1 INFO-VAX	Thu, 26 May 2005	Volume 2005 : Issue 292       Contents: Re: ABC problem with disk  Re: ABC problem with disk  Re: ABC problem with disk  Re: ABC problem with disk  Re: ABC problem with disk # Computerworld: "Ready for trouble?" ' Re: Computerworld: "Ready for trouble?" ' Re: Computerworld: "Ready for trouble?" ! DECwindows screen lock and PWDMIX % Re: DECwindows screen lock and PWDMIX P HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors but P Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors P Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors P Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors  HP PaRisc, IA64 and NSK news4 Re: HP Pulls Off a Respectable Second Fiscal Quarter4 Re: HP Pulls Off a Respectable Second Fiscal QuarterP Re: I want to ask you the most important question of your life. The question is: Re: Installing a printer Re: Installing a printer Re: Installing a printer SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: SEARCH : AND NOT ? Re: TCPIP$TELNETSYM @ Re: Third party Windows App to NFS map OpenVMS to Windows Server@ Re: Third party Windows App to NFS map OpenVMS to Windows Server  F ----------------------------------------------------------------------    Date: 26 May 2005 07:15:34 -0700) From: "Joe Sewell" <ultrajoe@spamcop.net> " Subject: Re: ABC problem with diskC Message-ID: <1117116934.731188.158840@o13g2000cwo.googlegroups.com>   E ABC is Archive Backup Client, as mentioned on the page you cite. This ' is version 4.0.0.0.1, if memory serves.    ------------------------------    Date: 26 May 2005 07:17:53 -0700) From: "Joe Sewell" <ultrajoe@spamcop.net> " Subject: Re: ABC problem with diskC Message-ID: <1117117073.906853.263060@g47g2000cwa.googlegroups.com>   B The complete backup command is generated from DCL code, so I'm not9 precisely sure what it is. I believe it's something line:    $ ABC INCREMENTAL E $1$DKB100:/IGNORE=INTERLOCK/LOG=(something)/SUMMARY=(something) ! I'd - have to figure out what the "something's" are   ? The logs don't show anything except that it stops at that disk.    ------------------------------    Date: 26 May 2005 07:23:30 -0700) From: "Joe Sewell" <ultrajoe@spamcop.net> " Subject: Re: ABC problem with diskC Message-ID: <1117117410.521767.149840@f14g2000cwb.googlegroups.com>    >Joe Sewell wrote:G >> Our organization has just gone to using ABC for networked backups of I >> machines scattered throughout the campus.  My particular machine seems J >> to have a problem.  It has a 9-bay StorageWorks tower with 7 18GB disksD >> in it.  One of those disks seems to put ABC into an infinite loopJ >> scanning for backup candidates.  If I manually run ABC INCREMENTAL withH >> the /ALIAS=3DCOPY qualifier, it works just fine.  The default doesn'tG >> have the /ALIAS qualifier, and I've seen it churn for *days* without  >> backing up a single file. > 8 >> ANALYZE/DISK/REPAIR/READ/RECOR=ADD shows no problems. > 4 >> Anybody have any ideas what else to look at here? > F >Just to clarify, is ABC the product known as Archive Backup Client asE >documented at http://www.rdperf.com/RDHTML/l=ADiterature_main.html ?    Yes, version 4.0.0.1.    ------------------------------    Date: 26 May 2005 07:20:31 -0700) From: "Joe Sewell" <ultrajoe@spamcop.net> " Subject: Re: ABC problem with diskB Message-ID: <1117117231.357434.64620@g43g2000cwa.googlegroups.com>  I >Archive backup client, eh? Haven't heard of that since about 1998 or so.  >Not recommended then, either. > I >Use VMS BACKUP to backup VMS disks, or risk losing your data and/or your 2 >ability to perform disaster recovery expediently.  G Not an option. This was installed by the system administrator (not me). D The only reason I even noticed its existence on my machine is due toG some system slowdown that I tracked down to this thing chewing up about  5% CPU time in kernel mode.    ------------------------------  % Date: Thu, 26 May 2005 11:29:30 -0400 - From: William Webb <william.w.webb@gmail.com> " Subject: Re: ABC problem with disk7 Message-ID: <8660a3a105052608292424b0f5@mail.gmail.com>   G On 26 May 2005 07:23:30 -0700, Joe Sewell <ultrajoe@spamcop.net> wrote:  > >Joe Sewell wrote:I > >> Our organization has just gone to using ABC for networked backups of K > >> machines scattered throughout the campus.  My particular machine seems L > >> to have a problem.  It has a 9-bay StorageWorks tower with 7 18GB disk= s F > >> in it.  One of those disks seems to put ABC into an infinite loopL > >> scanning for backup candidates.  If I manually run ABC INCREMENTAL wit= h J > >> the /ALIAS=3DCOPY qualifier, it works just fine.  The default doesn'tI > >> have the /ALIAS qualifier, and I've seen it churn for *days* without  > >> backing up a single file. > > : > >> ANALYZE/DISK/REPAIR/READ/RECOR=ADD shows no problems. > > 6 > >> Anybody have any ideas what else to look at here? > > H > >Just to clarify, is ABC the product known as Archive Backup Client asG > >documented at http://www.rdperf.com/RDHTML/l=ADiterature_main.html ?  >=20 > Yes, version 4.0.0.1.  >=20 >=20  F I'd try creating a VMS backup job that does a full backup on that disk and nothing else.   A If it works, and the ABC backup doesn't, then you've eliminated a / _whole lot of possibilities_ in one fell swoop.   : You've also made a point to your superiors in the process.   WWWebb --=20 C NOTE: This email address is only used for noncommerical VMS-related  correspondence. C All unsolicited commercial email will be deemed to be a request for 8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/    ------------------------------  % Date: Thu, 26 May 2005 09:26:01 +0200 - From: Didier Morandi <prenom.nom@freesurf.fr> , Subject: Computerworld: "Ready for trouble?"3 Message-ID: <42957a0a$0$6831$636a15ce@news.free.fr>   C http://www.computerworld.com.au/index.php/id;209495052;fp;16;fpid;0    ../..  State-of-the-art technology   H While opinions vary as to what constitutes state-of-the-art technology, B experts such as Karp of Enterprise Management Associates and Chip F Nickolett, a disaster recovery specialist at Comprehensive Consulting G Solutions, agree that clustering, SAN mirroring and replication are on  A the leading edge. However, they warn that these can be expensive  
 technologies.   H Among operating systems, OpenVMS and Unix seem to be favoured more than G others. Alpha/OpenVMS, for example, has built-in clustering technology  E that many companies use to mirror data between sites. Many financial  B institutions, including Commerzbank, the International Securities D Exchange and Deutsche Borse, rely on VMS-based mirroring to protect 0 their heavy-duty transaction-processing systems.  B Deutsche Borse, a German exchange for stocks and derivatives, has D deployed an OpenVMS cluster over two sites situated five kilometres G apart. It also uses Fibre Channel switches from Brocade Communications  E Systems and Cisco switches and routers in its network to ensure high  G availability. "DR is not about cold or warm backups, it's about having  I your data active and online no matter what," says Michael Gruth, head of  F systems and network support at Deutsche Borse. "That requires cluster * technology which is online at both sites." ../..    FYI.   D. --  1 Didier MORANDI - Expert informaticien - VMS / SAP 0   13 chemin du Gu, 1213 Petit-Lancy (GE) Suisse0 Tl. : +33(0)6 7983 6418 ~ www.didiermorandi.com   ------------------------------  % Date: Thu, 26 May 2005 04:21:55 -0400 ( From: Bill Todd <billtodd@metrocast.net>0 Subject: Re: Computerworld: "Ready for trouble?"= Message-ID: <vuednZPUV6a5GgjfRVn-pA@metrocastcablevision.com>    Didier Morandi wrote: E > http://www.computerworld.com.au/index.php/id;209495052;fp;16;fpid;0  >  > ../..  > State-of-the-art technology  > J > While opinions vary as to what constitutes state-of-the-art technology, D > experts such as Karp of Enterprise Management Associates and Chip H > Nickolett, a disaster recovery specialist at Comprehensive Consulting I > Solutions, agree that clustering, SAN mirroring and replication are on   > the leading edge.   F Gee - what do they suppose they were 20 years ago, then?  Well beyond K the imagination, let alone experience, of 'experts' such as these, I guess.   I Too bad the owners of this technology started losing appreciation for it  F about the time it started to become 'leading edge' in the eyes of the H less informed, back in the '90s.  I suppose that may have been somewhat H more true of VMS's owners than it was for NSK's - somewhat ironic given I that VMS was a far more general-purpose product and thus far more widely  F applicable (not to detract from NSK, mind you:  what it does, it does A exceedingly well and arguably significantly better than VMS can).    - bill   ------------------------------  % Date: Thu, 26 May 2005 06:15:30 -0700 # From: "Tom Linden" <tom@kednos.com> 0 Subject: Re: Computerworld: "Ready for trouble?"( Message-ID: <opsrdwf4tdzgicya@hyrrokkin>  H On Thu, 26 May 2005 04:21:55 -0400, Bill Todd <billtodd@metrocast.net>   wrote:   > Didier Morandi wrote: F >> http://www.computerworld.com.au/index.php/id;209495052;fp;16;fpid;0	 >>  ../..  >> State-of-the-art technologyA >>  While opinions vary as to what constitutes state-of-the-art   I >> technology, experts such as Karp of Enterprise Management Associates   H >> and Chip Nickolett, a disaster recovery specialist at Comprehensive  C >> Consulting Solutions, agree that clustering, SAN mirroring and   ' >> replication are on the leading edge.  > I > Gee - what do they suppose they were 20 years ago, then?  Well beyond   H > the imagination, let alone experience, of 'experts' such as these, I   > guess. > L > Too bad the owners of this technology started losing appreciation for it  I > about the time it started to become 'leading edge' in the eyes of the   K > less informed, back in the '90s.  I suppose that may have been somewhat   K > more true of VMS's owners than it was for NSK's - somewhat ironic given   L > that VMS was a far more general-purpose product and thus far more widely  I > applicable (not to detract from NSK, mind you:  what it does, it does   C > exceedingly well and arguably significantly better than VMS can).  >  > - billD I would be interested to see a comparison of VMS and NSK in terms ofC probability of failuer and availability.  I suspect that for most    applicationsK VMS does it just as well as NSK.  If I were Hurd, I would be trying to sell # NSK, while it still has some value.    ------------------------------    Date: 26 May 2005 08:51:32 -0700# From: "Galen" <gltackett@gmail.com> * Subject: DECwindows screen lock and PWDMIXC Message-ID: <1117122692.769994.314970@o13g2000cwo.googlegroups.com>   G (The text here is mostly a copy of my posting on the VMS ITRC forum, so . you'll want to skip this if you've read that.)  A In AUTHORIZE I have set /FLAG=PWDMIX for my acccount, and I set a ) password that has a period in the middle.   C Now, if the DECwindows session lock activates (by clicking the lock G button or due to inactivity) I can't unlock it! It tells me my password B is invalid and generates a security audit. Yet I am able to log inG elsewhere with exactly the same password. (The status show in the audit  is %LOGIN-F-INVPWD.)  E I have found that I can temporarily fix this by resetting my password E via AUTHORIZE. This works even if I set it to the same password I was F already using. Yet once the screen lock activates, the problem happens all over again.   D I am running VMS V7.3-2 and DECwindows V1.3-1 with all the ECOs that$ were available as of about May 14th.  G Maybe the DECwindows screen lock is getting tripped up by examining the 1 password itself, rather than letting VMS do this?   / (I later posted this to the ITRC topic as well:   G I should also say that this problem has been happening ever since I set < this system up in January, before most of the ECOs went on.)   ------------------------------  % Date: Thu, 26 May 2005 12:57:56 -0400 $ From: "PEN" <paul.nuneznosp@mhp.com>. Subject: Re: DECwindows screen lock and PWDMIX, Message-ID: <d74v6m$l7j$1@hplms2.hpl.hp.com>  	 Hi Galen,   / "Galen" <gltackett@gmail.com> wrote in message  = news:1117122692.769994.314970@o13g2000cwo.googlegroups.com...  [snip]C > In AUTHORIZE I have set /FLAG=PWDMIX for my acccount, and I set a + > password that has a period in the middle.  > E > Now, if the DECwindows session lock activates (by clicking the lock I > button or due to inactivity) I can't unlock it! It tells me my password D > is invalid and generates a security audit. Yet I am able to log inI > elsewhere with exactly the same password. (The status show in the audit  > is %LOGIN-F-INVPWD.) [snip]  E FWIW, I believe this is a known problem; I don't believe the fix (in  K loginout.exe) is in any eco as yet, however.  Here's the ident info on the  I patched image I received for the problem I reported (OpenVMS Alpha 7.3-2)   > Image Name                     Image Version        Link date 	 Linker ID S ------------------------- -------------------- ---------------------  ------------- E LOGINOUT                   X-49_70-3-8294    2-MAR-2005 14:07  A11-50    Paul     ------------------------------  + Date: Thu, 26 May 2005 07:50:27 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> Y Subject: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors but  ? Message-ID: <d73v42$djt$1@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com>    Hi,   K Please can someone help me with this Global Symbol/Image Activation problem J I'm having when porting some code from VAX/VMS? See small example (75 line reproducer) below.   In a nutshell : -   J 1) I have Sharable Image 1 (Share1) that has declared a Global, OverlayingH Psect (Global_Var) and a Global Constant (My_Sym). Share1 was written in! Macro and initializes Global_Var. I 2) When Linking Share1 I declare both Global_Var and My_Sym in the symbol  table using the options fileK 3) I have a second Shareable Image (Share2) that links against (Share1) and L calls a stub. Share2 also maps to Global_Var and references My_Sym and calls the stub (My_Sub) in Share1 L 4) Finally, I have an executable image (Test1) that links against Share2 and also maps to Global_Var C 4a) If it tries to reference My_Sym as an EXTERNAL constant it gets  Undefined Symbol  I If you run TEST1 you will see that Global_Var in SHARE1 and SHARE2 map to H the same area but in TEST1 it contains rubbish. What am I doing wrong??? Please Note: This works on VAX.   B I've come across this problem before with NEW code and was able toI code/design around it, but this is existing code that would require MAJOR ( (even if doable) work to code around :-(  J Please help if you can. Pointers to relevant documentation would be great.  > Can you get VMS support from HP if you're on the DSPP program?   Regards Richard Maher    $! $ on warning then exit* $ define/nolog share1 sys$login:share1.exe* $ define/nolog share2 sys$login:share2.exe $! $ create share1.mar            .title  Share One  my_sym == 1 ?         .psect  global_var,long,pic,ovr,rel,gbl,noshr,noexe,wrt /                         .ascii          "Hello" )                         .blkb           3   =         .psect  my_code,pic,con,rel,lcl,shr,exe,rd,nowrt,long   $         .call_entry     label=my_sub         movzbl  #1,r0          ret            .end $! $ macro share1.mar> $ link/share=sys$login:share1.exe share1.obj,sys$input/options1 symbol_vector = (                               - 1                 my_sub             = procedure, - 1                 global_var         = psect,     - 1                 my_sym             = data       -                  )  $! $ create share2.cob  identification division. program-id.    share2. data division. working-storage section.C 01  my_sym                  pic 9(9)   comp  value external my_sym. 0 01  global_var              pic x(5)   external. procedure division.  00. 4     display "Share2 sym = ", my_sym with conversion.(     display "Share2 gbl = ", global_var.     call "my_sub".     exit program.  end program share2.  $! $ cobol/lis share2.cob: $ link/share=sys$login:share2.exe share2.obj,sys$input/opt sys$login:share1.exe/share psect_attr=global_var,noshr 1 symbol_vector = (                               - 1                 share2             = procedure, - 1                 global_var         = psect,     - 1                 my_sym             = data       -                  )  $! $ create test1.cob identification division. program-id.    test1.  data division. working-storage section.D *01  my_sym                  pic 9(9)   comp  value external my_sym.0 01  global_var              pic x(5)   external. procedure division.  00. 4 *    display "Test1 sym = ", my_sym with conversion.'     display "Test1 gbl = ", global_var.      call "share2".
     stop run.  end program test1. $ cobol/lis test1.cob  $ link test1.obj,sys$input/opt psect_attr=global_var,noshr  sys$login:share2.exe/share $!   ------------------------------    Date: 26 May 2005 03:57:06 -0700 From: dooleys@snowy.net.auY Subject: Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors  C Message-ID: <1117105026.434787.304940@g14g2000cwa.googlegroups.com>   0 link /map is your friend in situations like this Phil   ------------------------------  % Date: Thu, 26 May 2005 09:45:33 -0400 $ From: "Hein" <hein.nomail@hp.nomail>Y Subject: Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors  , Message-ID: <42951b8a$1@usenet01.boi.hp.com>  J I heard about problems in this space mostly for fortran which changed the A default common block attribute from SHR on vax to NOSHR on Alpha.   J Admittedly I did not read you note in the full detail it warrant, but one  thing struch me as odd:   % --->      psect_attr=global_var,noshr    Should that not simply be shr ?    Hein.     ? "Richard Maher" <maher_rj@hotspamnotmail.com> wrote in message  9 news:d73v42$djt$1@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com...  > Hi,  > F > Please can someone help me with this Global Symbol/Image Activation 	 > problem L > I'm having when porting some code from VAX/VMS? See small example (75 line > reproducer) below. >  > In a nutshell : -  > L > 1) I have Sharable Image 1 (Share1) that has declared a Global, OverlayingJ > Psect (Global_Var) and a Global Constant (My_Sym). Share1 was written in# > Macro and initializes Global_Var. K > 2) When Linking Share1 I declare both Global_Var and My_Sym in the symbol  > table using the options fileJ > 3) I have a second Shareable Image (Share2) that links against (Share1)  > and I > calls a stub. Share2 also maps to Global_Var and references My_Sym and   > calls  > the stub (My_Sub) in Share1 K > 4) Finally, I have an executable image (Test1) that links against Share2   > and  > also maps to Global_Var E > 4a) If it tries to reference My_Sym as an EXTERNAL constant it gets  > Undefined Symbol > K > If you run TEST1 you will see that Global_Var in SHARE1 and SHARE2 map to J > the same area but in TEST1 it contains rubbish. What am I doing wrong???! > Please Note: This works on VAX.  > D > I've come across this problem before with NEW code and was able toK > code/design around it, but this is existing code that would require MAJOR * > (even if doable) work to code around :-( > L > Please help if you can. Pointers to relevant documentation would be great. > @ > Can you get VMS support from HP if you're on the DSPP program? >  > Regards Richard Maher  >  > $! > $ on warning then exit, > $ define/nolog share1 sys$login:share1.exe, > $ define/nolog share2 sys$login:share2.exe > $! > $ create share1.mar  >  >        .title  Share One
 > my_sym == 1 @ >        .psect  global_var,long,pic,ovr,rel,gbl,noshr,noexe,wrt0 >                        .ascii          "Hello"* >                        .blkb           3 > > >        .psect  my_code,pic,con,rel,lcl,shr,exe,rd,nowrt,long > % >        .call_entry     label=my_sub  >        movzbl  #1,r0 >        ret > 
 >        .end  > $! > $ macro share1.mar@ > $ link/share=sys$login:share1.exe share1.obj,sys$input/options3 > symbol_vector = (                               - 2 >                my_sub             = procedure, -2 >                global_var         = psect,     -2 >                my_sym             = data       - >                ) > $! > $ create share2.cob  > identification division. > program-id.    share2. > data division. > working-storage section.E > 01  my_sym                  pic 9(9)   comp  value external my_sym. 2 > 01  global_var              pic x(5)   external. > procedure division.  > 00. 5 >    display "Share2 sym = ", my_sym with conversion. ) >    display "Share2 gbl = ", global_var.  >    call "my_sub".  >    exit program. > end program share2.  > $! > $ cobol/lis share2.cob< > $ link/share=sys$login:share2.exe share2.obj,sys$input/opt > sys$login:share1.exe/share > psect_attr=global_var,noshr 3 > symbol_vector = (                               - 2 >                share2             = procedure, -2 >                global_var         = psect,     -2 >                my_sym             = data       - >                ) > $! > $ create test1.cob > identification division. > program-id.    test1.  > data division. > working-storage section.F > *01  my_sym                  pic 9(9)   comp  value external my_sym.2 > 01  global_var              pic x(5)   external. > procedure division.  > 00. 6 > *    display "Test1 sym = ", my_sym with conversion.( >    display "Test1 gbl = ", global_var. >    call "share2".  >    stop run. > end program test1. > $ cobol/lis test1.cob   > $ link test1.obj,sys$input/opt > psect_attr=global_var,noshr  > sys$login:share2.exe/share > $! >  >    ------------------------------  + Date: Thu, 26 May 2005 17:09:03 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> Y Subject: Re: HELP! Serious Alpha Image Activation problem. Symbol Resolution - No errors  ? Message-ID: <d74vre$cct$1@nwrdmz02.dmz.ncs.ea.ibs-infra.bt.com>    Et tu Hein?   L PLEASE shoot me down in flames on this, but dosen't SHR just create a Global" Section/ Process Shared data area?  5 Regards Richard "Can anyone do this on Itanium" Maher   L PS. YES IT IS MY PROBLEM! But 75 lines of reproducer just may be able to pin! it down a bit??? Whaddaya reakon?   / "Hein" <hein.nomail@hp.nomail> wrote in message & news:42951b8a$1@usenet01.boi.hp.com...K > I heard about problems in this space mostly for fortran which changed the C > default common block attribute from SHR on vax to NOSHR on Alpha.  > K > Admittedly I did not read you note in the full detail it warrant, but one  > thing struch me as odd:  > ' > --->      psect_attr=global_var,noshr  > ! > Should that not simply be shr ?  >  > Hein.  >  > @ > "Richard Maher" <maher_rj@hotspamnotmail.com> wrote in message; > news:d73v42$djt$1@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com...  > > Hi,  > > G > > Please can someone help me with this Global Symbol/Image Activation  > > problem I > > I'm having when porting some code from VAX/VMS? See small example (75  line > > reproducer) below. > >  > > In a nutshell : -  > > C > > 1) I have Sharable Image 1 (Share1) that has declared a Global, 
 OverlayingL > > Psect (Global_Var) and a Global Constant (My_Sym). Share1 was written in% > > Macro and initializes Global_Var. F > > 2) When Linking Share1 I declare both Global_Var and My_Sym in the symbol  > > table using the options fileK > > 3) I have a second Shareable Image (Share2) that links against (Share1)  > > and J > > calls a stub. Share2 also maps to Global_Var and references My_Sym and	 > > calls  > > the stub (My_Sub) in Share1 L > > 4) Finally, I have an executable image (Test1) that links against Share2 > > and  > > also maps to Global_Var G > > 4a) If it tries to reference My_Sym as an EXTERNAL constant it gets  > > Undefined Symbol > > J > > If you run TEST1 you will see that Global_Var in SHARE1 and SHARE2 map toL > > the same area but in TEST1 it contains rubbish. What am I doing wrong???# > > Please Note: This works on VAX.  > > F > > I've come across this problem before with NEW code and was able toG > > code/design around it, but this is existing code that would require  MAJOR , > > (even if doable) work to code around :-( > > G > > Please help if you can. Pointers to relevant documentation would be  great. > > B > > Can you get VMS support from HP if you're on the DSPP program? > >  > > Regards Richard Maher  > >  > > $! > > $ on warning then exit. > > $ define/nolog share1 sys$login:share1.exe. > > $ define/nolog share2 sys$login:share2.exe > > $! > > $ create share1.mar  > >  > >        .title  Share One > > my_sym == 1 B > >        .psect  global_var,long,pic,ovr,rel,gbl,noshr,noexe,wrt2 > >                        .ascii          "Hello", > >                        .blkb           3 > > @ > >        .psect  my_code,pic,con,rel,lcl,shr,exe,rd,nowrt,long > > ' > >        .call_entry     label=my_sub  > >        movzbl  #1,r0 > >        ret > >  > >        .end  > > $! > > $ macro share1.marB > > $ link/share=sys$login:share1.exe share1.obj,sys$input/options5 > > symbol_vector = (                               - 4 > >                my_sub             = procedure, -4 > >                global_var         = psect,     -4 > >                my_sym             = data       - > >                ) > > $! > > $ create share2.cob  > > identification division. > > program-id.    share2. > > data division. > > working-storage section.G > > 01  my_sym                  pic 9(9)   comp  value external my_sym. 4 > > 01  global_var              pic x(5)   external. > > procedure division.  > > 00. 7 > >    display "Share2 sym = ", my_sym with conversion. + > >    display "Share2 gbl = ", global_var.  > >    call "my_sub".  > >    exit program. > > end program share2.  > > $! > > $ cobol/lis share2.cob> > > $ link/share=sys$login:share2.exe share2.obj,sys$input/opt > > sys$login:share1.exe/share > > psect_attr=global_var,noshr 5 > > symbol_vector = (                               - 4 > >                share2             = procedure, -4 > >                global_var         = psect,     -4 > >                my_sym             = data       - > >                ) > > $! > > $ create test1.cob > > identification division. > > program-id.    test1.  > > data division. > > working-storage section.H > > *01  my_sym                  pic 9(9)   comp  value external my_sym.4 > > 01  global_var              pic x(5)   external. > > procedure division.  > > 00. 8 > > *    display "Test1 sym = ", my_sym with conversion.* > >    display "Test1 gbl = ", global_var. > >    call "share2".  > >    stop run. > > end program test1. > > $ cobol/lis test1.cob " > > $ link test1.obj,sys$input/opt > > psect_attr=global_var,noshr  > > sys$login:share2.exe/share > > $! > >  > >  >  >    ------------------------------  % Date: Thu, 26 May 2005 04:14:20 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> % Subject: HP PaRisc, IA64 and NSK news , Message-ID: <42958535.35BE7307@teksavvy.com>   Just a heads up.  E This or next week, HP is expected to unveil its new PaRisc chip which . will be available in servers/workstations etc.  E Also this or next week HP is supposed to unveil Tandem's NSK on IA64. H Supposedly the servers will use the "integrity" name but with a Non-Stop tagged on to it.  F Of the press I have read yesterday, articles mentioned the positive ofG HP simplifying its server products into 2 architectures, while doubting 5 the future of IA64 due to its low volume/small niche.     G Shouldn't we push HP to continue to sell Alphas for as long as it sells  PaRisc systems ?   ------------------------------    Date: 26 May 2005 01:30:25 -0700 From: icerq4a@spray.se= Subject: Re: HP Pulls Off a Respectable Second Fiscal Quarter C Message-ID: <1117096225.052076.239430@z14g2000cwz.googlegroups.com>    JF Mezei wrote:  > icerq4a@spray.se wrote: K > > Considering that HP is in a transition phase, I am surprised they do as J > > well as they do. IBM is not in a transition phase, so I expect them to< > > have better figures, anything else would be embarrasing. > 0 > Transition phase ? Perhaps Transition Decade ? > H > HP has been transitioning HP-UX for how long now ? 10 years ?  So, theH > IA64 came in very late in 2001 and was unusable, but HP customers haveD > known since the mid 1990s that PaRisc was dead. And they are still! > buying it over that IA64 thing.   F Real product usable for transition began to be available in 2002-2003.D I think it is first in 2005 that the Integrity line is or will be onF pair with 9000 on features, so lots of customers have had no reason toD switch yet. Their current stuff works just fine. Migrations have notC even yet begun. 2006-2008 is when PA-RISC users will begin to move. @ Most current Integrity sales are new systems and not systems for replacement.   ------------------------------  % Date: Thu, 26 May 2005 04:58:45 -0400 ( From: Bill Todd <billtodd@metrocast.net>= Subject: Re: HP Pulls Off a Respectable Second Fiscal Quarter = Message-ID: <qY-dncmtjb5bEgjfRVn-3Q@metrocastcablevision.com>    icerq4a@spray.se wrote:    ...   B > Most current Integrity sales are new systems and not systems for > replacement.  F Do you have an actual reference for that, or is it a 'fact' which you 0 just decided to pull out of some bodily orifice?  D For example, it would not take a *huge* migrating percentage of the D Tru64 base to have accounted for a sizable percentage of BCS Itanic G sales (given the modest magnitude of the latter) - and unlike the case  I with HP-UX (or even VMS), Tru64 customers who have decided to migrate at  G all have no reason to delay their move to HP-UX (however much they may  G regret its necessity) but rather considerable incentive to start using  G it for replacement systems as early as possible to see what they're up   against.   - bill   ------------------------------  % Date: Thu, 26 May 2005 17:50:54 +0200  From: unixSPAM@zeouane.orgY Subject: Re: I want to ask you the most important question of your life. The question is: : Message-ID: <1gx5zma.1kl6mqj1y5oistN%unixSPAM@zeouane.org>  E Shmuel (Seymour J.) Metz <spamtrap@library.lspace.org.invalid> wrote:   A > In <1gx2stc.12tka6x2nbc7rN%unixSPAM@zeouane.org>, on 05/24/2005 , >    at 05:06 PM, unixSPAM@zeouane.org said: > H > >Anyone know how to stop this lunatic ?  All methods considered, up to! > >and including use of firearms.  > D > Four horses. Four ropes. Four directions. Watch out for complaints > from the ASPCA.  > ' > Meanwhile complain to AOL and google.    As if AOL gives a s**t ...     --   unix@zeouane.org   ------------------------------    Date: 26 May 2005 07:56:50 -0500; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) ! Subject: Re: Installing a printer 3 Message-ID: <flRfBIs4AXBs@eisner.encompasserve.org>   ^ In article <Xns9661DF860172Fdcovmsrox@212.100.160.126>, "Doc." <doc@openvms-rocks.com> writes: > M > Personally I pity the people that have never worked on code printed on 132-  > column music-score paper.   E    We still have an LG08 that I bought when the boss said people kept F    telling here to get a laser printer and that's not what she wanted.D    It's so quiet I had a hell of a time telling when I got the queue.    and the reverse LAT connection setup right.  D    Its used mostly to help maintain software that was written in theH    80's on a processor from the 70's that needs to run until about 2010.@    (And its on a spacecraft so it's not going to be "upgraded".)G    This is done using tools written in the 70's in RATFOR on some kind  G    of IBM mainframe, the generated Fortran-IV for which was then ported @    to a VAX 8650, then a DEC 10000, and now runs on a little EV5    workstation.   J    The last time I ordered paper for the LG08 I heard it cost $600 a box, .    but we still have a good supply of ribbons.   ------------------------------   Date: 26 May 2005 13:13:26 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)! Subject: Re: Installing a printer + Message-ID: <3fm0bmF8gbinU1@individual.net>   3 In article <flRfBIs4AXBs@eisner.encompasserve.org>, > 	koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > F >    Its used mostly to help maintain software that was written in theJ >    80's on a processor from the 70's that needs to run until about 2010.B >    (And its on a spacecraft so it's not going to be "upgraded".)I >    This is done using tools written in the 70's in RATFOR on some kind PI >    of IBM mainframe, the generated Fortran-IV for which was then portedoB >    to a VAX 8650, then a DEC 10000, and now runs on a little EV5 >    workstation.e  G That's not the box that NASA needed a PDP-11 to read the telemetry from  a few uears back, is it?  :-)    bill   -- iJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   u   ------------------------------  % Date: Thu, 26 May 2005 11:31:35 -0400r- From: William Webb <william.w.webb@gmail.com> ! Subject: Re: Installing a printer:7 Message-ID: <8660a3a10505260831180bb4f8@mail.gmail.com>?  @ On 25 May 2005 19:57:09 GMT, Doc. <doc@openvms-rocks.com> wrote:& > %NEWS-I-NEWMSG, Dave Froble wrote in) > news:1199g6gfq8kfgd8@corp.supernews.comw >=20 > > MJS wrote: >=20H > >> I've been wondering about that, read the DCPS System Manager manualH > >> today but it only referred to Network or Serial printing. Can't say0 > >> I've ever heard of a serial printer though? > >u" > > That makes me feel really old. >=20L > Personally I pity the people that have never worked on code printed on 13= 2- > column music-score paper.h >=20& > Real listings have coffee rings. :-) >=20 >=20 > Doc. > --I > OpenVMS:     Eight out of ten hackers prefer *other* operating systems.SI > http://www.openvms-rocks.com    Deathrow Public-Access OpenVMS Cluster.- >=20  6 Real line printers have bands and rows of hammers. =20  L They can be used as musical instruments if you know what to print, and when= .w   WWWebb   --=20eC NOTE: This email address is only used for noncommerical VMS-relatedL correspondence.gC All unsolicited commercial email will be deemed to be a request fors8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/-   ------------------------------  % Date: Thu, 26 May 2005 04:38:27 -0400j- From: JF Mezei <jfmezei.spamnot@teksavvy.com>  Subject: SEARCH : AND NOT ?.+ Message-ID: <42958ADC.FEDF127@teksavvy.com>    Is there a way to do:O  1 SEARCH RECIPES.TXT  CAKE,VANILLA /MATCH=AND_NOT ?k  I In other words, I want all records that contain "CAKE" but NOT "VANILLA" a  E I know I can probably use a pipe to get all records that contain CAKEtH piked into a search that excludes anything with VANILLA. But it would be/ nice to be able to do this in a single command.f   Is there a trick ?    D (ok: consider a log file where you want to extract all records for aG certain date/hour, and want to exclude records with a certain text that*E you don't care to look at becase they are so numerous and normal, butR9 wish to focus on any record that may show some anomalies)    ------------------------------  % Date: Thu, 26 May 2005 10:57:30 +0200k& From: Paul Sture <paul.sture@decus.ch> Subject: Re: SEARCH : AND NOT ?H+ Message-ID: <3flhb5F8bi5mU1@individual.net>n   JF Mezei wrote:-   > Is there a way to do:  > 3 > SEARCH RECIPES.TXT  CAKE,VANILLA /MATCH=AND_NOT ?r > K > In other words, I want all records that contain "CAKE" but NOT "VANILLA" 5 > G > I know I can probably use a pipe to get all records that contain CAKE'J > piked into a search that excludes anything with VANILLA. But it would be1 > nice to be able to do this in a single command.n >  > Is there a trick ? >  > F > (ok: consider a log file where you want to extract all records for aI > certain date/hour, and want to exclude records with a certain text thatcG > you don't care to look at becase they are so numerous and normal, butc; > wish to focus on any record that may show some anomalies)e  8 I don't see an easy single step here, but something like  / $ search 'logfile' 'date_hour' /out=outfile.txta5 $ search outfile.txt 'date_hour','vanilla' /match=xork   should work8   ------------------------------  # Date: Thu, 26 May 2005 09:46:05 GMT0L From: winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) Subject: Re: SEARCH : AND NOT ?g6 Message-ID: <00A44525.86526B88@SSRL.SLAC.STANFORD.EDU>  [ In article <42958ADC.FEDF127@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:t >Is there a way to do: >B2 >SEARCH RECIPES.TXT  CAKE,VANILLA /MATCH=AND_NOT ? >tJ >In other words, I want all records that contain "CAKE" but NOT "VANILLA"  > F >I know I can probably use a pipe to get all records that contain CAKEI >piked into a search that excludes anything with VANILLA. But it would bes0 >nice to be able to do this in a single command. >i >Is there a trick ?> >A >EE >(ok: consider a log file where you want to extract all records for a H >certain date/hour, and want to exclude records with a certain text thatF >you don't care to look at becase they are so numerous and normal, but: >wish to focus on any record that may show some anomalies)  N I just needed to do this the other day, and I couldn't figure out a way to getN SEARCH to do  it in one image invocation. /MATCH=NAND absolutely didn't do it.  : (What I wanted was all the processes named STRING but not 1 STRING_WITH_MORE_STUFF. What I ended up doing waso  K $ PIP SHOW SYSTEM | SEA SYS$INPUT STRING_WITH_MORE_STUFF/match=NAND | sea -y%                     sys$input STRING o  O (I don't send the output anywhere because I'm really only interested in whether N I found anything, which I can get from the $STATUS.  But you could do /OUTPUT=L or have another entry in the pipe that deals with the records you found. TheM first SEA throws out all the records I know in advance I'm not interested in,eJ then the second SEA looks for my shorter string in the remaining records.)   -- Alant   ------------------------------  % Date: Thu, 26 May 2005 10:57:58 +0100 * From: "Richard Brodie" <R.Brodie@rl.ac.uk> Subject: Re: SEARCH : AND NOT ?s2 Message-ID: <d746j7$rtj$1@blackmamba.itd.rl.ac.uk>  Y "Alan Winston - SSRL Central Computing" <winston@SSRL.SLAC.STANFORD.EDU> wrote in messagem0 news:00A44525.86526B88@SSRL.SLAC.STANFORD.EDU...  ; > (What I wanted was all the processes named STRING but note > STRING_WITH_MORE_STUFF.h  2 In that case, a straight XOR search will work, no?   ------------------------------  % Date: Thu, 26 May 2005 12:07:54 +0200/& From: Paul Sture <paul.sture@decus.ch> Subject: Re: SEARCH : AND NOT ?p+ Message-ID: <3fllf5F8e77tU1@individual.net>f   Richard Brodie wrote:   [ > "Alan Winston - SSRL Central Computing" <winston@SSRL.SLAC.STANFORD.EDU> wrote in messageb2 > news:00A44525.86526B88@SSRL.SLAC.STANFORD.EDU... >  > ; >>(What I wanted was all the processes named STRING but notc >>STRING_WITH_MORE_STUFF.e >  > 4 > In that case, a straight XOR search will work, no? >  >  Example:   $ type x.txt sometext
 sometext plusd
 sometext pluss sometext+ $ sea x.txt sometext, plus/match=xor/numberr       1  sometext        4  sometexts   ------------------------------  % Date: Thu, 26 May 2005 12:09:25 +0200o& From: Paul Sture <paul.sture@decus.ch> Subject: Re: SEARCH : AND NOT ? + Message-ID: <3fllhvF8e77tU2@individual.net>s   Paul Sture wrote:h   > Richard Brodie wrote:  > + >> "Alan Winston - SSRL Central Computing" r4 >> <winston@SSRL.SLAC.STANFORD.EDU> wrote in message3 >> news:00A44525.86526B88@SSRL.SLAC.STANFORD.EDU...w >> >>= >>> (What I wanted was all the processes named STRING but not  >>> STRING_WITH_MORE_STUFF.m >> >> >>5 >> In that case, a straight XOR search will work, no?R >> >>
 > Example: >  > $ type x.txt
 > sometext > sometext pluse > sometext plusi
 > sometext- > $ sea x.txt sometext, plus/match=xor/numbere >      1  sometext >      4  sometext  K Oops. I didn't think of the case where "WITH_MORE_STUFF" might be variable.n   ------------------------------  # Date: Thu, 26 May 2005 10:10:18 GMTiL From: winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) Subject: Re: SEARCH : AND NOT ? 6 Message-ID: <00A44528.E8CC85B4@SSRL.SLAC.STANFORD.EDU>  _ In article <d746j7$rtj$1@blackmamba.itd.rl.ac.uk>, "Richard Brodie" <R.Brodie@rl.ac.uk> writes:  >lZ >"Alan Winston - SSRL Central Computing" <winston@SSRL.SLAC.STANFORD.EDU> wrote in message1 >news:00A44525.86526B88@SSRL.SLAC.STANFORD.EDU...a >e< >> (What I wanted was all the processes named STRING but not >> STRING_WITH_MORE_STUFF. >s3 >In that case, a straight XOR search will work, no?   C D'oh!  You're right there, for my special case.  (Wouldn't work forhK JF's case because he'd find records with "VANILLA" or "CHOCOLATE", and whatrG he wants is "CHOCOLATE" without "VANILLA".) I've now changed my processoI (which runs once a minute and checks on whether another process is still aM around) to use the /MATCH=XOR; you've saved me 1440 subprocess creations and 5! image activations a day.  Thanks!@   -- Alan    ------------------------------  % Date: Thu, 26 May 2005 13:07:08 -0400z$ From: "Hein" <hein.nomail@hp.nomail> Subject: Re: SEARCH : AND NOT ?t* Message-ID: <42954ac8@usenet01.boi.hp.com>  ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message 0% news:42958ADC.FEDF127@teksavvy.com...a > Is there a way to do:  > 3 > SEARCH RECIPES.TXT  CAKE,VANILLA /MATCH=AND_NOT ?h >rJ > In other words, I want all records that contain "CAKE" but NOT "VANILLA" > G > I know I can probably use a pipe to get all records that contain CAKE J > piked into a search that excludes anything with VANILLA. But it would be1 > nice to be able to do this in a single command.h >d > Is there a trick ?   Yeah, use perl:e  5 $ perl -ne "print if /cake/ && ! /vanilla/" < tmp.tmpc  " but the VMS solution is of course:  = $ pipe search tmp.tmp cake | searc/match=nor sys$pipe vanillag    F > (ok: consider a log file where you want to extract all records for aI > certain date/hour, and want to exclude records with a certain text thatmG > you don't care to look at becase they are so numerous and normal, but'; > wish to focus on any record that may show some anomalies)a  M Unlike SEARCH, a perl solution would gladly search for specific time windows S  beyond a crude "search ... "10:"J And it woudl make it much easier ot build in further heuristics, allowing . you to focus more readily on critical records.7 A starting point for a more complete solution could be:r  A $ type search.p  !-----------------------------------------------s $file = shift @ARGV; $with = shift @ARGV;H $without = shift @ARGV or die "Please provide all arguments: <filename>  <match> <not-match>  "; D open (FILE,"<$file") or die "Failed to open \"$file\" for input.  "; while (<FILE>) {
   $records++; "   if (/$with/i && ! /$without/i) {       $match++;g       print;       }m   } 5 print "$records records searched, $match matched.\n";SF $!--------------------------------------------------------------------$ $ perl search.p tmp.tmp cake vanilla cake	 noot cake  7 records searched, 2 matched. $i   fwiw,d Hein.    ------------------------------    Date: 26 May 2005 03:49:37 -0700 From: dooleys@snowy.net.au Subject: Re: TCPIP$TELNETSYMB Message-ID: <1117104577.073325.51260@f14g2000cwb.googlegroups.com>   phantasmus wrote:h > I tried this logicals + >   "TCPIP$TELNETSYM_RAW_TCP" [super] = "1"%* >   "TCPIP$TELNETSYM_RAW_TCP" [exec] = "1"7 >   "TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS" [super] = "17"R6 >   "TCPIP$TELNETSYM_SUPPRESS_FORMFEEDS" [exec] = "17" > with no success ... ;-((0 and ran the relevant shutdown/startup procedure? Phil   ------------------------------  % Date: Thu, 26 May 2005 09:07:45 +0200O- From: "Martin Vorlaender" <mv@pdv-systeme.de> I Subject: Re: Third party Windows App to NFS map OpenVMS to Windows Servera+ Message-ID: <3flatsF897prU1@individual.net>s  7 "Peter Weaver" <newsgroup@weaverconsulting.ca> wrote...  > Keith Cayemberg wrote: > > ...aJ > > [TCPIP] How to Map NFS Exported OpenVMS File Systems from Microsoft SF > >  >sS http://h18000.www1.hp.com/support/asktima/communications/CHAMP_SRC031029001146.html  > > ...P > E > $ SET MODE/EMULATE="Richard Maher on FMS" ! People may want to stopX > reading now.   :-)7  ! > Oh, how I HATE HP's web site!!!    Only the business part...m  B > Go to that article and it tells you to go to two other articles;C >   "[TCP/IP] How To Mount TCPIP Exported File Systems Using PCNFS" 5 >   "[TCP/IP] How to Setup an NFS Server Using TCPIP"  > E > But neither of these are links, they are just plain text. Go to thei > "Search" box  F Wrong move (note the button "business support center" below the searchG text field). You should use the search page that links from the OpenVMSlF home page (titled "Natural language search assistant"). From there you! find the articles pretty quickly.W   cu,e   Martin --  B  Your mouse has moved.      | Martin Vorlaender  |  OpenVMS rules!5  Windows must be restarted  | work: mv@pdv-systeme.de H  for the change to take     |   http://www.pdv-systeme.de/users/martinv/<  effect. Reboot now? [OK]   | home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 26 May 2005 14:30:11 +0200-0 From: Keith Cayemberg <keith.cayemberg@arcor.de>I Subject: Re: Third party Windows App to NFS map OpenVMS to Windows Server-B Message-ID: <4295c153$0$14743$9b4e6d93@newsread4.arcor-online.net>   Peter Weaver wrote:h > Keith Cayemberg wrote: >  >>...HH >>[TCPIP] How to Map NFS Exported OpenVMS File Systems from Microsoft SF >> > U > http://h18000.www1.hp.com/support/asktima/communications/CHAMP_SRC031029001146.htmls >  >>...  >  > E > $ SET MODE/EMULATE="Richard Maher on FMS" ! People may want to stopo > reading now.   Oh Oh!!  :-)   > ! > Oh, how I HATE HP's web site!!!n > B > Go to that article and it tells you to go to two other articles;I >         "[TCP/IP] How To Mount TCPIP Exported File Systems Using PCNFS"S; >         "[TCP/IP] How to Setup an NFS Server Using TCPIP"W > E > But neither of these are links, they are just plain text. Go to thelF > "Search" box and paste in "[TCP/IP] How to Setup an NFS Server UsingA > TCPIP" and click >>. Nothing happens because you need to enablelJ > JavaScript. Reload the page, allow JavaScript, paste again and click >>.G > This time you get a page saying that it can not find any reference to I > "[TCP/IP] How to Setup an NFS Server Using TCPIP" so you try again withLF > out the quotes. This time you get a page saying that it found 43,306I > entries and that it will show you the first 500 at 10 entries per page.oH > So you walk through the first 60 entries looking for an article titledJ > "How to Setup an NFS Server Using TCPIP" and give up! CAN SOMEONE PLEASE: > FIX THE SEARCH FEATURE AT HP SO IT WORK!!!!!!!!!!!!!!!!! >  > $ SET MODE/NOEMULATE >   9 Sorry about that I was in a bit of a hurry when I posted.e  = [TCP/IP] How To Mount TCPIP Exported File Systems Using PCNFSrS http://h18000.www1.hp.com/support/asktima/communications/CHAMP_SRC960729007496.htmle  / [TCP/IP] How to Setup an NFS Server Using TCPIP S http://h18000.www1.hp.com/support/asktima/communications/CHAMP_SRC960729007455.htmle  @ There are also other NFS related documents to be found using the* HP natural language search assistant at... http://h71029.www7.hp.com/    H By the way, one further remark to Jim Strehlow about something which he E probably already knows, but is so important for his application area s that I need to say it anyways.  I If you are using NFS for concurrent access to a database, index files or eB other object with shared internal structures, the existence of an D emulated file system on a network device may or may not provide the G necessary compatible locking mechanisms needed for coordinating access sD to these internal structures. This may depend on the quality of the C combination of NFS tools used on both sides, or it may not be even oC possible if one side doesn't provide a suitable or compatible lock c6 service that can be used by the file system emulation.  G Some other middleware product could provide this internal (potentially  H distributed) locking which may be included as part of the OS as is with G OpenVMS Clustering, or as an option of the database to be installed on iG both systems as with Oracle RAC, or as an add-on from a third party as   may be provided by Veritas.s  I The most common method of sharing DB data with other other systems is by  ? allowing only one system to have direct access to the database  H on-disk-structure, and using this one system to serve requests for data I to the other systems over an API (such as ODBC or JDBC), DB link service nB or TP Monitor. In this case NFS would no longer be needed and may H actually cause problems by tricking a system into believing the file is F on a local physical device and thus owned and coordinated only by the  local native locking mechanism.y  D As an example, I had a recent customer that had mysterious problems G accessing an Oracle Rdb database on an OpenVMS system from a Sybase DB tH using OmniConnect(ODBC) on a Windows system. The various bugcheck dumps H had many different messages, which appeared as if the connection to the E database was being interrupted at random moments asynchronous to the SG operation of the database itself. A long intensive search for problems eE in the lower level network and middleware services proved fruitless. i@ Even more mysterious was that the same code and product version ? combinations were working perfectly at another factory. Then a rG difference was discovered between the two sites, along with an unusual e+ usage of proxies, that explained the cause.r  H The site which was working correctly had configured their Rdb DB system H as a member of a larger OpenVMS cluster, and the site with the problems A had their Rdb DB on a stand-alone system separate from the local dI cluster. That alone would have been no problem when OmniConnect had been  G used to do the obvious direct connect to the Rdb system in both sites. pB However, it turned out that the programmer of the software on the F windows system had no account on the Rdb DB Systems at the two sites, I but did have an account on the cluster systems at the two sites. Knowing bH that there was a default DECnet Proxy for the cluster account to an Rdb G account on the Rdb systems, that person simply configured the DSN with tD the IP address for the cluster system, and an attach statement that C included the node name of Rdb DB system. Rdb and SQL/Services were yI installed on all systems. Due to the DECnet Proxy, to the cluster system eG the Rdb database appeared to be local and dutifully allowed access and aB locking of it's structures. But the database was also open on the I standalone system and being used by process reserving tables and records sA on the same database. Rdb uses the OpenVMS DLM (distributed lock yH manager) to coordinate access to it's internal DB structures, but since C the two systems were not in the same cluster, there was no cluster  E interconnect over which the two DLMs could inform or coordinated the  I locks between the two systems. So occasionally the locks and assumptions HD of exclusivity in the internal DB structures were being ignored and I collisions occurred. Considering that the access was through a user that oI had all privileges turned on, it was surprising that the database wasn't iI corrupted. I consider it a tribute to the correctness of Rdb transaction w0 model that the database survived this interlude.  C Incorrect Sybase Direct Connect Configuration (greatly simplified):e       WINDOWSr6 ┌───────────┐DC(ODBC)/tcp/ip& │  Sybase   ├──────┐ │ODBC Client│      │5 └───────────┘      │  VMSo7                     V  Cluster          VMS Stand-alone              ┌────────────────┐         ┌──────────────┐  serviceJ             │  SQL/Services  │         │ SQL/Services │<-bypassed!{             ├────────────────┤         ├──────────────┤ J             │    Rdb SQL     │         │   Rdb SQL    │<-bypassed!{             ├────────────────┤         ├──────────────┤ J             │ Rdb DB Engine  │   CI    │Rdb DB Engine │<-bypassed!{             ├────────────────┤ missing!├──────────────┤vX             │      DLM       ├───//────┤     DLM      │<-bypassed!{             ├────────────────┤         ├──────────────┤ I             │VMS I/O Services│   ┌────>│ I/O Services │ }             └────────────────┘   │     ├──────────────┤iW                     └────────────┘     │ Rdb DB File  │hW                  DECnet Network Proxy  ├──────────────┤ ;                                        │  HD Device   │-W                                        └──────────────┘  DC = Direct Connect  DLM = Distributed Lock Manager HD = Hard Disk= Note: Sybase on VMS doesn't use DLM, and consequently is not . Cluster-capable.    I The above NFS concerns are some that I would have when attempting to use yH a database from multiple systems. They are however theoretical for your I case since I have no personal experience attempting concurrent access to dG a DB made available on a system using NFS, and the database you select  % will also determine what is possible.t   cheers!    Keith Cayembergi   ------------------------------   End of INFO-VAX 2005.292 ************************