1 INFO-VAX	Tue, 13 Jul 2004	Volume 2004 : Issue 384       Contents:1 %SYSTEM-E-BADLOGIC on pmap_unset call VAX/VMS 6.2 " Re: cluster_config extreme badness" Re: cluster_config extreme badness" Re: cluster_config extreme badness$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed$ Re: DECnet Phase IV Time Sync Needed- Re: Getting the RAB from BASIC channel number - Re: Getting the RAB from BASIC channel number 3 Re: interfacing dec basic to web based applications # Re: MySQL 4.1.3-beta on ODS2 volume 5 Re: New vms install, problem with serial console opa0 , No Ethernet device on system (UCX 4.2 ECO 5)- odd things with SYS$GETUAI and uai$_user_data 1 Re: odd things with SYS$GETUAI and uai$_user_data 1 Re: odd things with SYS$GETUAI and uai$_user_data 1 Re: odd things with SYS$GETUAI and uai$_user_data + Re: OpenVMS 7.3-1 on Alpha Wokstation 500au + RE: OpenVMS 7.3-1 on Alpha Wokstation 500au + Re: OpenVMS 7.3-1 on Alpha Wokstation 500au / Re: OpenVMS license transfer  policies and fees 6 Re: OpenVMS Management Station -  cluster storage ????6 Re: OpenVMS Management Station -  cluster storage ????6 Re: OpenVMS Management Station -  cluster storage ???? Re: OpenVMS Positions - Europe< Re: OpenVMS probably the best designed o/s in  the market...  F ----------------------------------------------------------------------    Date: 12 Jul 2004 22:22:13 -07001 From: patrice.patron@euriware.Fr (Patrice PATRON) : Subject: %SYSTEM-E-BADLOGIC on pmap_unset call VAX/VMS 6.2= Message-ID: <ee236176.0407122122.332fdd0e@posting.google.com>    Hi, 6 Some code which wak working fine some years ago, raiseF %SYSTEM-E-BADLOGIC on pmap_unset call VAX/VMS 6.2 ... Any body has got	 an idea ?   
 Best Regards.    Patrice    ------------------------------    Date: 12 Jul 2004 12:57:09 -0500. From: frey@encompasserve.org (Lurker at Large)+ Subject: Re: cluster_config extreme badness 3 Message-ID: <sSU4kSLeHuak@eisner.encompasserve.org>   d In article <mfPwPawg7bI7@eisner.encompasserve.org>, frey@encompasserve.org (Lurker at Large) writes: > $DEL_ROOT:/ > $say "    Deleting directory tree ''ROOT'..."   F 	forgive me for replying to my own note, but i'm still thinking about N this.  I think what is meant here is to delete the [SYS1] or whatever cluster O root.  Somehow it's deleting the node root instead.  But I still don't see any  K harm to letting the cluster root stay there; maybe output a message to the  K sysman that it's safe to delete [sys1...] and leave it up to him to decide   whether or not to do that.F 	It's just too dangerous to go around deleting *.*;* in a comfile and L assuming it picks up the logicals and symbols correctly enough to be in the C right place.  And it's NOT the expected response to a control-Y....    --      - Sharon, lurker at Large% http://mywebpages.comcast.net/jcwoman    ------------------------------  % Date: Mon, 12 Jul 2004 20:16:13 +0200 , From: "Hans Vlems" <hvlems.dotweg@zonnet.nl>+ Subject: Re: cluster_config extreme badness * Message-ID: <2lg2rqFbu7hiU1@uni-berlin.de>  = "Lurker at Large" <frey@encompasserve.org> schreef in bericht - news:sSU4kSLeHuak@eisner.encompasserve.org... L > In article <mfPwPawg7bI7@eisner.encompasserve.org>, frey@encompasserve.org (Lurker at Large) writes:  > > $DEL_ROOT:1 > > $say "    Deleting directory tree ''ROOT'..."  > F > forgive me for replying to my own note, but i'm still thinking aboutG > this.  I think what is meant here is to delete the [SYS1] or whatever  cluster L > root.  Somehow it's deleting the node root instead.  But I still don't see any L > harm to letting the cluster root stay there; maybe output a message to theL > sysman that it's safe to delete [sys1...] and leave it up to him to decide > whether or not to do that.F > It's just too dangerous to go around deleting *.*;* in a comfile andI > assuming it picks up the logicals and symbols correctly enough to be in  the E > right place.  And it's NOT the expected response to a control-Y....  >  > --   >  >  - Sharon, lurker at Large' > http://mywebpages.comcast.net/jcwoman   E None of my V5 or V6 VMS systems is up right now so I'm guessing here: I Could it be that the correct value for the symbol ROOT was not yet set at  the  time you entered ^Y?2 More specifically: at what question did you hit ^Y   Hans   ------------------------------  + Date: Tue, 13 Jul 2004 04:09:42 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) + Subject: Re: cluster_config extreme badness ( Message-ID: <ccvna6$qha$1@pcls4.std.com>  E VMS Engineering would be very interested in learning of such badness. F Could you be more specific in exactly what you were doing?  I take it B adding a root for a new node.  Exactly where did you ^Y it (which J question)?  AFAIK this hadn't been seen before, now you say you have seen F it twice.  Can you reproduce it again?  I suggest entering an IPMT so  it gets taken care of. --   -Mike    ------------------------------  % Date: Mon, 12 Jul 2004 20:20:12 +0200 , From: "Hans Vlems" <hvlems.dotweg@zonnet.nl>- Subject: Re: DECnet Phase IV Time Sync Needed * Message-ID: <2lg339Fc6g17U1@uni-berlin.de>  ; "Arthur Cochrane" <FACochrane@Yahoo.com> schreef in bericht - news:CtiCr4naWW$G@eisner.encompasserve.org... I >     We have two VMS systems that are going to be moved to a DECnet omly J >     link between the data center and a remote location. Why? Because theH >     Blaster virus last year caused so much traffic on the network thatK >     important operations were affected between these two sites. Right now I >     these systems sync their time via NTP. Once these move that ability ; >     goes away. These systems are running DECnet Phase IV.  > H >     My question is there a time sync over DECnet network we can use to keepL >     these systems in sync with each other? If not we will have to use wallH >     clock time sync (to a set time via the wall clock every so often).  E As others have pointed out, there's no such thing in DECnet phase IV. C If you don't want to upgrade to phase V, put them in an NI cluster. E Share nothing and use a quorum disk on each node. This makes shutting I one node rather painless, but is definitely an unsupported configuration.    Hans   ------------------------------  % Date: Mon, 12 Jul 2004 14:02:09 -0500 ( From: brandon@dalsemi.com (John Brandon)- Subject: Re: DECnet Phase IV Time Sync Needed 1 Message-ID: <04071214020934@dscis6-0.dalsemi.com>   I >     We have two VMS systems that are going to be moved to a DECnet omly J >     link between the data center and a remote location. Why? Because theH >     Blaster virus last year caused so much traffic on the network thatK >     important operations were affected between these two sites. Right now I >     these systems sync their time via NTP. Once these move that ability ; >     goes away. These systems are running DECnet Phase IV.  >     M >     My question is there a time sync over DECnet network we can use to keep L >     these systems in sync with each other? If not we will have to use wallH >     clock time sync (to a set time via the wall clock every so often).  9 Your statement suggests that you are moving off of TCPIP.   K If you are already running NTP and have been for some time then your system ! clocks should be fairly accurate.   
 Therefore ...   E You could use a DECnet object to collect and compare the time between H the two servers.  Designate a primary node and if the time is off on the7 secondary node then set the time on that node to match.    or  M You could introduce a third server into the cluster running TCPIP and NTP and > use that as your time source - coupled with the DECnet object.   or  M You could intermittently bring up TCPIP/NTP once a month and let it run for a ) day or two then shutdown the TCPIP stack.        J*o*h*n B*r*a*n*d*o*n  VMS Systems Administrator * firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  # Date: Mon, 12 Jul 2004 18:57:37 GMT ; From: "Duncan Macdonald" <duncanmacdonald@blueyonder.co.uk> - Subject: Re: DECnet Phase IV Time Sync Needed = Message-ID: <BaBIc.1723$ep2.20634723@news-text.cableinet.net>   L If you have a spare serial port at each end then you could syncronize to theL serial output of a GPS at each site. This should give you accuracy to betterD than 20ms with trivial coding, better than 1ms with moderate coding.  * Any GPS with a NEMA output should suffice.   ------------------------------  % Date: Mon, 12 Jul 2004 11:58:41 -0700  From: Z <z@no.spam> - Subject: Re: DECnet Phase IV Time Sync Needed 0 Message-ID: <10f5nqjbatr4k23@corp.supernews.com>   Arthur Cochrane wrote:M >     My question is there a time sync over DECnet network we can use to keep L >     these systems in sync with each other? If not we will have to use wall   With each other?  Yes.   $ MCR SYSMAN+ SYSMAN> SET ENVIRONMENT/NODES=(node1,node2) " SYSMAN> CONFIGURATION SET TIME ... SYSMAN> EXIT  F Some caveats apply ... the login acct must be privileged (eg: SYSTEM),1 exist on both systems and have the same password.    ------------------------------    Date: 12 Jul 2004 13:04:49 -0700( From: bob@instantwhip.com (Bob Ceculski)- Subject: Re: DECnet Phase IV Time Sync Needed = Message-ID: <d7791aa1.0407121204.6c3ca446@posting.google.com>   : Sorry again, node B routine should have been named IT.COM!  = why not run TCPwares Decnet Phase IV over IP ... then you can > run Phase IV over IP plus allow NTP ... but you can use a task( to do this ... I do ... it is simple ...  * on node A create a daily com routine to do  D $ DATE = "B::""TASK=IT " + F$CVTIME(,,"MONTH") + F$CVTIME(,,"DAY") -D                          + F$CVTIME(,,"YEAR") + F$CVTIME(,,"HOUR") -6                          + F$CVTIME(,,"MINUTE") + """" $ TYPE 'DATE   then on node B create IT.COM  < $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$OUTPUT SYS$NET; $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$ERROR SYS$NET P $ P2 = F$EXTRACT(2,2,P1) + "-" + F$CVTIME(F$EXTRACT(0,2,P1),"ABSOLUTE","MONTH")-         + "-" + F$EXTRACT(4,4,P1)> $ P2 = P2 + ":" + F$EXTRACT(8,2,P1) + ":" + F$EXTRACT(10,2,P1) $ SET TIME = 'P2N $ WRITE SYS$OUTPUT "Time on node " + F$GETJPI("","NODENAME") + " set to " + P2 $ EXIT  G with the appropriate proxy setting set.  For nodes in other time zones, O say pacific, just put a "-03:00" in the first position of each time criteria in . F$CVTIME command when setting the DATE symbol.  E It is really slick because you just pass the current time to the task @ routine on the node you want to set the time on ... pure genius!   ------------------------------    Date: 12 Jul 2004 13:00:44 -0700( From: bob@instantwhip.com (Bob Ceculski)- Subject: Re: DECnet Phase IV Time Sync Needed = Message-ID: <d7791aa1.0407121200.19753005@posting.google.com>   g FACochrane@Yahoo.com (Arthur Cochrane) wrote in message news:<CtiCr4naWW$G@eisner.encompasserve.org>... E > We have two VMS systems that are going to be moved to a DECnet omly J >     link between the data center and a remote location. Why? Because theH >     Blaster virus last year caused so much traffic on the network thatK >     important operations were affected between these two sites. Right now I >     these systems sync their time via NTP. Once these move that ability ; >     goes away. These systems are running DECnet Phase IV.  >     M >     My question is there a time sync over DECnet network we can use to keep L >     these systems in sync with each other? If not we will have to use wallH >     clock time sync (to a set time via the wall clock every so often).  = why not run TCPwares Decnet Phase IV over IP ... then you can > run Phase IV over IP plus allow NTP ... but you can use a task( to do this ... I do ... it is simple ...  * on node A create a daily com routine to do  D $ DATE = "A::""TASK=IT " + F$CVTIME(,,"MONTH") + F$CVTIME(,,"DAY") -D                          + F$CVTIME(,,"YEAR") + F$CVTIME(,,"HOUR") -6                          + F$CVTIME(,,"MINUTE") + """" $ TYPE 'DATE   then on node B create SDATE.COM   < $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$OUTPUT SYS$NET; $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$ERROR SYS$NET P $ P2 = F$EXTRACT(2,2,P1) + "-" + F$CVTIME(F$EXTRACT(0,2,P1),"ABSOLUTE","MONTH")-         + "-" + F$EXTRACT(4,4,P1)> $ P2 = P2 + ":" + F$EXTRACT(8,2,P1) + ":" + F$EXTRACT(10,2,P1) $ SET TIME = 'P2N $ WRITE SYS$OUTPUT "Time on node " + F$GETJPI("","NODENAME") + " set to " + P2 $ EXIT  G with the appropriate proxy setting set.  For nodes in other time zones, O say pacific, just put a "-03:00" in the first position of each time criteria in . F$CVTIME command when setting the DATE symbol.  E It is really slick because you just pass the current time to the task @ routine on the node you want to set the time on ... pure genius!   ------------------------------    Date: 12 Jul 2004 13:02:43 -0700( From: bob@instantwhip.com (Bob Ceculski)- Subject: Re: DECnet Phase IV Time Sync Needed = Message-ID: <d7791aa1.0407121202.3d49c939@posting.google.com>   C Sorry about the above, but notice I used node "B" when creating the 5 date command below, not "A" ... still pure genius! :)     = why not run TCPwares Decnet Phase IV over IP ... then you can > run Phase IV over IP plus allow NTP ... but you can use a task( to do this ... I do ... it is simple ...  * on node A create a daily com routine to do  D $ DATE = "B::""TASK=IT " + F$CVTIME(,,"MONTH") + F$CVTIME(,,"DAY") -D                          + F$CVTIME(,,"YEAR") + F$CVTIME(,,"HOUR") -6                          + F$CVTIME(,,"MINUTE") + """" $ TYPE 'DATE   then on node B create SDATE.COM   < $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$OUTPUT SYS$NET; $ IF F$MODE() .EQS. "NETWORK" THEN DEFINE SYS$ERROR SYS$NET P $ P2 = F$EXTRACT(2,2,P1) + "-" + F$CVTIME(F$EXTRACT(0,2,P1),"ABSOLUTE","MONTH")-         + "-" + F$EXTRACT(4,4,P1)> $ P2 = P2 + ":" + F$EXTRACT(8,2,P1) + ":" + F$EXTRACT(10,2,P1) $ SET TIME = 'P2N $ WRITE SYS$OUTPUT "Time on node " + F$GETJPI("","NODENAME") + " set to " + P2 $ EXIT  G with the appropriate proxy setting set.  For nodes in other time zones, O say pacific, just put a "-03:00" in the first position of each time criteria in . F$CVTIME command when setting the DATE symbol.  E It is really slick because you just pass the current time to the task @ routine on the node you want to set the time on ... pure genius!   ------------------------------    Date: 12 Jul 2004 14:34:12 -0500, From: FACochrane@Yahoo.com (Arthur Cochrane)- Subject: Re: DECnet Phase IV Time Sync Needed 3 Message-ID: <fgVA2mhNAAnY@eisner.encompasserve.org>   b In article <CtiCr4naWW$G@eisner.encompasserve.org>, FACochrane@Yahoo.com (Arthur Cochrane) writes:I >     We have two VMS systems that are going to be moved to a DECnet omly J >     link between the data center and a remote location. Why? Because theH >     Blaster virus last year caused so much traffic on the network thatK >     important operations were affected between these two sites. Right now I >     these systems sync their time via NTP. Once these move that ability ; >     goes away. These systems are running DECnet Phase IV.  >     M >     My question is there a time sync over DECnet network we can use to keep L >     these systems in sync with each other? If not we will have to use wallH >     clock time sync (to a set time via the wall clock every so often).  G     Thanks for all the suggestions. More information. These two systems G     will be in one location on a hub and connected via a router that is I     setup to ONLY route DECnet to another location with DECnet and IP. So C     these two systems will only be able to use DECnet to the remote I     location. DECnet Plus is NOT an option. I am just looking for an easy I     way to keep the time somewhat in sync. The SYSMAN CONFIG SET TIME may H     be the way to start looking or I might look at some DCL task to taskK     routine. I just thought someone might have had a solution from the past B     when DECnet was king and before IP and NTP took over the work.         Arthur Cochrane    ------------------------------  % Date: Tue, 13 Jul 2004 02:47:18 +0200 1 From: Milan Jercic <milan.jercicNOS@PAMaccess.si> 6 Subject: Re: Getting the RAB from BASIC channel number' Message-ID: <40F33116.407@PAMaccess.si>    > > > Why do you need  that, what problem are you trying to solve?  9 Ups, I obviously exaggerated trying to center on essence.   E We have a customer with comparatively big  in-house application (~7k  N modules/routines, 90% BASIC, rest in Macro-32, even some C in fractions). Our Q task is to "prepare it" for IA64 port, and while doing this "it would be nice to  K also find a solution to overcome this silly BASIC limitation of 119 opened  , channels at the same time" (customer quote).  O As OPEN ... CONNECT is heavily used the 119 channels limitation was frequently  G hit and the code is flooded with OPEN/OPEN...CONNECT]/CLOSE statements.   N My suggestion was to develop a set of routines (in C) to perform I/O directly J with FAB/RAB and use them for new development. All this new stuff will be A introduced slowly which is good for stability but raises various  Q interoperability questions (e.g. OPENing another file on same channel closes the  S previous one and releases all locks. Not so if opening new one with SYS$OPEN, etc.)   Q Needless to say, I can't modify OPEN statements to introduce USEROPEN clause. In  M fact USEROPEN is already used in some places to flag RU journaling for newly   created files.  ? > I just thought of a cute way to get the FAB, (not the RAB)... = > Code up an extra OPEN with USEROPEN my_useropenm CONNECT #n K > The Useropen willl be called with the address of the fab for the old rab, D > and a new rab address, which is pretty much a copy of the old rab. >   G Nice trick, but CONNECT clause needs channel number. Which one to use?  H Especially if no one is using LIB${GET|FREE}_LUN (but for BASIC this is N applicable only in 119..100 range, so you have to perform additional check on  Alpha).    > H >>Are CCB$PUSH/CCB$POP meant for this? If so, how should be they called?E > They should not be called, those are internal routines. No support.  >   . Don't bother. I'm just looking for quick fiks.  & > Can you get VMS listings for basrtl?  Q Um, probably. But not for free/cheap. Have to check for price (any discounts via  8 DSPP?). It's hard to justify purchase just for this one.  @ > Just have a basic program call GETRFA and debug with SET/INTO.! > Expect it to do something like: * >    DBASIC$$CB_PUSH (.CHAN,    1,   CCB);D >    if this-that-and-the-other-thing are all valid: CH$MOVE (6, CCB > [RAB$W_RFA], .RFA) >    DBASIC$$CB_POP (CCB);K >   Hint: The RAB is (main) part of the CCB, it start with x4401, should be  > easy to find with a debugger.  >   ! I like c.o.v for hints like this!   $ Now, for DBASIC$$CB_PUSH(chn,n,ccb):  )    chn - channel number, by value (long?)   :    n - number of CCB addresses to return? by value (long?)  (    ccb - address of CCB, address by ref?  > How comes it that RAB is unaligned inside CCB (^x4401 offset)?     Many thanks!     Regards,  
        J.M   ------------------------------  % Date: Tue, 13 Jul 2004 01:24:34 -0400 4 From: "Hein RMS van den Heuvel" <hein_at_hp_dot_com>6 Subject: Re: Getting the RAB from BASIC channel number* Message-ID: <40f374a0@usenet01.boi.hp.com>  > "Milan Jercic" <milan.jercicNOS@PAMaccess.si> wrote in message! news:40F33116.407@PAMaccess.si...  > > @ > > Why do you need  that, what problem are you trying to solve? : F >this silly BASIC limitation of 119 opened  channels at the same time" (customer quote).   L Yeah, that is restictive isn't it. Specially since VMS is happy to do 299 or so for several years/   E > As OPEN ... CONNECT is heavily used the 119 channels limitation was 
 frequentlyI > hit and the code is flooded with OPEN/OPEN...CONNECT]/CLOSE statements.    Yuck.   E I supposed a close and future re-open on an open-connected channel is  relatively cheap. I Open-connect is relatively sophisticated. Do they really know/need to use  multiple rmsF stream contexts, or is it just a 'handy thing to do'. That is, do they< really have multiple get-next contexts within a single file?  F > My suggestion was to develop a set of routines (in C) to perform I/O directlyK > with FAB/RAB and use them for new development. All this new stuff will be B > introduced slowly which is good for stability but raises variousG > interoperability questions (e.g. OPENing another file on same channel 
 closes theE > previous one and releases all locks. Not so if opening new one with  SYS$OPEN, etc.)   A Yeah, but you'd be walking away froma lot of Basic's added value. 6 Then you might just as well program in C directly :-(.  G > Needless to say, I can't modify OPEN statements to introduce USEROPEN 
 clause. InH > fact USEROPEN is already used in some places to flag RU journaling for newly  created files.   & They are sophisticated userds! Goodie!  H > Nice trick, but CONNECT clause needs channel number. Which one to use?   Yuck...   D > Um, probably. But not for free/cheap. Have to check for price (any
 discounts via : > DSPP?). It's hard to justify purchase just for this one.  L I thought is was in the hundreds, not thousands of $$$. It may well be worth your time/effortF Maybe a support/sales person feels justified to help you to at least a basrtl listing zip?     K >> >   Hint: The RAB is (main) part of the CCB, it start with x4401, should  be  + >    chn - channel number, by value (long?) = The debugger will show you this soon enough, and don't forget  BASIC/LIST/MACHIN   A >> How comes it that RAB is unaligned inside CCB (^x4401 offset)?   G I wasn't clear. The CONTENTS starts with 4401 (as any rab does). (hint:  Debugger has a search :-)      Hein.    ------------------------------  % Date: Mon, 12 Jul 2004 21:40:58 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> < Subject: Re: interfacing dec basic to web based applications: Message-ID: <L4HIc.11130$TB3.356166@news20.bellglobal.com>  : "Craig Marburger" <craigm2@earthlink.net> wrote in message< news:Z0VFc.5195$oD3.5142@newsread1.news.pas.earthlink.net...D > I have data in rms databases and that uses dec basic.  How would I	 interface C > the database with the internet(http web based).  What software or 
 interfacesJ > would I use to accomplish this?  Is there any web links with information on9 > how to accomplish this.  I have tcpip, vms Version 7.2.  >  > Craig Marburger  > craig@marburgerdairy.com >  >   L We do this all the time. Our BASIC applications are first started by runningJ a DCL script under Apache directory [.cgi-bin] where we set a logical nameH called "FROM_WEB". When our apps start up they look for this logical andJ then send output to either FMS for interactive users or channel #0 for web apps.   K If starting from the web, you can either send a static web page to the user J or a dynamic web page built from BASIC statements. Using some system callsK will allow you to determine whether the browser was using either the GET or J POST method, then we return reports or whatever based upon choices made in the browser.  
 Neil Rieck   Kitchener/Waterloo/Cambridge,    Ontario, Canada.  8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------    Date: 12 Jul 2004 13:14:49 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) , Subject: Re: MySQL 4.1.3-beta on ODS2 volume3 Message-ID: <jKjkOGff5VxY@eisner.encompasserve.org>   f In article <cc5619f2.0407120835.75c5b450@posting.google.com>, jordan@ccs4vms.com (Rich Jordan) writes: > H > So what is happening out there in unix-land that all these things seem( > to have become so depressingly common?  D    UNIX will allow any byte in a file name, even if its not good forH    you.  Folks who don't know better, have no exposure to other systems,B    and don't know the first thing about portability emulate things'    they've seen, with all thier faults.    ------------------------------    Date: 12 Jul 2004 13:12:41 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) > Subject: Re: New vms install, problem with serial console opa03 Message-ID: <$p9x+bhXhnww@eisner.encompasserve.org>   f In article <22e43e71.0407120537.25a438c@posting.google.com>, rcwood0000@yahoo.com (Roy C Wood) writes: > A > I guess I should qualify my question. I don't neccessarily need 
 > decwindows.   D > I need to be able to login to do editing like with TPU or EDT (not > command line editing :*).   . > Is there a trick I am missing in my old age?  B    The VT52 emulation in some of the consoles will not respond to ,    set term/inq, nor is it VT100 compatable.  &    Did you try explicit set term/vt52?  A    DECwindows offers DECterm, which does a VT300 emulation (VT100 -    compatable), which is what most of us use.   A    Alternatively, get a used VT and plug it into the serial port.    ------------------------------  % Date: Mon, 12 Jul 2004 17:15:08 -0700  From: Z <z@no.spam> 5 Subject: No Ethernet device on system (UCX 4.2 ECO 5) 0 Message-ID: <10f6abshp42pe3f@corp.supernews.com>  = I'm trying to remove all current UCX (V4.2 ECO 5) settings on @ an inherited Alphaserver (OpenVMS 7.1-2) ... so I shut down UCX,B renamed the UCX$*.DAT files in SYS$SYSTEM and then ran UCX$CONFIG.  ? UCX$CONFIG recreated new .DAT files and all seemed fine until I C tried to define the Interface (in Core).  The interface is/was IE0, 7 but now I'm getting a repeatable error from UCX$CONFIG:    	"No Ethernet device on system"    and then UCX$CONFIG aborts.    Arrgh!   What am I doing wrong?   ------------------------------  % Date: Mon, 12 Jul 2004 14:57:32 -0500 ( From: brandon@dalsemi.com (John Brandon)6 Subject: odd things with SYS$GETUAI and uai$_user_data1 Message-ID: <04071214573260@dscis6-0.dalsemi.com>   L Using the SYS$GETUAI and SYS$SETUAI I am maintaining employee information in& the SYSUAF file (uai$_user_data area).  N The application I wrote up works for almost every account (1500 or so) howeverL does not work on a handful (12 or less).  The application READS (getuai) the? uai$_user_data then changes the value and then WRITES (setuai).   M I was racking my brains out on this and decided to delete the problem account B and re-add it.  Subsequently, the update to uai$_user_data worked.   Has anyone experienced this?     J*o*h*n B*r*a*n*d*o*n  VMS Systems Administrator * firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  # Date: Mon, 12 Jul 2004 22:06:32 GMT # From: hoff@hp.nospam (Hoff Hoffman) : Subject: Re: odd things with SYS$GETUAI and uai$_user_data0 Message-ID: <IXDIc.5711$_u.867@news.cpqcorp.net>  \ In article <04071214573260@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:N :...maintaining employee information in the SYSUAF file (uai$_user_data area). ..O :The application I wrote up works for almost every account (1500 or so) however M :does not work on a handful (12 or less).  The application READS (getuai) the @ :uai$_user_data then changes the value and then WRITES (setuai). : N :I was racking my brains out on this and decided to delete the problem accountC :and re-add it.  Subsequently, the update to uai$_user_data worked.  :  :Has anyone experienced this?     I   FWIW, I don't generally recommend using the UAF for the storage of user H   data, the existence of the itemcode aside.   This for various reasons,F   not the least of which is a complete lack of coordination across any$   potential users of UAI$_USER_DATA.  G   Check that there is not already somebody else using the SYSUAF record G   in other words, or something else odd about the particular UAF record 	   itself.   H   It is much safer -- and just as supported -- to use a private file forH   this and other data.  I use tools which regularly synchronize with theF   current records stored in the SYSUAF, keeping the side database file@   current with the status of the records in the main SYSUAF ile.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------    Date: 12 Jul 2004 18:14:04 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) : Subject: Re: odd things with SYS$GETUAI and uai$_user_data3 Message-ID: <5l2UQ8LzT806@eisner.encompasserve.org>   \ In article <04071214573260@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:N > Using the SYS$GETUAI and SYS$SETUAI I am maintaining employee information in( > the SYSUAF file (uai$_user_data area). > P > The application I wrote up works for almost every account (1500 or so) howeverN > does not work on a handful (12 or less).  The application READS (getuai) theA > uai$_user_data then changes the value and then WRITES (setuai).  > O > I was racking my brains out on this and decided to delete the problem account D > and re-add it.  Subsequently, the update to uai$_user_data worked.  F Are you certain you are using the first byte of the field as a count ?   ------------------------------    Date: 13 Jul 2004 00:04:38 -0500+ From: young_r@encompasserve.org (Rob Young) : Subject: Re: odd things with SYS$GETUAI and uai$_user_data3 Message-ID: <dCDxRpLUZOV9@eisner.encompasserve.org>   V In article <IXDIc.5711$_u.867@news.cpqcorp.net>, hoff@hp.nospam (Hoff Hoffman) writes:^ > In article <04071214573260@dscis6-0.dalsemi.com>, brandon@dalsemi.com (John Brandon) writes:P > :...maintaining employee information in the SYSUAF file (uai$_user_data area). > ..Q > :The application I wrote up works for almost every account (1500 or so) however O > :does not work on a handful (12 or less).  The application READS (getuai) the B > :uai$_user_data then changes the value and then WRITES (setuai). > : P > :I was racking my brains out on this and decided to delete the problem accountE > :and re-add it.  Subsequently, the update to uai$_user_data worked.  > :  > :Has anyone experienced this?  >  > K >   FWIW, I don't generally recommend using the UAF for the storage of user J >   data, the existence of the itemcode aside.   This for various reasons,H >   not the least of which is a complete lack of coordination across any& >   potential users of UAI$_USER_DATA. > I >   Check that there is not already somebody else using the SYSUAF record I >   in other words, or something else odd about the particular UAF record  >   itself.  > J >   It is much safer -- and just as supported -- to use a private file forJ >   this and other data.  I use tools which regularly synchronize with theH >   current records stored in the SYSUAF, keeping the side database fileB >   current with the status of the records in the main SYSUAF ile. >   < 	That field is a good place to stuff information that normalD 	prying eyes won't see.  Additionally, the owner field is not a goodD 	place as finger will show that.  Yes, go around and disable finger.  B 	But in the age of Identity Management - it is necessary to *know*< 	who you are disusering/removing.  As the employee count sky9 	- rockets , there are many opportunities for error short B 	of having identifiable information at the ready.  Create separate< 	tools, methods?  Why?  setuai/getuai and the uai$_user_data= 	field is there and numerous Googles don't even reveal a hint = 	of "unsafeness."  Separate tools/methods?  One more thing to = 	support and sell to management?  I don't think so.  Creative A 	solutions when looked at years later are often very very painful H 	to support (unless of course Hoff or equivalent are writing/documenting 	).     @ 	And oh... when that side database gets stomped on or corrupted.D 	A flat file?  How does it scale?  Insertions, deletions.  An index - 	file?  And who does that bit of programming?   ; 	Any hints as to why things aren't safe with setuai/getuai?    				Rob    ------------------------------    Date: 12 Jul 2004 12:54:11 -05004 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)4 Subject: Re: OpenVMS 7.3-1 on Alpha Wokstation 500au3 Message-ID: <GJOVaUByQ8pk@eisner.encompasserve.org>   Z In article <40ED8C1D.3070500@goatley.com>, Hunter Goatley <goathunter@goatley.com> writes: > Steve O'Hara wrote:  >  >> Hi, >>  H >> The odd thing is that the CD does actually boot i.e. it's finding theM >> bottstrap code on the start of the CD that contains the VMS banner........e: >> I'm waiting for SCSI CD to turn so that I can try that. >> eC > You said you set the os_type to OpenVMS.  Did you INIT the systemiA > afterward?  On my 600au, I had the same problem you did until Ia? > did the re-INIT so the system really knew it was OpenVMS.  At ? > that time, my dk devices became dq devices, and I was able toe+ > successfully boot VMS from the dq device.e > @ > My 500au has all DK devices, so I'm not sure if the above will > help you at all or not.  >  > Hunter  E I had an odd quirk with my 500au where it would only boot completely  D after a power on.  If I rebooted it without a power off, or changed B a console variable and then tried to reboot, it would hang at the C banner.  Powering it all the way off and then back on would get it dF to boot completely.  I never figured out why, but since the workaroundF worked, I didn't spend too much time trying to figure it out (I mostly cold boot the thing anyway).   ------------------------------  % Date: Mon, 12 Jul 2004 16:05:55 -0400i' From: "Main, Kerry" <kerry.main@hp.com>p4 Subject: RE: OpenVMS 7.3-1 on Alpha Wokstation 500auR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB394CED@tayexc19.americas.cpqcorp.net>   > -----Original Message-----@ > From: Marty Kuhrt [mailto:kuhrt@nospammy.encompasserve.org]=20 > Sent: July 12, 2004 1:54 PM- > To: Info-VAX@Mvb.Saic.Com-6 > Subject: Re: OpenVMS 7.3-1 on Alpha Wokstation 500au >=20> > In article <40ED8C1D.3070500@goatley.com>, Hunter Goatley=20" > <goathunter@goatley.com> writes: > > Steve O'Hara wrote:- > >=20 > >> Hi, > >>=202A > >> The odd thing is that the CD does actually boot i.e. it's=20e
 > finding the,? > >> bottstrap code on the start of the CD that contains the=20  > VMS banner........< > >> I'm waiting for SCSI CD to turn so that I can try that. > >>=20aE > > You said you set the os_type to OpenVMS.  Did you INIT the system C > > afterward?  On my 600au, I had the same problem you did until IUA > > did the re-INIT so the system really knew it was OpenVMS.  AtsA > > that time, my dk devices became dq devices, and I was able toi- > > successfully boot VMS from the dq device.o > >=20B > > My 500au has all DK devices, so I'm not sure if the above will > > help you at all or not.t > >=20
 > > Hunter >=20I > I had an odd quirk with my 500au where it would only boot completely=20eH > after a power on.  If I rebooted it without a power off, or changed=20F > a console variable and then tried to reboot, it would hang at the=20G > banner.  Powering it all the way off and then back on would get it=20tH > to boot completely.  I never figured out why, but since the workaroundH > worked, I didn't spend too much time trying to figure it out (I mostly > cold boot the thing anyway).   Marty,  C I had a similar issue with old DEC3000 and the problem was fixed by / increasing the SCSI_RESET console parameter.=20   G This is console delay before it starts to access the scsi devices after F an init or power on. As I recall, some scsi devices (e.g. Cdrom) wouldG respond to an INIT with self reset and would not be completed this self C check when the system was trying to access the device. Sometimes itHH worked, sometimes it did not. After increasing the delay, I never had an issue like this.   YMMV ..r   Regardsr  
 Kerry Main Senior ConsultantR HP Services Canada Voice: 613-592-4660e Fax: 613-591-4477  Email: kerryDOTmainAThpDOTcom . (remove the DOT's and AT for email address)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Mon, 12 Jul 2004 16:53:01 -0500t/ From: Chris Scheers <chris@applied-synergy.com>a4 Subject: Re: OpenVMS 7.3-1 on Alpha Wokstation 500au3 Message-ID: <40F3083D.E54C3D25@applied-synergy.com>t   Marty Kuhrt wrote: > F > I had an odd quirk with my 500au where it would only boot completelyE > after a power on.  If I rebooted it without a power off, or changed C > a console variable and then tried to reboot, it would hang at thepD > banner.  Powering it all the way off and then back on would get itH > to boot completely.  I never figured out why, but since the workaroundH > worked, I didn't spend too much time trying to figure it out (I mostly > cold boot the thing anyway).    . I've noticed this on my AlphaStation 255 also.  F I found that it was necessary to enter an INIT command before it would reboot.   A I'll have to go back and check on Kerry's SCSI_RESET suggestions.t  G -----------------------------------------------------------------------n$ Chris Scheers, Applied Synergy, Inc.  C Voice: 817-237-3360            Internet: chris@applied-synergy.com h   Fax: 817-237-3074    ------------------------------  % Date: Mon, 12 Jul 2004 15:06:20 -0400 # From: "John Smith" <a@nonymous.com>u8 Subject: Re: OpenVMS license transfer  policies and fees, Message-ID: <VNCdnfkMiq6sfG_dRVn-vw@igs.net>   Rich Jordan wrote:0 > "John Smith" <a@nonymous.com> wrote in message* > news:<qPKdnbroP-X1iXLdRVn-gg@igs.net>... >- >> >> Click here:= >> http://www.hp.com/hpinfo/execteam/email/fiorina/index.htmlt >>& >> Paste your complaint and elaborate. >> Send. >> > 
 > Done, sent.2  K One never knows what may occur when one writes to carly(tm). I actually gotfI a call back from a minion once to gather more information. Have not heard " back from them for a while though.   ------------------------------    Date: 12 Jul 2004 13:52:40 -07000 From: dave.baxter@bannerhealth.com (Dave Baxter)? Subject: Re: OpenVMS Management Station -  cluster storage ????o= Message-ID: <a3c44af1.0407121252.1cd4bb96@posting.google.com>o  u keithparris_NOSPAM@yahoo.com (Keith Parris) wrote in message news:<cf15391e.0407091430.e0e6cc1@posting.google.com>...gw > dave.baxter@bannerhealth.com (Dave Baxter) wrote in message news:<a3c44af1.0407071017.5d2612b4@posting.google.com>...m  G > >     As far as I am aware, I have my SAN/CLUSTER configured for higheI > > redundancy, and therefore I am trying to figure out why my system hadp **
 **   SNIP  ** > G > > When a disk controller fails, it is supposed to failover all of thep4 > > drives it has responsibility for to its partner. > G > Yes, it is supposed to. I've run into a few cases over the years whens
 > it hasn't, u     Thanks for your comments Keith, :     Just for information, here a summary of what we found:  D HSOF firmware V8.6F-7    (discovered that current patch level should be 13)  4 FMU indicated that the failed controller was due to 7                    "An error has occurred on the PDAL."aA which apparently indicates a problem on one of the 6 SCSI busses.   C Our HP engineer indicated that the most likely problem was that thetC failed controller/SCSI buss was refusing to release the LUN so thatr% the other controller could take over.s  F The effect of this was that the units never come out of Mount Verify. C   When the MVTIMEOUT (3600 secs) is reached, the SHADOW_SERVER will D try to eject the failed unit.    It will try to do this for 120 secs? (SHADOW_MBR_TMO), and if it fails it will take the whole volumec. offline.     This appears to be what happened.9      There is one big problem with this scenario which isoE inconsistent.    The first indication of "mount verify" was at ~19:10 A our time, which is consistent with the contoller failure (~19:09)r> however the shadow volumes started dropping off line very soon& (relatively) after that< ~19:20,19:30.  @      All I can think is that MVTIMEOUT was set much lower than IC think, (maybe dynamically by someone dealing with the problem in myiB absence).    Of course, after the system booted it was back to the default.  D Regarding the Controller failure, I have since discovered that thereB is a patch (#12) for this version of the HSOV "Title:   PersistentF Reservation and LUN hang.     This patch may have avoided my problem. 4 My contollers ar now all upgraded to patch level 13.   Thanks to all who commented.   Dave   ------------------------------    Date: 12 Jul 2004 13:53:11 -07000 From: dave.baxter@bannerhealth.com (Dave Baxter)? Subject: Re: OpenVMS Management Station -  cluster storage ???? = Message-ID: <a3c44af1.0407121253.49d36692@posting.google.com>3  u keithparris_NOSPAM@yahoo.com (Keith Parris) wrote in message news:<cf15391e.0407091430.e0e6cc1@posting.google.com>... w > dave.baxter@bannerhealth.com (Dave Baxter) wrote in message news:<a3c44af1.0407071017.5d2612b4@posting.google.com>...a  G > >     As far as I am aware, I have my SAN/CLUSTER configured for high-I > > redundancy, and therefore I am trying to figure out why my system hadm **
 **   SNIP  ** > G > > When a disk controller fails, it is supposed to failover all of theo4 > > drives it has responsibility for to its partner. > G > Yes, it is supposed to. I've run into a few cases over the years whenn
 > it hasn't,       Thanks for your comments Keith,l:     Just for information, here a summary of what we found:  D HSOF firmware V8.6F-7    (discovered that current patch level should be 13)  4 FMU indicated that the failed controller was due to 7                    "An error has occurred on the PDAL." A which apparently indicates a problem on one of the 6 SCSI busses.3  C Our HP engineer indicated that the most likely problem was that theCC failed controller/SCSI buss was refusing to release the LUN so thata% the other controller could take over.,  F The effect of this was that the units never come out of Mount Verify. C   When the MVTIMEOUT (3600 secs) is reached, the SHADOW_SERVER willoD try to eject the failed unit.    It will try to do this for 120 secs? (SHADOW_MBR_TMO), and if it fails it will take the whole volume:. offline.     This appears to be what happened.9      There is one big problem with this scenario which is$E inconsistent.    The first indication of "mount verify" was at ~19:10"A our time, which is consistent with the contoller failure (~19:09)E> however the shadow volumes started dropping off line very soon& (relatively) after that< ~19:20,19:30.  @      All I can think is that MVTIMEOUT was set much lower than IC think, (maybe dynamically by someone dealing with the problem in my B absence).    Of course, after the system booted it was back to the default.  D Regarding the Controller failure, I have since discovered that thereB is a patch (#12) for this version of the HSOV "Title:   PersistentF Reservation and LUN hang.     This patch may have avoided my problem. 4 My contollers ar now all upgraded to patch level 13.   Thanks to all who commented.   Dave   ------------------------------    Date: 12 Jul 2004 13:55:37 -07000 From: dave.baxter@bannerhealth.com (Dave Baxter)? Subject: Re: OpenVMS Management Station -  cluster storage ????r= Message-ID: <a3c44af1.0407121255.7597577c@posting.google.com>s  u keithparris_NOSPAM@yahoo.com (Keith Parris) wrote in message news:<cf15391e.0407091430.e0e6cc1@posting.google.com>...cw > dave.baxter@bannerhealth.com (Dave Baxter) wrote in message news:<a3c44af1.0407071017.5d2612b4@posting.google.com>...m > G > > When a disk controller fails, it is supposed to failover all of thei4 > > drives it has responsibility for to its partner. > G > Yes, it is supposed to. I've run into a few cases over the years wheneH > it hasn't, but that's the very reason why you shadow between differentA > controller pairs -- that should have covered that unlikely, buts > possible, eventD   Thanks for your comments Keith,q:     Just for information, here a summary of what we found:  D HSOF firmware V8.6F-7    (discovered that current patch level should be 13)  4 FMU indicated that the failed controller was due to 7                    "An error has occurred on the PDAL."hA which apparently indicates a problem on one of the 6 SCSI busses.T  C Our HP engineer indicated that the most likely problem was that the C failed controller/SCSI buss was refusing to release the LUN so thatT% the other controller could take over.D  F The effect of this was that the units never come out of Mount Verify. C   When the MVTIMEOUT (3600 secs) is reached, the SHADOW_SERVER willAD try to eject the failed unit.    It will try to do this for 120 secs? (SHADOW_MBR_TMO), and if it fails it will take the whole volumeH. offline.     This appears to be what happened.9      There is one big problem with this scenario which is"E inconsistent.    The first indication of "mount verify" was at ~19:101A our time, which is consistent with the contoller failure (~19:09)"> however the shadow volumes started dropping off line very soon& (relatively) after that< ~19:20,19:30.  @      All I can think is that MVTIMEOUT was set much lower than IC think, (maybe dynamically by someone dealing with the problem in mydB absence).    Of course, after the system booted it was back to the default.  D Regarding the Controller failure, I have since discovered that thereB is a patch (#12) for this version of the HSOV "Title:   PersistentF Reservation and LUN hang.     This patch may have avoided my problem. 4 My contollers ar now all upgraded to patch level 13.   Thanks to all who commented.   Dave    P --------------------------------------------------------------------------------   ------------------------------  % Date: Mon, 12 Jul 2004 15:08:48 -0400r# From: "John Smith" <a@nonymous.com>o' Subject: Re: OpenVMS Positions - EuropeT, Message-ID: <Pu2dnTDgmLVYfG_d4p2dnA@igs.net>   David M Smith wrote:C > On Mon, 12 Jul 2004 10:02:01 -0400, "John Smith" <a@nonymous.com>= > wrote: >M >> Bill Gunshannon wrote:D7 >>> In article <PYp2t7sSE3E9@eisner.encompasserve.org>,$A >>> koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: 8 >>>> In article <1Sibu84jMShu@eisner.encompasserve.org>,4 >>>> Kilgallen@SpamCop.net (Larry Kilgallen) writes: >>>>>TH >>>>> And when it comes right down to it, Object Oriented programming is' >>>>> _not_ the answer to all problems.( >>>>) >>>>    Is it the answer to any problems?T >>>4? >>> There are some of us, even in academia, who don't think so.2 >> >>D >> Is there a lucid discussion of the pro/con arguments of this that >> you could recommend?  >> > A > John, I read an interesting article called Objects Have Failed,z2 > presented at the OOPSLA convention in 2002, see: >o+ > http://oopsla.acm.org/fp/files/pan-1.htmlm >dC > You can find copies of this paper on the Internet by searching on F > Googol for the file name objectshavefailed.pdf. I don't know whetherA > a paper for the other side of the debate is available or not...a     Thanks.0   ------------------------------    Date: 12 Jul 2004 15:45:13 -0700, From: jeremybarker@email.com (Jeremy Barker)E Subject: Re: OpenVMS probably the best designed o/s in  the market...a= Message-ID: <5b86b9ee.0407121445.337d8575@posting.google.com>c  a mb301@hotmail.com (MB) wrote in message news:<1d08b916.0407090143.2e2eaeb8@posting.google.com>...k< > Does anyone have next week copy of Comperter Weekly yet... > (6-Jul-2004) > @ > OpenVMS is probably one of the best designed and most reliableC > operating systems on the market, so why is it unnoticed and under  > apprecicated?m   1. It doesn't run on a "PC".  @ 2. It doesn't come from a corportaion whose name begines with M.  F 3. Because it is reliable it doesn't give rise to "incidents" that hit the headlines.  1 4. You can't buy it at your local computer store.c   I could go on.  ! jb (ex-DEC, these days a lawyer).i   ------------------------------   End of INFO-VAX 2004.384 ************************stions. More information. These two systems G     will be in one location on a hub and connected via a router that is I     setup to ONLY route DECnet to another location with DECnet and IP. So C     these has been to persuade countries
to cooperate with the <NaviLink ID='31:2,41:1'>UN's</NaviLink> nuclear inspection regime and non-proliferation
agreements, and to <NaviLink ID='19:1'>deal</NaviLink> with signatory states who flout their commitments.</P>
<P>   He must now counsel caution from <NaviLink ID='11:1'>UN Security Council</NaviLink> members and seek more
time for inspectors to search for an alleged <NaviLink ID='15:1'>nuclear weapons</NaviLink> program in <NaviLink ID='9:3,25:6,26:5,3:7,14:8,22:9