1 INFO-VAX	Tue, 20 Dec 2005	Volume 2005 : Issue 706       Contents: Re: Another Backup suggestion " Re: calloc() fails, no more memory" Re: calloc() fails, no more memory" Re: calloc() fails, no more memory Re: Database access from COBOL RE: Database access from COBOL Re: Database access from COBOL Re: Database access from COBOL# Re: DECnet IV over Multinet Circuit ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted) ) Re: Deleting alias files (blocks deleted)  format bug in vmsperl  format bug in vmsperl  Re: format bug in vmsperl  Licenses MOD_UNITS question.  Re: Licenses MOD_UNITS question.  Re: Licenses MOD_UNITS question.  Re: Licenses MOD_UNITS question.  Re: Licenses MOD_UNITS question.  Re: Licenses MOD_UNITS question.& Re: Moving directories of files to dos& Re: Moving directories of files to dos& Re: Moving directories of files to dosP Re: OT - GETSYI (was:Re: PHONE error - Invalid specification of node or person. 2 OT: Humor: Best Windows Fix story I've ever heard.E Re: PHONE error - Invalid specification of node or person. Try again. E Re: PHONE error - Invalid specification of node or person. Try again. E Re: PHONE error - Invalid specification of node or person. Try again.  Re: Rules for joining a cluster  Update on my migration... :-( ! Re: Update on my migration... :-(   F ----------------------------------------------------------------------  + Date: Mon, 19 Dec 2005 22:29:16 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)& Subject: Re: Another Backup suggestion$ Message-ID: <do7c7p$927$3@online.de>  C In article <1135009159.323397.127380@g14g2000cwa.googlegroups.com>, ' "AEF" <spamsink2001@yahoo.com> writes:    H > Just to be more clear, and to second this request, SET PREFIX will notF > show you when the verification and backup passes start, or any other" > BACKUP messages for that matter.  E Yes, of course; I read too quickly and was thinking of a sequence of  B backup commands and wanting to know how much time each one needed.   ------------------------------  % Date: Mon, 19 Dec 2005 10:47:05 -0800  From: Z <Z@no.spam> + Subject: Re: calloc() fails, no more memory ) Message-ID: <GeDpf.4133$Vr.3200@fe06.lga>    Hans Blom wrote:G > I tried a loop calloc:ing 400000 pointers (1.6 Mb approx). The 383:rd : > attempt fails. errno = %x0c, vaxc$errno = 0. I think theE > programmer/developer should do a review of his program,  the calloc A > failing can't be the main problem, and then he is welcome back.   & Isn't errno = %0C an access violation?  ? Why would you get an accvio from running out of virtual memory?    ------------------------------  + Date: Mon, 19 Dec 2005 13:43:04 -0600 (CST) * From: sms@antinode.org (Steven M. Schweda)+ Subject: Re: calloc() fails, no more memory 2 Message-ID: <05121913430435_202488EC@antinode.org>   From: Z <Z@no.spam>    > Hans Blom wrote:I > > I tried a loop calloc:ing 400000 pointers (1.6 Mb approx). The 383:rd < > > attempt fails. errno = %x0c, vaxc$errno = 0. I think theG > > programmer/developer should do a review of his program,  the calloc C > > failing can't be the main problem, and then he is welcome back.  > ( > Isn't errno = %0C an access violation?  A    Don't do much work in C, do you?  "%x0c", often referred to as 7 "12" (or "twelve"), is entirely appropriate, as seen in 3 SYS$COMMON:[DECC$LIB.REFERENCE.DECC$RTLDEF]ERRNO.H:   J #define ENOMEM          12     /* Not enough core                       */  A > Why would you get an accvio from running out of virtual memory?   H    No reason at all (unless you actually tried to use the resulting null	 pointer).   H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode-org     Saint Paul  MN  55105-2547    ------------------------------    Date: 19 Dec 2005 15:56:30 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) + Subject: Re: calloc() fails, no more memory 3 Message-ID: <kvboEKbkCsC2@eisner.encompasserve.org>   [ In article <j617od.m1d.ln@news.hus-soft.de>, Albrecht Schlosser <ajs856@tiscali.de> writes:  >>  ) >> Isn't errno = %0C an access violation? B >> Why would you get an accvio from running out of virtual memory? >  > $ errno 12$ > Error number  12 : not enough core >   > "$ errno" does something like: >   I    SYSTEM-F-ACCVIO is a VMS message who's corresponding number is C (12). C    errno contains C RTL message numbers, not VMS message numbers.   E    VMS, C, Fortran, ... all have their own message numbering schemes.   D    errno 12 is "not enough core", which is not an access violation. A    More likely translated from a VMS error which had more detail.   G    perror() on VMS will tell you what the VMS error was, if errno comes     out as EVMSERR.   ------------------------------  % Date: Mon, 19 Dec 2005 13:37:39 -0500 ' From: Dave Froble <davef@tsoft-inc.com> ' Subject: Re: Database access from COBOL 0 Message-ID: <11qdvc66hh3jh92@corp.supernews.com>   Bill Gunshannon wrote:2 > In article <11qbp869ffr60fc@corp.supernews.com>,, > 	Dave Froble <davef@tsoft-inc.com> writes: > ! >>johnhreinhardt@yahoo.com wrote:  >> >>>Bob Koehler wrote:  >>>  >>>  >>> A >>>>  Use of Hobbyist license for educational institutions is not D >>>>  permitted, unless there have been changes since Ilast read it. >>>>= >>>>  Educational institutions are offered discount programs.  >>>  >>> D >>>That's still correct, however there is a free Educational License
 >>>Program >>> 2 >>>http://h71000.www7.hp.com/openvmsedu/index.html >>>  >>>Copied from the first page: >>> E >>>"The OpenVMS Educational License Program is patterned after the HP G >>>OpenVMS Hobbyist Program. It is offered in response to your requests F >>>that we make HP OpenVMS available to educational institutions at no >>>cost. >>> H >>>All types of educational institutions worldwide are eligible. All youC >>>have to do is request an access code over the Web, register your H >>>school, and once registered, anyone at your school can use the accessI >>>code to get a product authorization key (PAK) through e-mail. The base C >>>license for HP OpenVMS is included, and you may request over 100 H >>>layered product PAKs. The licenses are good from August to August andB >>>must be renewed annually. Anyone associated with an educational@ >>>institution-students, faculty or administrators-may use these9 >>>licenses. The pages listed below will get you started.  >>F >>I don't know what could be much better.  It's available.  It's free. >  > K > Well, I won't go into all this again as we did it to death some time ago, J > but it was obviously written by someone with absolutlely no knowledge or > experience in education.   >  > B >>>Note: All existing educational programs, such as the Campuswide< >>>Software License Grant (CSLG) program, remain in place. " >>> F >>>However, for what Bill wanted I think it would be an administrativeI >>>nightmare.  For each system you have to request one PAK for each piece J >>>of software you wish to use.  And they expire every August 31st.  WhichG >>>means you would be requesting somewhere around 10-15 PAKs per system > >>>per year.  For a 100 workstation lab or class that would be >>>unimaginable. >>J >>I seem to recall that when I was in school, students were informed what = >>texts, materials, and such would be required for a class.   J >>                                                          It was up tot I >>he student to acquire the textbooks, materials, and such.  I don't see  H >>anything different here.  The student is informed of what's required, G >>including written directions.  College students should be capable of  ( >>acquiring the required class material. >  > L > True, up to a point.  You can't tell them that all programming assignmentsJ > will be done using VMS Pascal and the student is responsible for buying,L > licensing, installing and administering VMS on the Alpha which the studentL > is also expected to purchase.  You could try, but I'll bet the school that- > tried it wouldn't be in business very long.  >  > bill >   H I've read where a PC is expected to be purchased by the student at some ) schools.  Don't know how that worked out.   G I thought we were discussing the students using a workstation that you  E provide?  I only suggested that the student, as part of the learning  G experience, could get the licenses.  I'd expect the computer and media  E to be already available.  At least, that is where the discussion was  B when you said you'd accept a truckload of VAXstation 4000 systems.  G The missing piece is how each student uses the individual license on a  H computer used by many.  I'm wondering whether each computer can get one = license for the year, and multiple people can use any of the  G workstations, using a cluster login.  That seems to be what you'd want.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 19 Dec 2005 15:06:36 -0500 ' From: "Main, Kerry" <Kerry.Main@hp.com> ' Subject: RE: Database access from COBOL R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB773765@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Dave Froble [mailto:davef@tsoft-inc.com]=20 ! > Sent: December 19, 2005 1:38 PM  > To: Info-VAX@Mvb.Saic.Com ) > Subject: Re: Database access from COBOL  >=20 > Bill Gunshannon wrote:4 > > In article <11qbp869ffr60fc@corp.supernews.com>,. > > 	Dave Froble <davef@tsoft-inc.com> writes: > >=20# > >>johnhreinhardt@yahoo.com wrote:  > >> > >>>Bob Koehler wrote:  > >>>  > >>>  > >>> C > >>>>  Use of Hobbyist license for educational institutions is not F > >>>>  permitted, unless there have been changes since Ilast read it. > >>>>? > >>>>  Educational institutions are offered discount programs.  > >>>  > >>> F > >>>That's still correct, however there is a free Educational License > >>>Program > >>> 4 > >>>http://h71000.www7.hp.com/openvmsedu/index.html > >>>   > >>>Copied from the first page: > >>> G > >>>"The OpenVMS Educational License Program is patterned after the HP > > >>>OpenVMS Hobbyist Program. It is offered in response to=20 > your requests H > >>>that we make HP OpenVMS available to educational institutions at no
 > >>>cost. > >>> ; > >>>All types of educational institutions worldwide are=20  > eligible. All you E > >>>have to do is request an access code over the Web, register your B > >>>school, and once registered, anyone at your school can use=20 > the access= > >>>code to get a product authorization key (PAK) through=20  > e-mail. The baseE > >>>license for HP OpenVMS is included, and you may request over 100 B > >>>layered product PAKs. The licenses are good from August to=20 > August andD > >>>must be renewed annually. Anyone associated with an educationalB > >>>institution-students, faculty or administrators-may use these; > >>>licenses. The pages listed below will get you started.  > >>H > >>I don't know what could be much better.  It's available.  It's free. > >=20 > >=20A > > Well, I won't go into all this again as we did it to death=20  > some time ago,B > > but it was obviously written by someone with absolutlely no=20 > knowledge or > > experience in education.=20  > >=20 > >=20D > >>>Note: All existing educational programs, such as the Campuswide> > >>>Software License Grant (CSLG) program, remain in place. " > >>> H > >>>However, for what Bill wanted I think it would be an administrative? > >>>nightmare.  For each system you have to request one PAK=20  > for each pieceB > >>>of software you wish to use.  And they expire every August=20 > 31st.  WhichA > >>>means you would be requesting somewhere around 10-15 PAKs=20  > per system@ > >>>per year.  For a 100 workstation lab or class that would be > >>>unimaginable. > >>@ > >>I seem to recall that when I was in school, students were=20 > informed what=20A > >>texts, materials, and such would be required for a class. =20 @ > >>                                                         =20 > It was up tot=20A > >>he student to acquire the textbooks, materials, and such. =20  > I don't see=20B > >>anything different here.  The student is informed of what's=20 > required,=20@ > >>including written directions.  College students should be=20 > capable of=20 * > >>acquiring the required class material. > >=20 > >=209 > > True, up to a point.  You can't tell them that all=20  > programming assignments 7 > > will be done using VMS Pascal and the student is=20  > responsible for buying, ? > > licensing, installing and administering VMS on the Alpha=20  > which the student A > > is also expected to purchase.  You could try, but I'll bet=20  > the school that / > > tried it wouldn't be in business very long.  > >=20 > > bill > >=20 >=20< > I've read where a PC is expected to be purchased by the=20 > student at some=20+ > schools.  Don't know how that worked out.  >=20B > I thought we were discussing the students using a workstation=20
 > that you=20 I > provide?  I only suggested that the student, as part of the learning=20 A > experience, could get the licenses.  I'd expect the computer=20  > and media=20I > to be already available.  At least, that is where the discussion was=20 D > when you said you'd accept a truckload of VAXstation 4000 systems. >=20> > The missing piece is how each student uses the individual=20 > license on a=20 @ > computer used by many.  I'm wondering whether each computer=20 > can get one=20A > license for the year, and multiple people can use any of the=20 @ > workstations, using a cluster login.  That seems to be what=20
 > you'd want.  >=20    ? Fwiw, lets not forget that where larger numbers of licenses are ; required, I am sure some custom arrangements could be made.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------   Date: 20 Dec 2005 01:12:05 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)' Subject: Re: Database access from COBOL , Message-ID: <40p435F195ufqU1@individual.net>  0 In article <11qdvc66hh3jh92@corp.supernews.com>,* 	Dave Froble <davef@tsoft-inc.com> writes: > J > I've read where a PC is expected to be purchased by the student at some + > schools.  Don't know how that worked out.   I Schools have often worked out agreements with companies that allowed them G to offer really good deals on computers but the only school I even knew E that mandated every student buying a computer was West Point.  Hardly F your traditional school (they mandate a lot of other things that otherC schools could never get away with, too).  But then, rather than the H student paying for his education, the student is actually paid to attend; so making them "buy" a computer was somewhat of a misnomer.    > I > I thought we were discussing the students using a workstation that you  G > provide?  I only suggested that the student, as part of the learning  ' > experience, could get the licenses.     D One, they are not here to learn how to deal with stupid bureaucracy.E Two, when they screw it up and can't do their assignments what do you 4 think the administration would have to say about it?C Three, If the University owns the workstation, who do you think the 1 administration is going to want maintaining them? > Should I go on with all the other things wrong with this idea.  I >                                      I'd expect the computer and media  G > to be already available.  At least, that is where the discussion was  D > when you said you'd accept a truckload of VAXstation 4000 systems. > I > The missing piece is how each student uses the individual license on a  J > computer used by many.  I'm wondering whether each computer can get one ? > license for the year, and multiple people can use any of the  I > workstations, using a cluster login.  That seems to be what you'd want.   F And now you are getting into what people who do this for a living have7 been saying was wrong with the program from the start.     bill   --  J 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>       ------------------------------  % Date: Mon, 19 Dec 2005 23:53:31 -0500 ' From: Dave Froble <davef@tsoft-inc.com> ' Subject: Re: Database access from COBOL 0 Message-ID: <11qf3etdbllqb59@corp.supernews.com>   Main, Kerry wrote: >>-----Original Message-----1 >>From: Dave Froble [mailto:davef@tsoft-inc.com]  ! >>Sent: December 19, 2005 1:38 PM  >>To: Info-VAX@Mvb.Saic.Com ) >>Subject: Re: Database access from COBOL  >> >>Bill Gunshannon wrote: >>3 >>>In article <11qbp869ffr60fc@corp.supernews.com>, - >>>	Dave Froble <davef@tsoft-inc.com> writes:  >>>  >>> # >>>>johnhreinhardt@yahoo.com wrote:  >>>> >>>> >>>>>Bob Koehler wrote:  >>>>>  >>>>>  >>>>>  >>>>> B >>>>>> Use of Hobbyist license for educational institutions is notE >>>>>> permitted, unless there have been changes since Ilast read it.  >>>>>>> >>>>>> Educational institutions are offered discount programs. >>>>>  >>>>> F >>>>>That's still correct, however there is a free Educational License >>>>>Program >>>>> 4 >>>>>http://h71000.www7.hp.com/openvmsedu/index.html >>>>>   >>>>>Copied from the first page: >>>>> G >>>>>"The OpenVMS Educational License Program is patterned after the HP < >>>>>OpenVMS Hobbyist Program. It is offered in response to  >> >>your requests  >>H >>>>>that we make HP OpenVMS available to educational institutions at no
 >>>>>cost. >>>>> 9 >>>>>All types of educational institutions worldwide are   >> >>eligible. All you  >>E >>>>>have to do is request an access code over the Web, register your @ >>>>>school, and once registered, anyone at your school can use  >> >>the access >>; >>>>>code to get a product authorization key (PAK) through   >> >>e-mail. The base >>E >>>>>license for HP OpenVMS is included, and you may request over 100 @ >>>>>layered product PAKs. The licenses are good from August to  >> >>August and >>D >>>>>must be renewed annually. Anyone associated with an educationalB >>>>>institution-students, faculty or administrators-may use these; >>>>>licenses. The pages listed below will get you started.  >>>>H >>>>I don't know what could be much better.  It's available.  It's free. >>>  >>> > >>>Well, I won't go into all this again as we did it to death  >> >>some time ago, >>? >>>but it was obviously written by someone with absolutlely no   >> >>knowledge or >> >>>experience in education.  >>>  >>>  >>> D >>>>>Note: All existing educational programs, such as the Campuswide> >>>>>Software License Grant (CSLG) program, remain in place. " >>>>> H >>>>>However, for what Bill wanted I think it would be an administrative= >>>>>nightmare.  For each system you have to request one PAK   >> >>for each piece >>@ >>>>>of software you wish to use.  And they expire every August  >> >>31st.  Which >>? >>>>>means you would be requesting somewhere around 10-15 PAKs   >> >>per system >>@ >>>>>per year.  For a 100 workstation lab or class that would be >>>>>unimaginable. >>>>> >>>>I seem to recall that when I was in school, students were  >> >>informed what  >>? >>>>texts, materials, and such would be required for a class.   = >>>>                                                           >> >>It was up tot  >>? >>>>he student to acquire the textbooks, materials, and such.    >> >>I don't see  >>@ >>>>anything different here.  The student is informed of what's  >> >>required,  >>> >>>>including written directions.  College students should be  >>
 >>capable of   >>* >>>>acquiring the required class material. >>>  >>> 6 >>>True, up to a point.  You can't tell them that all  >> >>programming assignments  >>4 >>>will be done using VMS Pascal and the student is  >> >>responsible for buying,  >>< >>>licensing, installing and administering VMS on the Alpha  >> >>which the student  >>> >>>is also expected to purchase.  You could try, but I'll bet  >> >>the school that  >>. >>>tried it wouldn't be in business very long. >>>  >>>bill  >>>  >>: >>I've read where a PC is expected to be purchased by the  >>student at some + >>schools.  Don't know how that worked out.  >>@ >>I thought we were discussing the students using a workstation  >>that you  G >>provide?  I only suggested that the student, as part of the learning  ? >>experience, could get the licenses.  I'd expect the computer   >>and media G >>to be already available.  At least, that is where the discussion was  D >>when you said you'd accept a truckload of VAXstation 4000 systems. >>< >>The missing piece is how each student uses the individual  >>license on a  > >>computer used by many.  I'm wondering whether each computer  >>can get one ? >>license for the year, and multiple people can use any of the  > >>workstations, using a cluster login.  That seems to be what 
 >>you'd want.  >> >  >  > A > Fwiw, lets not forget that where larger numbers of licenses are = > required, I am sure some custom arrangements could be made.   E S**t!  Now you ARE going to hold his hand.  You'll show up to do the   installation also?   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 19 Dec 2005 17:24:09 -0500 - From: William Webb <william.w.webb@gmail.com> , Subject: Re: DECnet IV over Multinet CircuitI Message-ID: <8660a3a10512191424s7520b129p497646200d45b013@mail.gmail.com>   5 On 12/14/05, Dave Froble <davef@tsoft-inc.com> wrote:  > Bill Gunshannon wrote: >  > > L > > Our networking people installed a (industry standard) firewall here som= e = > > time ago that has turned TCPIP in to a stateful protocol.   	  [ snip ]  > @ > Anyone described as a 'network person' should be a slave.  TheG > alternative is the chaos we now see where the inmates run the asylum.   D Judging from what I've seen over time concerning how such people areC often (mis)treated and (over)worked, anyone described as a 'network  person' probably IS a slave.  " Or at best, an indentured servant.   WWWebb   >  > --6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > 170 Grimplin Road  > Vanderbilt, PA  15486  >      --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: Mon, 19 Dec 2005 23:10:51 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>2 Subject: Re: Deleting alias files (blocks deleted)0 Message-ID: <%5Hpf.904$%z1.605@news.cpqcorp.net>  H DELETE checks the directory bit and has done so for six+ major releases G of OpenVMS.  If the directory bit is set, the directory file cannot be  8 deleted until the contents of the directory are deleted.  G DELETE does not check the alias bit (and there isn't one of those), it  G deletes the file.  Prior to recent OpenVMS versions, there was no real  H way to check for multiple aliases for a particular file within OpenVMS; % there was no count of aliases around.   I For the general case reported, if tracts of the OpenVMS system disk were  H deleted, then the system disk is arguably corrupt.   Various core parts I of the OpenVMS system disk structure are built on alias entries starting  I with the advent of the cluster common system disk -- which around around  D V3.x or so and eventually became standard structure for system disk B volumes -- and each root on the cluster common system disk shares B directories with any other root present and with the common root. H Aiming an errant DELETE at a system file or system directory can easily ( cause collateral damage, in other words.  G If some or all of the core parts of the particular OpenVMS system disk  F are now gone, restore from your most recent off-line BACKUP/IMAGE, or  reinstall OpenVMS.  H If SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE was deleted (or other core OpenVMS F files), ANALYZE/DISK will not show any errors associated with that -- H this as ANALYZE/DISK is intended to verify the volume structure and not A the volume contents.  You can potentially restore SYSBOOT.EXE to  F SYS$COMMON:[SYSEXE]SYSBOOT.EXE (if that was the target for the DELETE C and that file was an alias entry), and you can remove any detritus  F associated with the SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE entry -- unless C you are using this disk with and have a MicroVAX I or VAXstation I   series system.  I It is possible that the entry in SYSBOOT.EXE was long ago removed or the  H system disk was once erroneously restored, and that the entry was never E connected when the DELETE arrived.  If so, the particular DELETE was  D potentially harmless, save for VAX systems using the version of VMB F built into the those few ancient VAX systems that have assumptions of C the V2-vintage "flat" system disk structures.  AFAIK, this is only  H ancient copies of VMB.EXE found out in equivalently ancient VAX console @ media, and maybe the version of VMB found in the MicroVAX I and E VAXstation I KA610/KD32 class processor.  Anything circa V3 or later  E will look for the common root copy of the SYSBOOT.EXE image and will   boot from there.  < Prior to the DELETE of SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE, a G DIRECTORY/FILE could have been used to verify the files are different;  H that the files are not alias entries based on the displayed FID values. E   (Files that are alias entries can and will have the same FID value    when resident on the same disk.)  H I would also personally tend to consider removing the privileges of any C user that performed this or similar DELETE on a production OpenVMS  B system that I was responsible for, in the most blunt of phrasing. H Evidence of a full off-line system BACKUP/IMAGE made immediately before E any such DELETE activity and/or substantial past evidence of careful  I system operations might potentially be applied as evidence to attempt to  A soften my initial response of MODIFY /PRIV=(NOALL,TMPMBX,NETMBX)  D /DEFPRIV=(NOALL,TMPMBX,NETMBX); in an attempt to dissuade me from a E reaction that would tend to follow such actions as a DELETE aimed at  H system-related files on a production system disk.  That evidence may or E may not actually prevent my removal of system privileges (within the  I production environment) from the user involved, of course.  (This is the  G "delete system privileges first, and ask questions later" reaction, of  G course, and other system managers of production systems may or may not  ! subscribe to this same approach.)   D If this is a hobbyist, non-production, target-practice or a testing D system, well, have at -- a valuable lesson has been learned by some H number of folks involved, and the necessary recovery or re-installation H can now be be commenced.   (If this is an isolated copy of SYSBOOT.EXE, H and a current SYSBOOT is still present in the SYS$COMMON:[SYSEXE] area,  then no recovery is required.)  G If this was a production OpenVMS system that is now being recovered, I  I would also tend to start the recovery with a full off-line BACKUP/IMAGE.  ?   This on the off change that something in the recovery causes  G additional problems, and given there is apparently no current archival   copy available.    ------------------------------  % Date: Mon, 19 Dec 2005 20:02:31 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 2 Subject: Re: Deleting alias files (blocks deleted), Message-ID: <43A75826.C1B04A10@teksavvy.com>   Hoff Hoffman wrote: H > If some or all of the core parts of the particular OpenVMS system diskG > are now gone, restore from your most recent off-line BACKUP/IMAGE, or  > reinstall OpenVMS.    * That is the problem !  They are not gone !  [ DFU DIR/ALIAS said that [000000]SYSBOOT.EXE was an alias for [VMS$COMMON.SYSEXE]SYSBOOT.EXE   H I did a DELETE [000000]SYSBOOT.EXE; (thinking it would simply remove theD directory entry). But it indicated it deleted many blocks.  However,E [VMS$COMMON.SYSEXE]SYSBOOT.EXE was still there, CHECKSUM produced the H same number as a checksum of SYSBOOT.EXE on a different system disk. And ANA/DISK revealed no problems.  I > If SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE was deleted (or other core OpenVMS G > files), ANALYZE/DISK will not show any errors associated with that --     D If I deleted [000000]SYSBOOT.EXE, wouldn't ANA/DISK then report that> [VMS$COMMON.SYSEXE]SYSBOOT.EXE is some rogue directory entry ?  E Wouldn't any access to the file (COPY, CHECKSUM etc) fail since there F would either not be an INDEXF.SYS entry or there would be one pointing to free blocks ?  F > reaction that would tend to follow such actions as a DELETE aimed at4 > system-related files on a production system disk.   G It was a DELETE aimed ate cleaning up that 18 year old system disk that G had accumulated lint and had some file entries in [0000000] that didn't 8 belong there with no recollection of how they got there.    , Is there a way to ensure that tbe blocks in F [VMS$COMMON.SYSEXE]SYSBOOT.EXE  are in fact allocated to that file andF not blocks that were freed when the alias was deleted and the contentsG of that file will eventually change when those blocks are given to some  other file ?  . (isn't that what ANA/DISK/REPAIR checks for ?)   ------------------------------    Date: 19 Dec 2005 18:10:55 -0800$ From: "AEF" <spamsink2001@yahoo.com>2 Subject: Re: Deleting alias files (blocks deleted)B Message-ID: <1135044655.395018.29230@g43g2000cwa.googlegroups.com>   JF Mezei wrote:  > Hoff Hoffman wrote: J > > If some or all of the core parts of the particular OpenVMS system diskI > > are now gone, restore from your most recent off-line BACKUP/IMAGE, or  > > reinstall OpenVMS. >  > , > That is the problem !  They are not gone ! > ] > DFU DIR/ALIAS said that [000000]SYSBOOT.EXE was an alias for [VMS$COMMON.SYSEXE]SYSBOOT.EXE  >   0 It's a problem that you didn't lose SYSBOOT.EXE?  J > I did a DELETE [000000]SYSBOOT.EXE; (thinking it would simply remove theF > directory entry). But it indicated it deleted many blocks.  However,G > [VMS$COMMON.SYSEXE]SYSBOOT.EXE was still there, CHECKSUM produced the J > same number as a checksum of SYSBOOT.EXE on a different system disk. And  > ANA/DISK revealed no problems.  D I'd be a little more careful before deleting files like SYSBOOT.EXE.G Why are you deleting this? Use SET FILE/REMOVE to remove alias entries.   F But it sounds like you WANT to delete SYSBOOT.EXE! Why? Don't you need# that? This is a system disk, right?    > K > > If SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE was deleted (or other core OpenVMS I > > files), ANALYZE/DISK will not show any errors associated with that --  >  > F > If I deleted [000000]SYSBOOT.EXE, wouldn't ANA/DISK then report that@ > [VMS$COMMON.SYSEXE]SYSBOOT.EXE is some rogue directory entry ?  C Only if you deleted the primary entry. That would delete the actual C header and data of the file, leaving an orphaned alias in some .DIR D file. ANAL/DISK would complain about that! But you deleted the aliasD and fortunately still have the primary entry along with the data and file header.   > G > Wouldn't any access to the file (COPY, CHECKSUM etc) fail since there H > would either not be an INDEXF.SYS entry or there would be one pointing > to free blocks ? > H > > reaction that would tend to follow such actions as a DELETE aimed at5 > > system-related files on a production system disk.  > I > It was a DELETE aimed ate cleaning up that 18 year old system disk that I > had accumulated lint and had some file entries in [0000000] that didn't : > belong there with no recollection of how they got there.  D Removing an alias frees up less than one block of disk space. If youE have to resort to deleting important system files to clean up a disk, < how did it ever work with adequate space in the first place?  - > Is there a way to ensure that tbe blocks in H > [VMS$COMMON.SYSEXE]SYSBOOT.EXE  are in fact allocated to that file andH > not blocks that were freed when the alias was deleted and the contentsI > of that file will eventually change when those blocks are given to some  > other file ?  G This question makes no sense. If the blocks are there, you've got them. E If you free only an alias, you still have the primary directory entry G which means you also have the file's data and header. And if you remove E the last directory entry via SET FILE/REMOVE, you will still have the A file's data and header which are recoverable by ANAL/DISK/REPAIR.    > 0 > (isn't that what ANA/DISK/REPAIR checks for ?)  > ANAL/DISK doesn't check for a valid system disk. If it finds aF directory entry pointing to a deleted file header, it will report that? and repair it if you ask. If it finds a file header that has no F directory entry, it will report that and optionally repair it, putting it in [SYSLOST].  D If you're going to play with deleting aliases like this, you need toG learn more about how they work. Check the File Applications manual. And E I'd not mess with SYSBOOT.EXE in the first place unless I were really  sure what I was doing.   ------------------------------  % Date: Mon, 19 Dec 2005 21:01:01 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>2 Subject: Re: Deleting alias files (blocks deleted)+ Message-ID: <43A773EC.8EF8E9F8@comcast.net>    JF Mezei wrote:  >  > Hoff Hoffman wrote: J > > If some or all of the core parts of the particular OpenVMS system diskI > > are now gone, restore from your most recent off-line BACKUP/IMAGE, or  > > reinstall OpenVMS. > , > That is the problem !  They are not gone ! > ] > DFU DIR/ALIAS said that [000000]SYSBOOT.EXE was an alias for [VMS$COMMON.SYSEXE]SYSBOOT.EXE  > J > I did a DELETE [000000]SYSBOOT.EXE; (thinking it would simply remove theF > directory entry). But it indicated it deleted many blocks.  However,G > [VMS$COMMON.SYSEXE]SYSBOOT.EXE was still there, CHECKSUM produced the J > same number as a checksum of SYSBOOT.EXE on a different system disk. And  > ANA/DISK revealed no problems.  H I'd have to wonder what SYSBOOT.EXE was doing in the MFD. Since the fileG exists in the expected path, this was likely a garbage file, anyway. If @ no blocks are trashed according to ANAL/DISK, you should be o.k.   > [snip]I > It was a DELETE aimed ate cleaning up that 18 year old system disk that I > had accumulated lint and had some file entries in [0000000] that didn't : > belong there with no recollection of how they got there.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  # Date: Tue, 20 Dec 2005 04:26:10 GMT   From: John Santos <john@egh.com>2 Subject: Re: Deleting alias files (blocks deleted)* Message-ID: <CJLpf.5218$vJ4.4003@trnddc07>  
 AEF wrote: > JF Mezei wrote:  >  >>Hoff Hoffman wrote:  >>I >>>If some or all of the core parts of the particular OpenVMS system disk H >>>are now gone, restore from your most recent off-line BACKUP/IMAGE, or >>>reinstall OpenVMS.  >> >>, >>That is the problem !  They are not gone ! >>] >>DFU DIR/ALIAS said that [000000]SYSBOOT.EXE was an alias for [VMS$COMMON.SYSEXE]SYSBOOT.EXE  >> >  > 2 > It's a problem that you didn't lose SYSBOOT.EXE? >  > J >>I did a DELETE [000000]SYSBOOT.EXE; (thinking it would simply remove theF >>directory entry). But it indicated it deleted many blocks.  However,G >>[VMS$COMMON.SYSEXE]SYSBOOT.EXE was still there, CHECKSUM produced the J >>same number as a checksum of SYSBOOT.EXE on a different system disk. And  >>ANA/DISK revealed no problems. >  > F > I'd be a little more careful before deleting files like SYSBOOT.EXE.I > Why are you deleting this? Use SET FILE/REMOVE to remove alias entries.  > H > But it sounds like you WANT to delete SYSBOOT.EXE! Why? Don't you need% > that? This is a system disk, right?  >  > J >>>If SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE was deleted (or other core OpenVMSH >>>files), ANALYZE/DISK will not show any errors associated with that -- >> >>F >>If I deleted [000000]SYSBOOT.EXE, wouldn't ANA/DISK then report that@ >>[VMS$COMMON.SYSEXE]SYSBOOT.EXE is some rogue directory entry ? >  > E > Only if you deleted the primary entry. That would delete the actual E > header and data of the file, leaving an orphaned alias in some .DIR F > file. ANAL/DISK would complain about that! But you deleted the aliasF > and fortunately still have the primary entry along with the data and > file header. >   B On a VAX?  Never!  As others have posted, deleting the file by any@ of its aliases releases the file header, and deallocates all theC blocks, leaving any other directory entries as dangling pointers to D a deleted file.  This has always been true, ever since alias entriesH have been possible, VMS V1.0, IIRC.  (SET FILE/ENTER and SET FILE/REMOVEB came about *long* before the cluster common directory structure of V3.7 or so.)  D Only on Alpha VMS 7.3(-1?) or later, on an ODS-5 disk with hardlinksA enabled is the above true.  Ans AFAIK, it doesn't distinguish the A primary entry from alias entries, it just doesn't delete the file " until the link count goes to zero.     > G >>Wouldn't any access to the file (COPY, CHECKSUM etc) fail since there H >>would either not be an INDEXF.SYS entry or there would be one pointing >>to free blocks ? >> >>G >>>reaction that would tend to follow such actions as a DELETE aimed at 4 >>>system-related files on a production system disk. >>I >>It was a DELETE aimed ate cleaning up that 18 year old system disk that I >>had accumulated lint and had some file entries in [0000000] that didn't : >>belong there with no recollection of how they got there. >  > F > Removing an alias frees up less than one block of disk space. If youG > have to resort to deleting important system files to clean up a disk, > > how did it ever work with adequate space in the first place? >  > - >>Is there a way to ensure that tbe blocks in H >>[VMS$COMMON.SYSEXE]SYSBOOT.EXE  are in fact allocated to that file andH >>not blocks that were freed when the alias was deleted and the contentsI >>of that file will eventually change when those blocks are given to some  >>other file ? >  > I > This question makes no sense. If the blocks are there, you've got them. G > If you free only an alias, you still have the primary directory entry I > which means you also have the file's data and header. And if you remove G > the last directory entry via SET FILE/REMOVE, you will still have the C > file's data and header which are recoverable by ANAL/DISK/REPAIR.  >  > 0 >>(isn't that what ANA/DISK/REPAIR checks for ?) >  > @ > ANAL/DISK doesn't check for a valid system disk. If it finds aH > directory entry pointing to a deleted file header, it will report thatA > and repair it if you ask. If it finds a file header that has no H > directory entry, it will report that and optionally repair it, putting > it in [SYSLOST]. > F > If you're going to play with deleting aliases like this, you need toI > learn more about how they work. Check the File Applications manual. And G > I'd not mess with SYSBOOT.EXE in the first place unless I were really  > sure what I was doing.   That I agree with :-)      --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Mon, 19 Dec 2005 23:59:44 -0500 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: Deleting alias files (blocks deleted)0 Message-ID: <11qf3qkk7sd2h93@corp.supernews.com>   Hoff Hoffman wrote:  > J > I would also personally tend to consider removing the privileges of any E > user that performed this or similar DELETE on a production OpenVMS  D > system that I was responsible for, in the most blunt of phrasing. J > Evidence of a full off-line system BACKUP/IMAGE made immediately before G > any such DELETE activity and/or substantial past evidence of careful  K > system operations might potentially be applied as evidence to attempt to  C > soften my initial response of MODIFY /PRIV=(NOALL,TMPMBX,NETMBX)  F > /DEFPRIV=(NOALL,TMPMBX,NETMBX); in an attempt to dissuade me from a G > reaction that would tend to follow such actions as a DELETE aimed at  3 > system-related files on a production system disk.   H Me thinks that Steve is reaching for the large mallet normally used for G smashing fingers, and in extreme cases, flattens the top of heads.  :-)    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 20 Dec 2005 00:13:10 -0500 ' From: Dave Froble <davef@tsoft-inc.com> 2 Subject: Re: Deleting alias files (blocks deleted)0 Message-ID: <11qf4jr1a4omv17@corp.supernews.com>   JF Mezei wrote:  > Hoff Hoffman wrote:  > H >>If some or all of the core parts of the particular OpenVMS system diskG >>are now gone, restore from your most recent off-line BACKUP/IMAGE, or  >>reinstall OpenVMS. >  >  > , > That is the problem !  They are not gone ! > ] > DFU DIR/ALIAS said that [000000]SYSBOOT.EXE was an alias for [VMS$COMMON.SYSEXE]SYSBOOT.EXE   I My recolection is that VMS$COMMON is the primary directory structure for  F the VMS files, and the others are the alias directories.  Still, many A references are to the alias directories.  Seems reasonable to me.   J > I did a DELETE [000000]SYSBOOT.EXE; (thinking it would simply remove theF > directory entry). But it indicated it deleted many blocks.  However,G > [VMS$COMMON.SYSEXE]SYSBOOT.EXE was still there, CHECKSUM produced the J > same number as a checksum of SYSBOOT.EXE on a different system disk. And  > ANA/DISK revealed no problems.  5 As Steve stated, that is NOT the purpose of ANA/DISK!   I >>If SYS$SYSDEVICE:[SYSEXE]SYSBOOT.EXE was deleted (or other core OpenVMS G >>files), ANALYZE/DISK will not show any errors associated with that --  >  >  > F > If I deleted [000000]SYSBOOT.EXE, wouldn't ANA/DISK then report that@ > [VMS$COMMON.SYSEXE]SYSBOOT.EXE is some rogue directory entry ?  
 See above.  G > Wouldn't any access to the file (COPY, CHECKSUM etc) fail since there H > would either not be an INDEXF.SYS entry or there would be one pointing > to free blocks ? >  > F >>reaction that would tend to follow such actions as a DELETE aimed at4 >>system-related files on a production system disk.  >  > I > It was a DELETE aimed ate cleaning up that 18 year old system disk that I > had accumulated lint and had some file entries in [0000000] that didn't : > belong there with no recollection of how they got there.  F When doing so, one assumes the responsibility of knowing EXACTLY what J one is doing, and any errors/mistakes are solely attributed to the guilty.  . > Is there a way to ensure that tbe blocks in H > [VMS$COMMON.SYSEXE]SYSBOOT.EXE  are in fact allocated to that file andH > not blocks that were freed when the alias was deleted and the contentsI > of that file will eventually change when those blocks are given to some  > other file ? > 0 > (isn't that what ANA/DISK/REPAIR checks for ?)  G I'm not sure that you have a problem, nor can I say you don't.  From a  8 V7.2 VAX system, user directories removed from the list:  ! Directory SYS$SYSDEVICE:[000,000]   F 000000.DIR;1                    2/2       [1,1]       (RWED,RWED,RE,E)D BACKUP.DIR;1                    1/1       [BACKUP]    (RWE,RWE,RE,E)E BACKUP.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,) E BADBLK.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,) E BADLOG.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,) E BITMAP.SYS;1                  512/512     [1,1]       (RWED,RWED,RE,) E CONTIN.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,) E CORIMG.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,)  .  .  . E INDEXF.SYS;1                 9831/40066   [1,1]       (RWED,RWED,RE,) E SECURITY.SYS;1                  1/1       [1,1]       (RWED,RWED,RE,) E SYS0.DIR;1                      1/1       [SYSTEM]    (RWE,RWE,RE,RE) D SYSE.DIR;1                      1/1       [1,1]       (RWE,RWE,RE,E)= SYSEXE.DIR;1                    1/1       [SYSTEM]    (R,R,,) E SYSMAINT.DIR;1                  1/1       [SYSTEM]    (RWE,RWE,RE,RE) D TCPIP$FINGER.DIR;1              1/1       [TCPIP$AUX  (RWE,RWE,RE,E)D TCPIP$FTP.DIR;1                 1/1       [TCPIP$AUX  (RWE,RWE,RE,E)D TCPIP$ROUTED.DIR;1              1/1       [SYSTEM]    (RWE,RWE,RE,E)D TCPIP$TFTP.DIR;1                1/1       [TCPIP$AUX  (RWE,RWE,RE,E)D TCPIP$TFTP_ROOT.DIR;1           1/1       [TCPIP$AUX  (RWE,RWE,RE,E)E VMS$COMMON.DIR;1                3/3       [SYSTEM]    (RWE,RWE,RE,RE) E VOLSET.SYS;1                    0/0       [1,1]       (RWED,RWED,RE,)   ? I don't see SYSBOOT in [0,0].  If your system is still booting  D successfully and not showing any problems, maybe you're Ok.  If you E haven't re-booted yet, a good backup of anything you want is advised.   F Maybe since you're having fun with your new systems, now might be the E time to do a fresh install of VMS and then a restore of whatever you  F wish to retain.  That's the only foolproof method I know to clear out  old lint and cruft.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------    Date: 19 Dec 2005 21:14:30 -0800$ From: "AEF" <spamsink2001@yahoo.com>2 Subject: Re: Deleting alias files (blocks deleted)C Message-ID: <1135055670.045920.190600@o13g2000cwo.googlegroups.com>    John Santos wrote: > AEF wrote: > > JF Mezei wrote:  > >  [...] H > >>If I deleted [000000]SYSBOOT.EXE, wouldn't ANA/DISK then report thatB > >>[VMS$COMMON.SYSEXE]SYSBOOT.EXE is some rogue directory entry ? > >  > > G > > Only if you deleted the primary entry. That would delete the actual G > > header and data of the file, leaving an orphaned alias in some .DIR H > > file. ANAL/DISK would complain about that! But you deleted the aliasH > > and fortunately still have the primary entry along with the data and > > file header. > >  > D > On a VAX?  Never!  As others have posted, deleting the file by any   Yes.  B > of its aliases releases the file header, and deallocates all theE > blocks, leaving any other directory entries as dangling pointers to  > a deleted file.    Not on my VAX systems!   Example (run on VAX/VMS v6.1):  
 $ TYPE A.A A.A  $ DIREC/FILE  $ Directory DISK$DATA1:[FELDMAN.DCL.A]    A.A;1                (44834,5,0)   Total of 1 file.
 $ TYPE A.A A.A  $ SET FILE A.A/ENTER=B.B $ DIREC/FILE  $ Directory DISK$DATA1:[FELDMAN.DCL.A]    A.A;1                (44834,5,0)  B.B;1                (44834,5,0)   Total of 2 files. 
 $ DEL B.B;D %DELETE-I-FILDEL, DISK$DATA1:[FELDMAN.DCL.A]B.B;1 deleted (4 blocks) $ DIREC/FILE  $ Directory DISK$DATA1:[FELDMAN.DCL.A]    A.A;1                (44834,5,0)   Total of 1 file.
 $ TYPE A.A A.A  $   5 > This has always been true, ever since alias entries J > have been possible, VMS V1.0, IIRC.  (SET FILE/ENTER and SET FILE/REMOVED > came about *long* before the cluster common directory structure of > V3.7 or so.)  F Interesting. (I did qualify my statement with ISTM.) For what purpose?  They're a pain to keep track of.   > F > Only on Alpha VMS 7.3(-1?) or later, on an ODS-5 disk with hardlinksC > enabled is the above true.  Ans AFAIK, it doesn't distinguish the C > primary entry from alias entries, it just doesn't delete the file $ > until the link count goes to zero.  F But on my VAX systems, deleting any alias leaves the file alone as you? can see above. Deleting the primary entry leaves dangling alias  entries.   ------------------------------    Date: 19 Dec 2005 14:14:12 -0800" From: ggl.20.jjpon@spamgourmet.com Subject: format bug in vmsperlC Message-ID: <1135030452.007053.216130@z14g2000cwz.googlegroups.com>    Greetings all,  + I need some help with format under vmsperl. 2 Please look at the following sample perl program::   #!/usr/bin/perl   " my $str = "First line" . chr(13) .?           "Second line should wrap at thirty chars" . chr(13) .             "Next line". chr(13) .           "Last line". chr(13);    format STDOUT = $     ^<<<<<<<<<<<<<<<<<<<<<<<<<<<<<~~ $str .    write;    E On Activestate perl v5.8.7 (running on Windows XP) I get the expected  outputt:     First line     Second line should wrap at     thirty characters 
     Next line 
     Last line   8 But on VMS perl v5.6.1, it gives me insufficient output:     First line     Second line should wrap at     thirty characters 
     Next line   : The VMS platform is OpenVMS v7.3-2 on an Alphaserver DS10.  / Can somebody please try it and confirm the bug. 0 Maybe even provide a work around or fix to this?   Thanks in advance,   Juna   ------------------------------    Date: 19 Dec 2005 14:39:12 -0800L From: "ggl.20.jjpon@spamgourmetDOT-SEE-OH-EM" <ggl.20.jjpon@spamgourmet.com> Subject: format bug in vmsperlC Message-ID: <1135031951.915544.256350@g47g2000cwa.googlegroups.com>    Greetings all,  + I need some help with format under vmsperl. 2 Please look at the following sample perl program::   #!/usr/bin/perl   " my $str = "First line" . chr(13) .?           "Second line should wrap at thirty chars" . chr(13) .             "Next line". chr(13) .           "Last line". chr(13);    format STDOUT = $     ^<<<<<<<<<<<<<<<<<<<<<<<<<<<<<~~ $str .    write;  E On Activestate perl v5.8.7 (running on Windows XP) I get the expected  outputt:     First line     Second line should wrap at     thirty characters 
     Next line 
     Last line   8 But on VMS perl v5.6.1, it gives me insufficient output:     First line     Second line should wrap at     thirty characters 
     Next line   : The VMS platform is OpenVMS v7.3-2 on an Alphaserver DS10.  / Can somebody please try it and confirm the bug. 0 Maybe even provide a work around or fix to this?   Thanks in advance,   Juna   ------------------------------  % Date: Mon, 19 Dec 2005 22:37:03 -0600 6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler>" Subject: Re: format bug in vmsperlD Message-ID: <craigberry-3A236B.22370319122005@news.isp.giganews.com>  C In article <1135031951.915544.256350@g47g2000cwa.googlegroups.com>, H  "ggl.20.jjpon@spamgourmetDOT-SEE-OH-EM" <ggl.20.jjpon@spamgourmet.com>   wrote:    > Greetings all, > - > I need some help with format under vmsperl. 4 > Please look at the following sample perl program:: >  > #!/usr/bin/perl  > $ > my $str = "First line" . chr(13) .A >           "Second line should wrap at thirty chars" . chr(13) . " >           "Next line". chr(13) .! >           "Last line". chr(13);  >  > format STDOUT = & >     ^<<<<<<<<<<<<<<<<<<<<<<<<<<<<<~~ > $str > .  >  > write; > G > On Activestate perl v5.8.7 (running on Windows XP) I get the expected 
 > outputt: >     First line  >     Second line should wrap at >     thirty characters  >     Next line  >     Last line  > : > But on VMS perl v5.6.1, it gives me insufficient output: >     First line  >     Second line should wrap at >     thirty characters  >     Next line  > < > The VMS platform is OpenVMS v7.3-2 on an Alphaserver DS10. > 1 > Can somebody please try it and confirm the bug. 2 > Maybe even provide a work around or fix to this?  G I can reproduce it with Perl 5.6.2 on OpenVMS Alpha v8.2, but not with  C any version of Perl 5.8.x that I have handy.  5.8.0 came out about  F three and a half years ago, so you are overdue for an upgrade.  5.8.7 @ is current, and there is an HP-supported binary kit of 5.8.6 at:  G http://h71000.www7.hp.com/openvms/products/ips/apache/csws_modperl.html    ------------------------------  % Date: Mon, 19 Dec 2005 14:18:31 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> % Subject: Licenses MOD_UNITS question. , Message-ID: <43A70720.A5CA3680@teksavvy.com>  A On my new 4000-6000, I tried to load my original (real) licenses.    Many are MOD_UNITS.     " C was at 50 units, with MOD_UNITS.  ; Si I figured I could simply do a LICENCE MODIFY C/UNITS=300   H The command worked. But LICENSE LOAD C yields an error message saying it isn't enough units.   # What exactly does MOD_UNITS allow ?   F If I do LICENCE LIST/FULL C, it still shows the original 50 with a new# line below "Modified Units" at 300.   D I am curious as to why this isn't working. (I can take the hobbyyist licence if all elase fails).    , --------------------------------------------  9 LICENSE LOAD DVNETEND loads succesfully with 0 units. But E @SYS$MANAGER:STARTNET  complains about operation requiring a licence, D just after the message announcing "Circuit ISA-0" nessage. I can setC host to myself, but cannot have inbound or outbound traffic between  nodes.    D Any explanation for this ?  This is a 0 units license. So going from: MVII to a 4000-6000 shoudln't have been a problem, right ?  F I have DVNETRTG licence which is disabled. And this was prevent on the& MVII and didn't cause a problem there.   Other licenses seem to work.   ------------------------------  % Date: Mon, 19 Dec 2005 15:20:06 -0500 ' From: Ken Robinson <kenrbnsn@gmail.com> ) Subject: Re: Licenses MOD_UNITS question. H Message-ID: <7dd80f60512191220i7fbd335ek82aa4b91f63dd8b3@mail.gmail.com>  ; On 12/19/05, JF Mezei <jfmezei.spamnot@teksavvy.com> wrote: C > On my new 4000-6000, I tried to load my original (real) licenses.  >  > Many are MOD_UNITS.  >  > $ > C was at 50 units, with MOD_UNITS. > ? > Si I figured I could simply do a LICENCE MODIFY C/UNITS=3D300  > J > The command worked. But LICENSE LOAD C yields an error message saying it > isn't enough units.   C If you issue the command "$ sho lic/charge", you will see what each F type of license requires for your machine. The use "licence modify" to# bump it up to at least that amount.   % > What exactly does MOD_UNITS allow ?   H It's not what MOD_UNITS allows, but what is required by the type of CPU.     Ken Robinson   ------------------------------  % Date: Mon, 19 Dec 2005 15:46:00 -0800 # From: "Tom Linden" <tom@kednos.com> ) Subject: Re: Licenses MOD_UNITS question. ( Message-ID: <ops111myy0zgicya@hyrrokkin>  . On Mon, 19 Dec 2005 18:10:58 -0500, JF Mezei  % <jfmezei.spamnot@teksavvy.com> wrote:    > Ken Robinson wrote: @ >> > Si I figured I could simply do a LICENCE MODIFY C/UNITS=300 > F >> If you issue the command "$ sho lic/charge", you will see what eachI >> type of license requires for your machine. The use "licence modify" to & >> bump it up to at least that amount. >  > D > This licence was and F activity code. Started off at 50 units. Did( > MOD/UNITS=300, and that wasn't enough. > E > SHOW LIC/CHARGE revealed that it needed 300 units for that machine.  > D > Seems that 300 wasn't enough. I put in 400 and then LICENSE LOAD CE > worked. Go figure.  Perhaps a MOD_UNITS decucts the original amount # > (which would have left 250 only).  > H > Now, I have to figure out why the DVNETEND at 0 units isn't enough for > that node.   Try 4800   ------------------------------  % Date: Mon, 19 Dec 2005 20:05:47 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ) Subject: Re: Licenses MOD_UNITS question. , Message-ID: <43A758E9.2A361174@teksavvy.com>   Tom Linden wrote: J > > Now, I have to figure out why the DVNETEND at 0 units isn't enough for > > that node. > 
 > Try 4800  @ Actually LICENSE LOAD DVNETEND works fine. It is @STARTNET whichH complains about a missing licence just after having defined the circuit.  9 NCP LIST EXEC CHAR shows it is defined as nonrouting IV.        C Is there a way to put the LMF into some debug mode to find out what $ license DECNET is expecting to see ?   ------------------------------  % Date: Mon, 19 Dec 2005 20:31:16 -0500 ) From: "Neil Rieck" <n.rieck@sympatico.ca> ) Subject: Re: Licenses MOD_UNITS question. 9 Message-ID: <x9Jpf.1838$%N1.189110@news20.bellglobal.com>   ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  & news:43A758E9.2A361174@teksavvy.com... > Tom Linden wrote: K >> > Now, I have to figure out why the DVNETEND at 0 units isn't enough for  >> > that node.  >> >> Try 4800  > B > Actually LICENSE LOAD DVNETEND works fine. It is @STARTNET whichJ > complains about a missing licence just after having defined the circuit. > : > NCP LIST EXEC CHAR shows it is defined as nonrouting IV. >  >  > E > Is there a way to put the LMF into some debug mode to find out what & > license DECNET is expecting to see ? > M More than a dozen years ago I was working on a project in Montreal. We moved  F and combined a bunch of licences from some older VAXes (which we were L retiring) onto a new VAX-6430. At that time the "C" compiler wouldn't start L because of insufficient license units. The "old master" who was running the G project just walked over to the console and used bumped up the license  J weight by using a command similar to "$lic mod /units=4800 c". Well I was K shocked. He told me it was perfectly legal but would only work on licenses  3 which had MOD_UNITS visible in the OPTIONS display.   L I never did get a good explanation of why it was legal. I just assumed that I maybe all the licenses with MOD_UNITS enabled added up to some value and  K that we were borrowing unused units from some grand account. Can anyone at  # HP shed some light on this subject?   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Tue, 20 Dec 2005 00:48:07 -0500 ' From: Dave Froble <davef@tsoft-inc.com> ) Subject: Re: Licenses MOD_UNITS question. 0 Message-ID: <11qf6l84ltqcg53@corp.supernews.com>   JF Mezei wrote:  > Ken Robinson wrote:  > > >>>Si I figured I could simply do a LICENCE MODIFY C/UNITS=300 >  > E >>If you issue the command "$ sho lic/charge", you will see what each H >>type of license requires for your machine. The use "licence modify" to% >>bump it up to at least that amount.  >  >  > D > This licence was and F activity code. Started off at 50 units. Did( > MOD/UNITS=300, and that wasn't enough. > F > SHOW LIC/CHARGE revealed that it needed 300 units for that machine.  > D > Seems that 300 wasn't enough. I put in 400 and then LICENSE LOAD CE > worked. Go figure.  Perhaps a MOD_UNITS decucts the original amount # > (which would have left 250 only).  > H > Now, I have to figure out why the DVNETEND at 0 units isn't enough for > that node.  - Needs to be 230 if I'm remembering correctly.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------    Date: 19 Dec 2005 12:22:52 -0800' From: "tadamsmar" <tadamsmar@yahoo.com> / Subject: Re: Moving directories of files to dos B Message-ID: <1135023771.937395.85550@g43g2000cwa.googlegroups.com>   Bob Koehler wrote:o > In article <1134963458.237511.107100@g47g2000cwa.googlegroups.com>, "tadamsmar" <tadamsmar@yahoo.com> writes: K > > I want to move some directories that are used for java development from ? > > a Alpha VMS system to a DOS/Windows system.  What is a good E > > way to do that?  I there a good cross-platform zip/unzip program?  > 8 >    My condolences.  ZIP works fine.  Watch out for -a.  G I'm not leaving VMS.   I just need some flexibility in maintaining this  Java application.   I F have a pretty short window (hours or at most a day or two) for loading	 data into E the app every March.  Look at the app and you will see why the window 	 is short:    www.poologic.com   > G > > I want to move the .java files and various ascii data files read by G > > the dos programs.  I will have to convert a few .com files to .bat.  > >    ------------------------------    Date: 19 Dec 2005 16:00:19 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) / Subject: Re: Moving directories of files to dos 3 Message-ID: <1PVUvcWLB30P@eisner.encompasserve.org>   l In article <1135023771.937395.85550@g43g2000cwa.googlegroups.com>, "tadamsmar" <tadamsmar@yahoo.com> writes: >  > Bob Koehler wrote:p >> In article <1134963458.237511.107100@g47g2000cwa.googlegroups.com>, "tadamsmar" <tadamsmar@yahoo.com> writes:L >> > I want to move some directories that are used for java development from@ >> > a Alpha VMS system to a DOS/Windows system.  What is a goodF >> > way to do that?  I there a good cross-platform zip/unzip program? >>9 >>    My condolences.  ZIP works fine.  Watch out for -a.  > I > I'm not leaving VMS.   I just need some flexibility in maintaining this  > Java application.   I H > have a pretty short window (hours or at most a day or two) for loading > data into G > the app every March.  Look at the app and you will see why the window  > is short:  >  > www.poologic.com >   G    OK, then you should be able to move just the .class and data files.  =    No need to recompile Java.   (Yes, I've really done this).    ------------------------------  % Date: Tue, 20 Dec 2005 00:19:50 -0500 ' From: Dave Froble <davef@tsoft-inc.com> / Subject: Re: Moving directories of files to dos 0 Message-ID: <11qf508j4vgt155@corp.supernews.com>   tadamsmar wrote: > Bob Koehler wrote: > o >>In article <1134963458.237511.107100@g47g2000cwa.googlegroups.com>, "tadamsmar" <tadamsmar@yahoo.com> writes:  >>J >>>I want to move some directories that are used for java development from> >>>a Alpha VMS system to a DOS/Windows system.  What is a goodD >>>way to do that?  I there a good cross-platform zip/unzip program? >>8 >>   My condolences.  ZIP works fine.  Watch out for -a. >  > I > I'm not leaving VMS.   I just need some flexibility in maintaining this  > Java application.   I H > have a pretty short window (hours or at most a day or two) for loading > data into G > the app every March.  Look at the app and you will see why the window  > is short:  >  > www.poologic.com >  > F >>>I want to move the .java files and various ascii data files read byF >>>the dos programs.  I will have to convert a few .com files to .bat. >>>  >  >   F If you have network connectivity, what's wrong with FTP.  Some of the I better products will automatically detect text vs binary.  Not much of a  G problem in any case.  Set the mode for each type of file and just copy  H them.  FTPing VMS .COM files to windows .BAT files could be done in the  FTP command.   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 20 Dec 2005 10:22:17 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> Y Subject: Re: OT - GETSYI (was:Re: PHONE error - Invalid specification of node or person.  1 Message-ID: <do7pu3$g7h$1@news-02.connect.com.au>    Hi Paul,  7 > The space filling is consistent with the returns from < > f$getsyi("NODENAME") and f$getsyi("SCSNODE") respectively.   Hoorah!   = >  From SYSGEN do a SHOW SCSNODE, and you will see that it is ? > an 8 byte space filled field, consistent with the results for - > Syi$_scsnode and f$getsyi("SCSNODE") above.    And?  J > As to the null filling, I would guess that is down to the author of thatH > piece of code thinking that initializing the buffer to remove previousF > "junk" not a bad idea (heavens knows how many bugs I've corrected inJ > program maintenance which were caused by failing to initialize buffers).  H Yep, if VMS was a self-imposed nanny-state then you'd have to save those@ programmers from themselves at ever opportunity. But wait! GoingK alphabetically, I just looked up sys$asctim and it says it written 23 bytes G and guess what? IT'S WRITTEN 23 BYTES. How bizarre!!! I mean, how could A anyone complain about stomping all over a pre-initialized buffer?   E Ok, but what your suggesting is not unreasonable, "Left-Justified and J Space-Filled" is common behaviour for output strings and if that's what itJ was doing then, it might be nice if it was documented, but I wouldn't haveH any complaints. But where the hell does "Left-Justified and NULL-Filled" come from on V.M.S.?  K Is this a section of VMS that was re-written in C? It's nice to have a null J at the end of everything isn't it? printf? Is there a C person in the room> who can explain what function (strcpy(), memcpy()?) does what?  + > down to the author of that  piece of code   J Who knows? Maybe the next author will like &s or s? Nothin' like personalH expression - pink with purple poker-dots? Yep, standards, convention and# predictability have had their time.   H > Perhaps I'm being dense today, but I don't see why you are passing 100H > bytes to retrieve Syi$_scsnode. If you are calling from COBOL, why not > call using BUFF(1:6) ?  H For Pete's sake, the 100 was the exaggerated case, yes I am only using 6K bytes. But if I use syi$_nodename and the SCSNODE is less than 6 bytes long J then I get these bollocks "nulls" for free. (As I said, not the end of the" world but a nuisance nonetheless,)  J So getting back to my original point, is everyone happy with the following9 $getsyi item code descriptions from the current VMS docs?   
 SYI$_NODENAME J Returns, as a character string, the name of the [scsnode?,decnet?,galaxy?] node in the buffer specified in the item list. G Because this name can include up to 15 [why 15?] characters, the buffer  length field1 in the item descriptor should specify 15 (bytes).  SYI$_SCSNODEE Returns the Galaxy instance name. [Does it really?] Supported only on  AlphaServer systems thatL support partitioning. [My system doesn't support partitioning, I'm stuffed!]  J If anyone would like to provide an alternative definitions then that would be just great.   Regards Richard Maher   3 "Paul Sture" <paul.sture@decus.ch> wrote in message & news:40nnnhF1b5s4fU1@individual.net... > Richard Maher wrote: > > Hi,  > >  > > Examples Correction: -; > > Syi$_nodename writes the 3 bytes scsnode plus 97 nulls. G > > Syi$_scsnode writes 3 bytes scsnode? + 5 bytes of spaces + 92 nulls  (Room  > > for "::"?) > >  > 7 > The space filling is consistent with the returns from < > f$getsyi("NODENAME") and f$getsyi("SCSNODE") respectively. > = >  From SYSGEN do a SHOW SCSNODE, and you will see that it is ? > an 8 byte space filled field, consistent with the results for - > Syi$_scsnode and f$getsyi("SCSNODE") above.  > I > However, the SYSGEN help says "Specify SCSNODE as a string of up to six  > characters." > J > As to the null filling, I would guess that is down to the author of thatH > piece of code thinking that initializing the buffer to remove previousF > "junk" not a bad idea (heavens knows how many bugs I've corrected inJ > program maintenance which were caused by failing to initialize buffers). > H > Perhaps I'm being dense today, but I don't see why you are passing 100H > bytes to retrieve Syi$_scsnode. If you are calling from COBOL, why not > call using BUFF(1:6) ?   ------------------------------  % Date: Mon, 19 Dec 2005 15:05:56 -0500 - From: William Webb <william.w.webb@gmail.com> ; Subject: OT: Humor: Best Windows Fix story I've ever heard. I Message-ID: <8660a3a10512191205q6dce7ef5re463bce7914f3f40@mail.gmail.com>   F The tech writer who is currently underemployed as our admin was having< home computer problems.  XP refused to let her update Norton? Anti-Virus, and she was getting spurious "Messages from WIndows B Registry" which strongly hinted that there was some spyware/trojan activity going on.  F I burned her a CD with three freeware anti-spyware products (said "use7 all three") and three free firewalls (said "pick one").    Gave her this on Friday.  ! Today she tells us the following:   C After three attempts to re-install OS, reboot cable modem, etc, she A gives up in frustration and decides to go and do some work on the  wiring in her spare bedroom.  D She slips, blows the main fuse in the house and is without power for twenty minutes.   H When she manages to get power restored, everything comes up and is fine.  L First time I've ever heard of somebody fixing Windows by rebooting their HO= USE.   WWWebb   --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: Mon, 19 Dec 2005 22:22:43 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)N Subject: Re: PHONE error - Invalid specification of node or person. Try again.$ Message-ID: <do7brj$927$1@online.de>  3 In article <CFJ+MEiDP$C7@eisner.encompasserve.org>, > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:   F > In article <do0rgj$ril$2@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    M > > On a related note, when will PHONE run over TCPIP?  MONITOR can run over  / > > TCPIP if there is no DECnet; why not PHONE?  > C >    The usual solution is to use talk, which already runs over IP.  > & >    Why solve the same problem twice?  I Because PHONE is an integral part of VMS.  I long for the days when most  H everything one needed was there, already installed, just worked and was G well documented.  Third-party products are OK for some things and I do  I use them but I would like to see VMS avoid becoming a "box of tool" like   unix.    ------------------------------    Date: 19 Dec 2005 21:57:25 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) N Subject: Re: PHONE error - Invalid specification of node or person. Try again.3 Message-ID: <$LwUL1uRUXkB@eisner.encompasserve.org>   w In article <do7brj$927$1@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: 5 > In article <CFJ+MEiDP$C7@eisner.encompasserve.org>, @ > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:  > G >> In article <do0rgj$ril$2@online.de>, helbig@astro.multiCLOTHESvax.de 5 > (Phillip Helbig---remove CLOTHES to reply) writes:   > N >> > On a related note, when will PHONE run over TCPIP?  MONITOR can run over 0 >> > TCPIP if there is no DECnet; why not PHONE? >>  D >>    The usual solution is to use talk, which already runs over IP. >>  ' >>    Why solve the same problem twice?  > + > Because PHONE is an integral part of VMS.   @ So is DECnet Phase V, which will run over TCP/IP (of any flavor)& allowing the use of PHONE over TCP/IP.   ------------------------------  % Date: Tue, 20 Dec 2005 00:42:08 -0500 ' From: Dave Froble <davef@tsoft-inc.com> N Subject: Re: PHONE error - Invalid specification of node or person. Try again.0 Message-ID: <11qf6a35639cf20@corp.supernews.com>  / Phillip Helbig---remove CLOTHES to reply wrote: 5 > In article <CFJ+MEiDP$C7@eisner.encompasserve.org>, @ > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:  >  > F >>In article <do0rgj$ril$2@online.de>, helbig@astro.multiCLOTHESvax.de > 5 > (Phillip Helbig---remove CLOTHES to reply) writes:   >  > L >>>On a related note, when will PHONE run over TCPIP?  MONITOR can run over . >>>TCPIP if there is no DECnet; why not PHONE? >>C >>   The usual solution is to use talk, which already runs over IP.  >>& >>   Why solve the same problem twice? >  > K > Because PHONE is an integral part of VMS.  I long for the days when most  J > everything one needed was there, already installed, just worked and was I > well documented.  Third-party products are OK for some things and I do  K > use them but I would like to see VMS avoid becoming a "box of tool" like   > unix.  >   G PHONE has limited usefulness.  Unless an application has a mailbox set  G up to intercept PHONE, and process the message within the application,  G you either get a user's screen screwed up, or you block incoming PHONE  ) calls.  Yes, at times it has been useful.   I Why do you want a TCPIP transport?  If you appriciate the integral parts  A of VMS, then use DECnet, one of the best of those integral parts.   H If you're going to pick and choose, isn't that a bit like the Unix "box  of tool" in some ways?   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Mon, 19 Dec 2005 20:58:14 +0800  From: prep@prep.synonet.com ( Subject: Re: Rules for joining a cluster- Message-ID: <877ja1p7kp.fsf@prep.synonet.com>   / JF Mezei <jfmezei.spamnot@teksavvy.com> writes:    ...   C > In terms of satellite booting, is it correct to say that the only F > security is that of the ethernet address of the satellite node which= > must be entered in LANCP or NCP to trigger a MOP download ?    Not quite...  B > Once this is accepted, the satelite implicitely is accepted as a0 > member of the cluster since it shares the same7 > cluster_authorization.dat file used by the boot node.    Yep.  F > So impersonating an ethernet address is all that is needed to join aE > cluster. Correct ? (I can understand why conversational booting can  > be disabled for MOP booting).   D No, you don't have to fake an address at all. Just request a load of7 NISCA and pass it the root parameters you want today...   F The only thing the address does is provide a short circuit id to avoidF having to stick a load of fruit in the boot command line. Oh, and some@ machine will choke on it as well, so you avoid that one as well.       --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Mon, 19 Dec 2005 14:09:58 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> & Subject: Update on my migration... :-(, Message-ID: <43A7051F.D6C9039A@teksavvy.com>  
 Murphy's law.   D Overnight, I migrated all except the serial ports from my all mighty Microvax II to a 4000-600.    & MVII -> VELO    (has 2 disks. 2 votes)> 4000 -> VELO2   (has 6 DSSI disks 0 votes satellite from BIKE)% BIKE -> 3100    ( has 1 disk, 1 vote)    Goal:    MVII -> VELO1  0 votes 4000 -> VELO   2 votes; BIKE -> 3100   1 votes. (the 4000-200s will be added later)   G However, during the transition, I want Bike to have 2 votes since it is 1 the main anchor, running the web server and email  until the enw VELO has settled.   A Murphy: couldn't figure out how to raise Bike's votes from 1 to 2 F without rebooting. But before that, I had to figure out what was wrongE with the disk, VELO2 just woudln't boot from disk. The delete/alias I E had used made me very suspect about integrity of the disk, and had to H scrounge around to find out how to use writeboot.exe. Turns out that theG previous onwer had left some bootflags, so "B" went to a different root  than SYS0.    H A mistake I made was to try to boot VELO2 "in production". I should haveG limited it to service disks. But going "production" meant that I had to F update each queue meant to run on VELO to now be /ON=VELO2::  And whenD VELO2 became VELO, I had to do the same work backwards, changing the queues to be back to VELO.  F I also had to switch the IPs around. This wasn't that bad. The problemB was that VELO2 inherited the TCPIP configuration from VELO and theF interface was quite different (from QE0 to EZ0).  But once you have itG working, you can dynamically change the IP while the system is running. ( Just make sure you purge the ARP caches.  @ Switching DECNET around wasnt't too tough once both were up. ButH initially, before VELO2 booted on its own, I had to redefine SYS$SYSROOTF to point to the new disk to make sure that it would update that decnetB database. You forget a colon in the DEFINE ?  You can't invoke anyE images anymore :-(  Thankfully DEFINE still works. Hint: the specific G portion of sys$sysroot is trans=(conc,term), but the SYSCOMMON: isn't.    1 The things you learn in the middle of the night !     D The thing about clustering is that each node has a lot of memory. SoA VELO2's disks were remembered as VELO2's even after VELO2 stopped P existing and it was now VELO offering the same disks. (marked host unavailable).  F In fact, at one point, VELO (the 4000) would hang during the boot justD after having joined the cluster. Lots of time wasted on this. In theH end, it seemed that BIKE had gotten too confused with a now long list ofE devices served by nodes which no longer existed as well as nodes that E had switched names.  I was able to keep the cluster instance alive by H having the 4000 stuck during the boot as well as the MVII booted while I@ rebooted the 3100, and it freed up the 4000 to continue to boot.  H But that wasn't all... the 3100 hadn't had its startup procedure updatedE to reflect the new drives, the new location for the queue manager etc K and that caused quite a mess !!!! (I hadn't planned on rebooting that one).   G At one point, had to move the serial console cable from the 4000 to the D port of the 3100 so I could type in commands.  At one point the 3100D just didn't want me to touch TTA2. SET HOST/DTE would just hang. NotF quite sure what caused it to unjam, but it did after much testing. ButD then, when I got cluster problems, quorum was temporarily lost and IG couldn'ty use a decterm to type commands to the OPA0 of the 4000. So my  trusty PSION PDA came in handy.   D I aklso had problem with licenses (different post) so enabled LAT toD access the nodes from decterms.  Just when I though all was fixed, IG logged in with my username and tried ALLIN1 to no avail. In fact, there H were no logicals defined at all. So I went to the OPA0 decterm, manuallyE rant the Allin1 startup preoceure and it all seemed to work. But from D the other decterm, it still was devoid of logicals. Started to thinkH that perhaps some of the modparams.dat parameters weern't sufficient forE the system logical name table.  Turns out that the MVII was, for some G reason, still broadcasting "VELO" service, even though it was now VELO1 > and my DECTERM kept connecting to the MVII instead of the 40000 (explaining why I wasn't seeing any logicals :-)    E My All Mighty Microvax II may have been slow, but it  caused very vew G headaches over a very long period. Now, this one big change has been an E interesting learning experience. Lots of little things that go wrong.   F And in a real life environment, one  might be able to plan such a moveP on paper really well, but there are many unforeseeen glitches that cause delays.   ------------------------------  % Date: Tue, 20 Dec 2005 00:45:34 -0500 ' From: Dave Froble <davef@tsoft-inc.com> * Subject: Re: Update on my migration... :-(0 Message-ID: <11qf6igis0ta543@corp.supernews.com>   JF Mezei wrote:  > Murphy's law.  > F > Overnight, I migrated all except the serial ports from my all mighty > Microvax II to a 4000-600.   > ( > MVII -> VELO    (has 2 disks. 2 votes)@ > 4000 -> VELO2   (has 6 DSSI disks 0 votes satellite from BIKE)' > BIKE -> 3100    ( has 1 disk, 1 vote)  >  > Goal:  >  > MVII -> VELO1  0 votes > 4000 -> VELO   2 votes= > BIKE -> 3100   1 votes. (the 4000-200s will be added later)  > I > However, during the transition, I want Bike to have 2 votes since it is 3 > the main anchor, running the web server and email ! > until the enw VELO has settled.  > C > Murphy: couldn't figure out how to raise Bike's votes from 1 to 2 H > without rebooting. But before that, I had to figure out what was wrongG > with the disk, VELO2 just woudln't boot from disk. The delete/alias I G > had used made me very suspect about integrity of the disk, and had to J > scrounge around to find out how to use writeboot.exe. Turns out that theI > previous onwer had left some bootflags, so "B" went to a different root 
 > than SYS0.   > J > A mistake I made was to try to boot VELO2 "in production". I should haveI > limited it to service disks. But going "production" meant that I had to H > update each queue meant to run on VELO to now be /ON=VELO2::  And whenF > VELO2 became VELO, I had to do the same work backwards, changing the > queues to be back to VELO. > H > I also had to switch the IPs around. This wasn't that bad. The problemD > was that VELO2 inherited the TCPIP configuration from VELO and theH > interface was quite different (from QE0 to EZ0).  But once you have itI > working, you can dynamically change the IP while the system is running. * > Just make sure you purge the ARP caches. > B > Switching DECNET around wasnt't too tough once both were up. ButJ > initially, before VELO2 booted on its own, I had to redefine SYS$SYSROOTH > to point to the new disk to make sure that it would update that decnetD > database. You forget a colon in the DEFINE ?  You can't invoke anyG > images anymore :-(  Thankfully DEFINE still works. Hint: the specific I > portion of sys$sysroot is trans=(conc,term), but the SYSCOMMON: isn't.   > 3 > The things you learn in the middle of the night !  >  > F > The thing about clustering is that each node has a lot of memory. SoC > VELO2's disks were remembered as VELO2's even after VELO2 stopped R > existing and it was now VELO offering the same disks. (marked host unavailable). > H > In fact, at one point, VELO (the 4000) would hang during the boot justF > after having joined the cluster. Lots of time wasted on this. In theJ > end, it seemed that BIKE had gotten too confused with a now long list ofG > devices served by nodes which no longer existed as well as nodes that G > had switched names.  I was able to keep the cluster instance alive by J > having the 4000 stuck during the boot as well as the MVII booted while IB > rebooted the 3100, and it freed up the 4000 to continue to boot. > J > But that wasn't all... the 3100 hadn't had its startup procedure updatedG > to reflect the new drives, the new location for the queue manager etc M > and that caused quite a mess !!!! (I hadn't planned on rebooting that one).  > I > At one point, had to move the serial console cable from the 4000 to the F > port of the 3100 so I could type in commands.  At one point the 3100F > just didn't want me to touch TTA2. SET HOST/DTE would just hang. NotH > quite sure what caused it to unjam, but it did after much testing. ButF > then, when I got cluster problems, quorum was temporarily lost and II > couldn'ty use a decterm to type commands to the OPA0 of the 4000. So my ! > trusty PSION PDA came in handy.  > F > I aklso had problem with licenses (different post) so enabled LAT toF > access the nodes from decterms.  Just when I though all was fixed, II > logged in with my username and tried ALLIN1 to no avail. In fact, there J > were no logicals defined at all. So I went to the OPA0 decterm, manuallyG > rant the Allin1 startup preoceure and it all seemed to work. But from F > the other decterm, it still was devoid of logicals. Started to thinkJ > that perhaps some of the modparams.dat parameters weern't sufficient forG > the system logical name table.  Turns out that the MVII was, for some I > reason, still broadcasting "VELO" service, even though it was now VELO1 @ > and my DECTERM kept connecting to the MVII instead of the 40002 > (explaining why I wasn't seeing any logicals :-) >  > G > My All Mighty Microvax II may have been slow, but it  caused very vew I > headaches over a very long period. Now, this one big change has been an G > interesting learning experience. Lots of little things that go wrong.  > H > And in a real life environment, one  might be able to plan such a moveR > on paper really well, but there are many unforeseeen glitches that cause delays.  C Which is why you test the move before touching production machines.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------   End of INFO-VAX 2005.706 ************************