it was a very large integer: (a*256*256*256 + b*256*256 + c*256   +d)::.  ; The acct's login.com only decoded the old UCX-style format.   # Anyway, thank you for all the help.    ------------------------------  # Date: Tue, 11 Apr 2006 18:57:49 GMT 7 From: John Malmberg <malmberg@dskwld.zko.hp.compaq.dec> ' Subject: Re: GLIB and GTK for OpenVMS ? 2 Message-ID: <N_S_f.6175$WD4.3428@news.cpqcorp.net>   Klaus-D. Bohn wrote:
 > Hello John,  > K > sometimes, between to finish my new product for OpenVMS, i try to compile 9 > the new version. But i get to much "CC-W-PTRMISMATCH1". L > I don't know if i have set the right defines. Do you have a hint or how do > you compile this new version?   H That warning is because older versions of GCC and other C compilers did 1 not diagnose signed/unsigned pointer mis-matches.   I It really is a bug in the original source because under ANSI rules, such  " mis-matches are no longer allowed.  E The warning can be inhibited and the HP OpenVMS compiler will do the   right thing in this case.   C I use the /FIRST_INCLUDE feature to include a header file with the  H OpenVMS specific stuff like below when I need to get code maintained by  someone else to build.  $ #pragma message disable ptrmismatch1  E If it is my code, I either fix the declarations of the variables, or  E apply the appropriate casts with a comment of why the cast is needed.    -John ! malmberg@dskwld.zko.hp.compaq.dec  Personal Opinion Only    ------------------------------  + Date: Tue, 11 Apr 2006 19:19:15 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)% Subject: host-based minimerge and VAX $ Message-ID: <e1gvfj$t1t$1@online.de>  H In article <odmg9SwTK$Qc@cuebid.zko.hp.com>, brooks@cuebid.zko.hp.nospam (Rob Brooks) writes:    4 > (Phillip Helbig---remove CLOTHES to reply) writes: > F > > Apparently host-based MINIMERGE is a no-go since the shadow set inG > > question has two members each of which has a direct connection to a I > > (different) VAX.  (I'm still confused about this; see the new thread  $ > > "host-based minimerge and VAX".) > 4 > That thread hasn't appeared yet to my news reader.   It has now.  :-)  I > That a disk is directly attached to a VAX does *not* disqualify it from J > using HBMM.  If the disk is *mounted* on a VAX, HBMM won't work, but theL > VAX can serve devices to Alphas with no problem.  To the Alpha, the VAX is > just another MSCP controller.    OK, that makes sense.   D I think I understand MINICOPY now and successfully use it for shadowC sets where the members have direct connections ONLY to VAXes.  This B works as long as the MOUNT and DISMOUNT commands both come from an ALPHA.    F It seems that MINIMERGE is more restrictive.  For example, I can have I shadow sets mounted on a VAX and still use MINICOPY while that's not the   case with MINIMERGE.  D However, the documentation makes it sound LESS restrictive.  QuotingD from the 8.2 documentation (which I assume applies to the 7.3-2 HBMM kit):   C    Whereas minicopy requires that all cluster members have minicopy E    support, HBMM requires only that all cluster members support write ;    bitmaps but does not require that they all support HBMM.   G I think this is rather confusing.  First, while the part about HBMM is  F true, the fact that it doesn't work if the shadow set is mounted on a G VAX severely limits its use.  Second, as I said I do use MINICOPY with  E disks with direct connections only to ALPHA (just MOUNT and DISMOUNT  . have to come from ALPHA).  However, HELP says:      DISMOUNT            /POLICY     9            /POLICY=[NO]MINICOPY[=(OPTIONAL)] (Alpha only)      G which makes it rather clear that this is ALPHA only.  So in what sense  A does "minicopy require[s] that all cluster members have minicopy  	 support"?   H To sum up, it appears that I can use HBMM from ALPHA on any shadow sets I in the cluster, wherever their members are served from, as long as these  % shadow sets are not mounted on a VAX.    ------------------------------    Date: 11 Apr 2006 16:06:05 -0500. From: brooks@cuebid.zko.hp.nospam (Rob Brooks)) Subject: Re: host-based minimerge and VAX , Message-ID: <9UA6PrFCsG5Z@cuebid.zko.hp.com>  2 (Phillip Helbig---remove CLOTHES to reply) writes:  F > I think I understand MINICOPY now and successfully use it for shadowE > sets where the members have direct connections ONLY to VAXes.  This D > works as long as the MOUNT and DISMOUNT commands both come from an	 > ALPHA.   > H > It seems that MINIMERGE is more restrictive.  For example, I can have K > shadow sets mounted on a VAX and still use MINICOPY while that's not the   > case with MINIMERGE. > F > However, the documentation makes it sound LESS restrictive.  QuotingF > from the 8.2 documentation (which I assume applies to the 7.3-2 HBMM > kit):  > E >    Whereas minicopy requires that all cluster members have minicopy G >    support, HBMM requires only that all cluster members support write = >    bitmaps but does not require that they all support HBMM.   E I agree that statement is a bit confusing.  It is, however, accurate.   C Minicopy (first supported on V7.3, then backported to V7.2-2 [which G actually came out *after* V7.3]) REQUIRES that ALL cluster members must G run a version of VMS that supports Minicopy.  If you boot (for example) L a V7.2-1 (or VAX V7.2) node into a cluster that has active minicopy bitmaps,J the bitmaps get deleted IMMEDIATELY.  That those "non-player" nodes didn'tJ have any minicopy-active shadow sets mounted is irrelevant.  The fact thatI there was a "non-player" node in the cluster prevents the use of minicopy M across the whole cluster.  I have no idea what happens when a minicopy bitmap I is being used to drive a minicopy operation and a "non-player" node joins * the cluster during that minicopy recovery.  J HBMM is far more forgiving!  While HBMM does not run on OpenVMS Alpha V7.3K or V7.3-1 (or OpenVMS VAX at all), you can have those nodes in your cluster K and still use HBMM if those "non-player" nodes do not have any HBMM-capable L shadow sets mounted.  So yes, HBMM is technically less restrictive regarding! node composition of your cluster.     I > I think this is rather confusing.  First, while the part about HBMM is  H > true, the fact that it doesn't work if the shadow set is mounted on a I > VAX severely limits its use.  Second, as I said I do use MINICOPY with  G > disks with direct connections only to ALPHA (just MOUNT and DISMOUNT  0 > have to come from ALPHA).  However, HELP says: > 
 >    DISMOUNT  >      >      /POLICY >     ; >            /POLICY=[NO]MINICOPY[=(OPTIONAL)] (Alpha only)  >     I > which makes it rather clear that this is ALPHA only.  So in what sense  C > does "minicopy require[s] that all cluster members have minicopy   > support"?   C 	It's actually Alpha and I64 (I'm sure there are still places where & 	we need to update the documentation.)  0 	Try running minicopy on OpenVMS Alpha V7.2-1 -)  J > To sum up, it appears that I can use HBMM from ALPHA on any shadow sets K > in the cluster, wherever their members are served from, as long as these  ' > shadow sets are not mounted on a VAX.    	Yes.   ? 	If what I've written is still confusing, keep firing away and  $ 	I'll try to answer your questions!    --    H Rob Brooks    VMS Engineering -- Exec Group     brooks!cuebid.zko.hp.com   ------------------------------  % Date: Wed, 12 Apr 2006 07:31:04 +0300 7 From: "Guy Peleg" <guy.peleg@remove_this_header@hp.com> 5 Subject: Re: How does a fix become a published patch? , Message-ID: <443c828c$1@usenet01.boi.hp.com>  8 "Steve Matzura" <number6@speakeasy.net> wrote in message2 news:sifk32dghgml5t6h03r8fe8rqd22pl91p8@4ax.com...@ > Four-and-a-half years ago, my employer, through me and anotherD > coworker, reported a problem to HP regarding the creation of imageF > savesets of system disks on volumes starting at and larger than 9gb.? > It took them nearly four years to figure it out, with about a C > year-and-a-half of no action mostly due to us not pushing it, but C > eventually the problem was found.  Seems there was a problem with D > putting VMB.EXE out further than LBN 2^24-1 (= 16,777,215) becauseE > BACKUP wasn't using all 32 bits (not to mention a quadword) to mark = > this location.  Anyway, the problem was determined and some " > replacement images issued to us. > H > Shortly after we reported success with the replacement images, I fullyG > expected to see them incorporated into a patch, and then to have that C > patch included in a mandatory service update. No such luck. Was I F > supposed to file some special paperwork with HP to make this happen?G > I remember in the old days, you could file an SPR and either see your B > problem show up in a patch kit or at least an article in the VAXE > Professional soon thereafter.  I figure, if this problem plagued us H > for as long as it did, others must be experiencing it, too, and should" > be told there *is* a fix for it. >   < I'm the one who fixed the BACKUP problem and as far as I can, remember this was reported against VAX V7.3.  A The fix has been checked into the VAX V7.3 remedial stream (Alpha D & IA64 will not see this problem). The fix will see the day of light: with the next BACKUP ECO for VAX V7.3. There are other VAXA fixes that are in the works right now, once completed an ECO will % be created....probably another month.    Guy    ------------------------------  # Date: Tue, 11 Apr 2006 18:33:10 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>0 Subject: Re: Internal to External PID conversion1 Message-ID: <GDS_f.6172$_E4.278@news.cpqcorp.net>    Alan Frisbie wrote: A > Is there any way in DCL to convert an Internal Process ID (PID) $ > to the corresponding External PID?  ;    Here are the usual suspects for PID-related conversions:   C    EXE$CVT_EPID_TO_IPID, EXE$CVT_IPID_TO_EPID, EXE$CVT_IPID_TO_KTB, H EXE$CVT_IPID_TO_PCB, EXE_STD$CVT_EPID_TO_IPID, EXE_STD$CVT_IPID_TO_EPID, EXE_STD$CVT_IPID_TO_PCB   1    I think you will find EXE$CVT_IPID_TO_EPID or  I EXE_STD$CVT_IPID_TO_EPID to your liking.  These are for the non-standard  5 and call-style linkages, respectively, and obviously.    ------------------------------  + Date: Tue, 11 Apr 2006 20:55:33 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 0 Subject: Re: Internal to External PID conversion' Message-ID: <e1h545$um$1@pcls4.std.com>   . Hoff Hoffman <hoff-remove-this@hp.com> writes:   >Alan Frisbie wrote:B >> Is there any way in DCL to convert an Internal Process ID (PID)% >> to the corresponding External PID?   < >   Here are the usual suspects for PID-related conversions:  D >   EXE$CVT_EPID_TO_IPID, EXE$CVT_IPID_TO_EPID, EXE$CVT_IPID_TO_KTB,I >EXE$CVT_IPID_TO_PCB, EXE_STD$CVT_EPID_TO_IPID, EXE_STD$CVT_IPID_TO_EPID,  >EXE_STD$CVT_IPID_TO_PCB  2 >   I think you will find EXE$CVT_IPID_TO_EPID or J >EXE_STD$CVT_IPID_TO_EPID to your liking.  These are for the non-standard 6 >and call-style linkages, respectively, and obviously.  > FWIW, EXE_STD$CVT_EPID_TO_IPID is callable from user mode, so E EXE_STD$CVT_IPID_TO_EPID might be as well.  You'll still need to link B it /SYSEXE which will mean relinking to run on different versions.   ------------------------------  + Date: Wed, 12 Apr 2006 01:33:05 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 0 Subject: Re: Internal to External PID conversion( Message-ID: <e1hlch$4q8$1@pcls4.std.com>  6 Alan Frisbie <Usenet02_REMOVE@Flying-Disk.com> writes:  A >> FWIW, EXE_STD$CVT_EPID_TO_IPID is callable from user mode, so  H >> EXE_STD$CVT_IPID_TO_EPID might be as well.  You'll still need to linkE >> it /SYSEXE which will mean relinking to run on different versions.   H >Honest, I really do RTFM!   I was asking about a way to do it from DCL.F >I was hoping to save a bit of coding in case someone knew a DCL trick >that I didn't.   A Oh, missed the DCL part, I just followed up on the EXE$ routines.   ? What can you do with an IPID from DCL?  This might be somewhat   interesting.   ------------------------------  % Date: Tue, 11 Apr 2006 20:31:42 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>0 Subject: Re: Internal to External PID conversion6 Message-ID: <443C587E.ECB996D9@NeOaSrPtAhMlNiOnWk.net>   Michael Moroney wrote: > 0 > Hoff Hoffman <hoff-remove-this@hp.com> writes: >  > >Alan Frisbie wrote:D > >> Is there any way in DCL to convert an Internal Process ID (PID)' > >> to the corresponding External PID?  > > > >   Here are the usual suspects for PID-related conversions: > F > >   EXE$CVT_EPID_TO_IPID, EXE$CVT_IPID_TO_EPID, EXE$CVT_IPID_TO_KTB,K > >EXE$CVT_IPID_TO_PCB, EXE_STD$CVT_EPID_TO_IPID, EXE_STD$CVT_IPID_TO_EPID,  > >EXE_STD$CVT_IPID_TO_PCB > 3 > >   I think you will find EXE$CVT_IPID_TO_EPID or K > >EXE_STD$CVT_IPID_TO_EPID to your liking.  These are for the non-standard 8 > >and call-style linkages, respectively, and obviously. > ? > FWIW, EXE_STD$CVT_EPID_TO_IPID is callable from user mode, so G > EXE_STD$CVT_IPID_TO_EPID might be as well.  You'll still need to link D > it /SYSEXE which will mean relinking to run on different versions.  D Of course, getting the source code to DCL and actually modifying it,. then building it will be the *REAL* challenge!   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  + Date: Wed, 12 Apr 2006 02:01:23 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 0 Subject: Re: Internal to External PID conversion( Message-ID: <e1hn1j$q37$1@pcls4.std.com>  B "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net> writes:   >Michael Moroney wrote:  >>  @ >> FWIW, EXE_STD$CVT_EPID_TO_IPID is callable from user mode, soH >> EXE_STD$CVT_IPID_TO_EPID might be as well.  You'll still need to linkE >> it /SYSEXE which will mean relinking to run on different versions.   E >Of course, getting the source code to DCL and actually modifying it, / >then building it will be the *REAL* challenge!   F Well, at least if EXE_STD$CVT_IPID_TO_EPID is callable from user mode,0 it should work just fine in supervisor mode :-).   ------------------------------    Date: 11 Apr 2006 12:10:58 -0700" From: dave.baxter@bannerhealth.comJ Subject: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1144782657.916911.119910@u72g2000cwu.googlegroups.com>         Thats my question!....    Dave   ------------------------------  # Date: Tue, 11 Apr 2006 20:02:11 GMT & From: John Reagan <john.reagan@hp.com>N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).2 Message-ID: <7XT_f.6182$DH4.5853@news.cpqcorp.net>  # dave.baxter@bannerhealth.com wrote:  >      Thats my question!....  >  > Dave >   ? So do you have a 32P system that is max-ed out?  The number of  ? applications that run well on even a 32P system is pretty rare.   G The SPD has 32P limit.  If you have real need, you should contact your   account person.    --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------    Date: 11 Apr 2006 14:04:42 -0700" From: dave.baxter@bannerhealth.comN Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1144789482.423095.267950@i39g2000cwa.googlegroups.com>   F I am running 2 x M32's (currently both 24-way) which will be maxed out@ by spring 2007.   I am looking at my upgrade paths for 2007, and" upgrade to M64 is one possibility.   Dave.    ------------------------------  % Date: Tue, 11 Apr 2006 20:41:43 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).6 Message-ID: <443C5AD7.D0214F55@NeOaSrPtAhMlNiOnWk.net>  # dave.baxter@bannerhealth.com wrote:  > H > I am running 2 x M32's (currently both 24-way) which will be maxed outB > by spring 2007.   I am looking at my upgrade paths for 2007, and$ > upgrade to M64 is one possibility. >  > Dave.    Dave,   E You're not alone. Most large Cerner sites are looking at "hitting the G wall", pretty much at mach speeds prior to the arrival of 64-CPU Alphas F (if they are even on the map), and long before Cerner will be preparedE to certify either V8.2 or I64. Exactly when we could expect to see 64 E CPU I64 SuperDomes, and whether they'd measure up to 32 CPU Alphas is B anyone's guess (especially Intel's!), Cerner considerations aside.  G The AlphaServer website does show 64 CPU Alphas. My preliminary feelers G have indicated negative awareness of this among the HP field personnel.   F HP management have been apprised since roughly -5 days. No response to date.    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  % Date: Tue, 11 Apr 2006 22:39:52 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64)., Message-ID: <443C685B.15ADF7BD@teksavvy.com>   "David J. Dachtera" wrote:G > You're not alone. Most large Cerner sites are looking at "hitting the I > wall", pretty much at mach speeds prior to the arrival of 64-CPU Alphas ! > (if they are even on the map),      D Consider that Alpha end of sales in coming in a few months. ConsiderE that HP will do everythinb it can to push people off Alpha , hoping a C percantage choose IA64. So by default, I don't think you'll see any H budgets to certify VMS for 64 processor Alpha systems. (especially if itG means that such customers will have to wait even longer before they can  move to those IA64 things.  H My guess is that such customers should deal with VMS management and  useF "show me the money" arguments to convince them that they stand to loseD too much of they don't provide the support for 64 CPU systems. We'veF seen the argument "customers have to make the business case to justifyM engineers working on X" quite often here and I suspect the same applies here.    ------------------------------  + Date: Tue, 11 Apr 2006 17:40:01 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply). Subject: Re: possible to avoid a shadow merge?$ Message-ID: <e1gplh$j23$1@online.de>  : In article <e1eg4s$1k7$1@news.doit.wisc.edu>, Carl Karcher. <karcher.nomorespam@waisman.wisc.edu> writes:   1 > Phillip Helbig---remove CLOTHES to reply wrote:  > M > > I'm not getting a merge of the system disk, but rather of the disk where   > > SYSUAF etc are.  > K > You'll need to do some things in syshutdwn.com to close the files so the  I > volume dismounts cleanly. Once that happens you won't get a merge. I'd  J > have syshutdwn.com spawn a command prompt so you can see what files are I > open (by what process) at that time. Rightlist is usually one of them.   > You will need to at least: > ! > SET SERVER SECURITY_SERVER/EXIT  > C > As of VMS 7.2-1, shutdown.com will stop other processes that had  B > rightslist open (audit server, decw$*, smiserver) for you. Your  > situation may vary.   D It might be just the security server.  I'll try putting this in.  ItH seems to me that this situation (a member has mounted shadow sets served? entirely by other members and one has cluster-common files on a ; non-system disk) is so common that a factory-supplied basic = SYSHUTDWN.COM should stop those applications with open files.   F A workaround would be to dismount one member of this shadow set, shut E down the node, and bring back the other member with a minicopy.  But   this is not easy to automate.   B Apparently host-based MINIMERGE is a no-go since the shadow set inC question has two members each of which has a direct connection to a E (different) VAX.  (I'm still confused about this; see the new thread    "host-based minimerge and VAX".)   ------------------------------    Date: 11 Apr 2006 14:34:14 -0500. From: brooks@cuebid.zko.hp.nospam (Rob Brooks). Subject: Re: possible to avoid a shadow merge?, Message-ID: <odmg9SwTK$Qc@cuebid.zko.hp.com>  2 (Phillip Helbig---remove CLOTHES to reply) writes:  D > Apparently host-based MINIMERGE is a no-go since the shadow set inE > question has two members each of which has a direct connection to a G > (different) VAX.  (I'm still confused about this; see the new thread  " > "host-based minimerge and VAX".)  2 That thread hasn't appeared yet to my news reader.  G That a disk is directly attached to a VAX does *not* disqualify it from H using HBMM.  If the disk is *mounted* on a VAX, HBMM won't work, but theJ VAX can serve devices to Alphas with no problem.  To the Alpha, the VAX is just another MSCP controller.    --    H Rob Brooks    VMS Engineering -- Exec Group     brooks!cuebid.zko.hp.com   ------------------------------  + Date: Tue, 11 Apr 2006 19:06:28 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply). Subject: Re: possible to avoid a shadow merge?$ Message-ID: <e1gunj$rlb$1@online.de>  D In article <e1gplh$j23$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    < > In article <e1eg4s$1k7$1@news.doit.wisc.edu>, Carl Karcher0 > <karcher.nomorespam@waisman.wisc.edu> writes:  > 3 > > Phillip Helbig---remove CLOTHES to reply wrote:  > > O > > > I'm not getting a merge of the system disk, but rather of the disk where   > > > SYSUAF etc are.  > > M > > You'll need to do some things in syshutdwn.com to close the files so the  K > > volume dismounts cleanly. Once that happens you won't get a merge. I'd  L > > have syshutdwn.com spawn a command prompt so you can see what files are K > > open (by what process) at that time. Rightlist is usually one of them.   > > You will need to at least: > > # > > SET SERVER SECURITY_SERVER/EXIT  > > E > > As of VMS 7.2-1, shutdown.com will stop other processes that had  D > > rightslist open (audit server, decw$*, smiserver) for you. Your  > > situation may vary.    Here's the problem:   ; Files accessed on device DSA510: on 11-APR-2006 20:45:28.09   # Process name      PID     File name .                 00000000  [000000]INDEXF.SYS;1@ OPCOM           21C00211  [SYSTEM.MANAGER]OPERATOR_NADILA.LOG;10@ AUDIT_SERVER    21C00212  [SYSTEM.MANAGER]VMS$AUDIT_SERVER.DAT;17 JOB_CONTROL     21C00213  [SYSTEM.EXE]QMAN$MASTER.DAT;1 7 AUDIT_SERVER    21C00212  [SYSTEM.EXE]VMS$OBJECTS.DAT;1 6 SECURITY_SERVER 21C00218  [SYSTEM.EXE]RIGHTSLIST.DAT;1  A I think the operator log is closed automatically by the standard  F shutdown procedure.  SET SERVER SECURITY_SERVER/EXIT will take care ofE [SYSTEM.EXE]RIGHTSLIST.DAT.  SET AUDIT/SERVER=EXIT will take care of  G VMS$AUDIT_SERVER.DAT and VMS$OBJECTS.DAT.  Since the standard shutdown  H procedures stops the queues, that means QMAN$MASTER.DAT should be OK as I well.  I'll give it a try.  (SMISERVER can have RIGHTSLIST.DAT open, but  G that should be covered by what you said above; if not, I can just kill  5 it (which won't work for SECURITY and AUDIT servers.)    ------------------------------  + Date: Tue, 11 Apr 2006 19:41:31 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply). Subject: Re: possible to avoid a shadow merge?$ Message-ID: <e1h0pa$ucb$1@online.de>  D In article <e1gunj$rlb$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    C > I think the operator log is closed automatically by the standard  H > shutdown procedure.  SET SERVER SECURITY_SERVER/EXIT will take care ofG > [SYSTEM.EXE]RIGHTSLIST.DAT.  SET AUDIT/SERVER=EXIT will take care of  I > VMS$AUDIT_SERVER.DAT and VMS$OBJECTS.DAT.  Since the standard shutdown  J > procedures stops the queues, that means QMAN$MASTER.DAT should be OK as K > well.  I'll give it a try.  (SMISERVER can have RIGHTSLIST.DAT open, but  I > that should be covered by what you said above; if not, I can just kill  7 > it (which won't work for SECURITY and AUDIT servers.)   C OK, I've now shut the node down with no merge!  Here's the relevant G code, which I added after the normal site-specific shutdown stuff (for   layered products etc):  " $  SET SERVER SECURITY_SERVER/EXIT $  SET AUDIT/SERVER=EXIT $! kill the SMI server/ $  @ CLUSTER_MANAGER:STOP_PROCESS.COM SMISERVER   # CLUSTER_MANAGER:STOP_PROCESS.COM is   8 $  PIPE SHOW SYSTEM/PROCESS='P1' | SEA SYS$PIPE 'P1' | -3     (READ SYS$PIPE REC ; DEFINE/NOLOG/JOB PID &REC) ' $  PID = F$EXTRACT(0,8,F$TRNLNM("PID"))  $  DEASSIGN/JOB PID  $  IF PID .NES. "%SEARCH-" $  THEN  $    STOP/ID='PID' $  ENDIF  G The above code is sufficient, but is it all necessary?  In particular,  H are ALL applications which access RIGHTSLIST.DAT closed as of VMS 7.2-2?B (I don't want to test myself since the merge takes a few hours to 
 complete).  I Is there anything else which could have files open but just happened not   to in my case?   ------------------------------  % Date: Tue, 11 Apr 2006 20:27:29 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>. Subject: Re: possible to avoid a shadow merge?6 Message-ID: <443C5781.C3E1F1BE@NeOaSrPtAhMlNiOnWk.net>  / Phillip Helbig---remove CLOTHES to reply wrote:  > F > In article <e1gunj$rlb$1@online.de>, helbig@astro.multiCLOTHESvax.de4 > (Phillip Helbig---remove CLOTHES to reply) writes: > D > > I think the operator log is closed automatically by the standardJ > > shutdown procedure.  SET SERVER SECURITY_SERVER/EXIT will take care ofH > > [SYSTEM.EXE]RIGHTSLIST.DAT.  SET AUDIT/SERVER=EXIT will take care ofJ > > VMS$AUDIT_SERVER.DAT and VMS$OBJECTS.DAT.  Since the standard shutdownK > > procedures stops the queues, that means QMAN$MASTER.DAT should be OK as L > > well.  I'll give it a try.  (SMISERVER can have RIGHTSLIST.DAT open, butJ > > that should be covered by what you said above; if not, I can just kill9 > > it (which won't work for SECURITY and AUDIT servers.)  > E > OK, I've now shut the node down with no merge!  Here's the relevant H > code, which I added after the normal site-specific shutdown stuff (for > layered products etc): > $ > $  SET SERVER SECURITY_SERVER/EXIT > $  SET AUDIT/SERVER=EXIT > $! kill the SMI server1 > $  @ CLUSTER_MANAGER:STOP_PROCESS.COM SMISERVER  > % > CLUSTER_MANAGER:STOP_PROCESS.COM is  > : > $  PIPE SHOW SYSTEM/PROCESS='P1' | SEA SYS$PIPE 'P1' | -5 >     (READ SYS$PIPE REC ; DEFINE/NOLOG/JOB PID &REC) ) > $  PID = F$EXTRACT(0,8,F$TRNLNM("PID"))  > $  DEASSIGN/JOB PID  > $  IF PID .NES. "%SEARCH-"	 > $  THEN  > $    STOP/ID='PID'
 > $  ENDIF  G Great stuff! I'll be likely to use that in a common files consolidation C project at work related to multiple system disks (rolling upgrades,  etc.).  H However, I have my own STOP_PROCESS.COM. It uses F$CONTEXT() and F$PID()= to find processes by name rather than parsing command output.    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  + Date: Tue, 11 Apr 2006 17:50:21 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)+ Subject: Re: Shadowed System disk questions $ Message-ID: <e1gq8t$j23$2@online.de>  F In article <72596$443bbc28$50db5015$22910@news.hispeed.ch>, Paul Sture' <paul.sture.nospam@hispeed.ch> writes:    . > > I'll back up Alan's advice (pun intended). > > K > > A few things to think about once you have system disk shadowing enabed:    I generally agree, but...   F > o - Before applying a boatload of patches (or other work which couldJ >      render your system disk unusable), add a third member, wait for theC >      shadow copy to complete, and dismount it before applying the E >      patches. If you don't have a third disk handy, simply drop the I >      second one to act as a backup. As long as you have followed Alan's G >      advice about startup procedures, if the second member is already J >      dismounted at shutdown, it won't come in at the next reboot. If youH >      have multiple disks defined in bootdev_def, make sure you specifyH >      the desired disk in the boot command to make sure you boot to theI >      correct one. THIS IS IMPORTANT - it is all too easy to boot to the B >      wrong disk and potentially lose your work with a subsequentK >      shadow copy. I'm perhaps stating the obvious, but once caught, twice 3 >      shy (and that was 10 years ago for me!). :-)   B Note that you cannot dismount one member of a shadow set on a liveA system and expect to get a clean copy.  You need to shut down the F system, physically disconnect a member, and bring it back up with one G less member.  What I usually do before this shutdown is set VAXCLUSTER  E to 0.  In my startup procedure in SYLOGICALS.COM, if VAXCLUSTER is 0  I then I don't mount disks served by other nodes, don't start applications  @ I only use in a cluster etc.  Normally, I apply patches, change G VAXCLUSTER back to 2 and reboot.  When all looks OK, I add the removed  I member back in.  (Using a third member would be better, especially since  I one can then use it to make a clone for another system.  If one has this  G process down pat, it is less work than upgrading several system disks.)   I > o - If you have a set of patches which advise a reboot between each one I >      (warning of system instability if you don't), boot into standalone F >      mode to apply them. Since working at the console can be a pain,H >      I prefer to create a command procedure beforehand to do this, and2 >      take a printout of it with me just in case.  @ While this is the recommendation, has ANYONE EVER had a problem @ installing a batch of patches then rebooting once after all are B installed?  (Obvious exceptions are patches to PCSI.)  After all, + consider what happens with an UPDATE patch.   D > o - VMS upgrades require that you switch system disk shadowing offG >      before applying the upgrade. Don't forget this step (I did once, & >      which is why I mention it). :-)  G True, but if you boot from CD shadowing won't be switched on.  You can  I then MOUNT/OVER=SHADOW the target disk for the upgrade AND MAKE SURE YOU   BOOT FROM THAT.    ------------------------------  % Date: Wed, 12 Apr 2006 02:23:02 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> + Subject: Re: Shadowed System disk questions ; Message-ID: <1d45e$443c4868$50db5015$23147@news.hispeed.ch>   / Phillip Helbig---remove CLOTHES to reply wrote: H > In article <72596$443bbc28$50db5015$22910@news.hispeed.ch>, Paul Sture) > <paul.sture.nospam@hispeed.ch> writes:   >  > - >>>I'll back up Alan's advice (pun intended).  >>> J >>>A few things to think about once you have system disk shadowing enabed: >  >  > I generally agree, but...  >  > F >>o - Before applying a boatload of patches (or other work which couldJ >>     render your system disk unusable), add a third member, wait for theC >>     shadow copy to complete, and dismount it before applying the E >>     patches. If you don't have a third disk handy, simply drop the I >>     second one to act as a backup. As long as you have followed Alan's G >>     advice about startup procedures, if the second member is already J >>     dismounted at shutdown, it won't come in at the next reboot. If youH >>     have multiple disks defined in bootdev_def, make sure you specifyH >>     the desired disk in the boot command to make sure you boot to theI >>     correct one. THIS IS IMPORTANT - it is all too easy to boot to the B >>     wrong disk and potentially lose your work with a subsequentK >>     shadow copy. I'm perhaps stating the obvious, but once caught, twice 3 >>     shy (and that was 10 years ago for me!). :-)  >  > D > Note that you cannot dismount one member of a shadow set on a live( > system and expect to get a clean copy.  I Then why have you gone to the trouble to move SYSUAF etc off your system  A disk? Those files at the very minimum will be safe. What else of  I importance is left? Obviously you need to know your own system to answer     this one.   I You can also close applications by running manually closing batch queues  F and running SYSHUTDWN.COM to ensure they aren't using the system disk.   >  You need to shut down theH > system, physically disconnect a member, and bring it back up with one  > less member.    ? No way, except under exceptional circumstances. Rebooting with  G SHADOW_SYS_DISK=0 suffices. Why much with hardware when you don't need  3 to? Anyway, the disk farm could be many miles away.   ;    What I usually do before this shutdown is set VAXCLUSTER G > to 0.  In my startup procedure in SYLOGICALS.COM, if VAXCLUSTER is 0  K > then I don't mount disks served by other nodes, don't start applications  B > I only use in a cluster etc.  Normally, I apply patches, change I > VAXCLUSTER back to 2 and reboot.  When all looks OK, I add the removed   > member back in.   5 That's one way, but why not simply do a minimum boot?   ; >  (Using a third member would be better, especially since  K > one can then use it to make a clone for another system.  If one has this  I > process down pat, it is less work than upgrading several system disks.)  >    True.   I >>o - If you have a set of patches which advise a reboot between each one I >>     (warning of system instability if you don't), boot into standalone F >>     mode to apply them. Since working at the console can be a pain,H >>     I prefer to create a command procedure beforehand to do this, and2 >>     take a printout of it with me just in case. >  > B > While this is the recommendation, has ANYONE EVER had a problem B > installing a batch of patches then rebooting once after all are D > installed?  (Obvious exceptions are patches to PCSI.)  After all, - > consider what happens with an UPDATE patch.  >   I With the exception of PCSI, no, but logging out and in again solves that   one. No reboot required.   > D >>o - VMS upgrades require that you switch system disk shadowing offG >>     before applying the upgrade. Don't forget this step (I did once, & >>     which is why I mention it). :-) >  > I > True, but if you boot from CD shadowing won't be switched on.  You can  K > then MOUNT/OVER=SHADOW the target disk for the upgrade AND MAKE SURE YOU   > BOOT FROM THAT.  >   3 Please read the Installation Notes on that subject.    ------------------------------  + Date: Tue, 11 Apr 2006 17:51:23 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)7 Subject: Re: SMTP: stop sending "no such user" messages $ Message-ID: <e1gqar$j23$3@online.de>  C In article <06041023444317_202002AF@antinode.org>, sms@antinode.org  (Steven M. Schweda) writes:   / > From: JF Mezei <jfmezei.spamnot@teksavvy.com>  > F > > Have you tried extracting the TCPIP$SMTP_RECEIVER.EXE from the 5.5I > > distribution and trying to run it on your non 8.2 system to see if it C > > works ?  That is the image which would be doing teh checking of B > > usernames during the SMTP negotiation when receiving messages. > G >    As I recall, I did, and it behaved badly (hung, died, failed, ...) B > and I stopped getting e-mail, so I yanked it out pretty quickly.  I Release notes say that 5.5 is only for 8.2 and up.  Not a risk I want to   take.    ------------------------------  + Date: Wed, 12 Apr 2006 01:55:16 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 7 Subject: Re: SMTP: stop sending "no such user" messages ( Message-ID: <e1hmm4$4q8$2@pcls4.std.com>    david20@alpha2.mdx.ac.uk writes:    L >Note that checking the HELO argument is about the only check that the RFCs 8 >explicitly say cannot be used to reject a mail message.   >From RFC 1123   >"= > The HELO receiver MAY verify that the HELO parameter really C > corresponds to the IP address of the sender. However the receiver N > MUST NOT refuse to accept a message, even if the sender's HELO command fails > verification.  >. >. >.K > Note also that the HELO argument is still required to have valid <domain> P > syntax, since it will appear in a Received: line; otherwise, a 501 error is to
 > be sent.   >"  G Kind of contradictory.  The last part checks for syntax and rejects the / HELO if it fails, contradicting the first part.   C Anyway, lots of spammers use HELO 127.0.0.1 or HELO <my.ip.address> C where <my.ip.address> is the IP address of the _recipient_.  If you ? compare these to the actual IP address it's really coming from, B (if that's actually necessary, configuration dependent), rejecting6 these will reject lots of spam and nothing legitimate.   ------------------------------  # Date: Wed, 12 Apr 2006 03:39:09 GMT   From: John Santos <john@egh.com>7 Subject: Re: SMTP: stop sending "no such user" messages * Message-ID: <xD__f.9918$Fp4.2905@trnddc01>   Michael Moroney wrote:" > david20@alpha2.mdx.ac.uk writes: >  >  > M >>Note that checking the HELO argument is about the only check that the RFCs  9 >>explicitly say cannot be used to reject a mail message.  >  >  >>From RFC 1123  >  >  >>" = >>The HELO receiver MAY verify that the HELO parameter really C >>corresponds to the IP address of the sender. However the receiver N >>MUST NOT refuse to accept a message, even if the sender's HELO command fails >>verification.  >>.   ? Spurious comma?  "..., even if ..." implies the first clause is A always true, with the second clause just being an extreme example D of when it holds.  "... even if ..." (without the comma) is a littleC ambiguous, but can imply "just because", i.e. the 2nd clause is not 0 sufficient reason for falsifying the 1st clause.     >>.  >>. K >>Note also that the HELO argument is still required to have valid <domain> P >>syntax, since it will appear in a Received: line; otherwise, a 501 error is to
 >>be sent. >  >  >>"  >  > I > Kind of contradictory.  The last part checks for syntax and rejects the 1 > HELO if it fails, contradicting the first part.   B What I think they mean is it is okay to reject for a syntacticallyF invalid HELO message (protocol violation), but not for an non-existent; or spurious, but syntactically correct address on the HELO. D "HELO foo.bar.invalid" is valid, but "HELO foo.b%$<BEL><FF>*.com" is% not and can be legitimately rejected.   E > Anyway, lots of spammers use HELO 127.0.0.1 or HELO <my.ip.address> E > where <my.ip.address> is the IP address of the _recipient_.  If you A > compare these to the actual IP address it's really coming from, D > (if that's actually necessary, configuration dependent), rejecting8 > these will reject lots of spam and nothing legitimate.  C Good point.  Though you have to make sure you aren't rejecting mail > from the local host or from the local domain!  Otherwise Pine,E Mozilla, etc. won't work if you are running them on your SMTP server.    --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Tue, 11 Apr 2006 14:37:54 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> & Subject: Re: Using the Serial Terminal, Message-ID: <443BF783.622EC844@teksavvy.com>   "doypapio@gmail.com" wrote: H > My VT displays certain outputs or fields...for example, name, address,G > height, weight...etc. on the screen. My concern now is how to be able E > to use any one of these values say for another command procedure or F > maybe output it to a file triggered for example by pressing the "DO" > key.  ; You would need to modify your application program for that.   E With a VT terminal, about the only way you can capture a screen is to D use the local "print screen" button and have the printer serial portE connected to the VMS machine with a program that captures whatever is C sent out of the printer port to a file.  aka: 2 serial lines to the G terminal, one that runs the application, and one that captures whatever  you send to "print screen".   E The other way would be to replace the VT terminal with an X terminal, G and at that point, you can do select/copy/paste from one window running # yoru application to another window.   F Note that some of the last generation of VT terminals are able to haveA multiple sessions and have some rudimentary cut/paste capability.    ------------------------------   End of INFO-VAX 2006.202 ************************