1 INFO-VAX	Tue, 05 Dec 2006	Volume 2006 : Issue 668       Contents:0 Re: Cluster connection lost when one link fails?4 Re: Is the DCL SEARCH utility callable from FORTRAN?& Moving privileged from VMS 7.2 to 8.x?* Re: Moving privileged from VMS 7.2 to 8.x?* RE: Moving privileged from VMS 7.2 to 8.x?* RE: Moving privileged from VMS 7.2 to 8.x? Re: recursive copy in VMS  Re: recursive copy in VMS  Re: recursive copy in VMS   F ----------------------------------------------------------------------   Date: 4 Dec 2006 12:09:50 -0800  From: "Verne" <verne@wvnet.edu> 9 Subject: Re: Cluster connection lost when one link fails? C Message-ID: <1165262990.856691.109900@n67g2000cwd.googlegroups.com>    Malcolm Dunnett wrote:< > "Volker Halle" <volker_halle@hotmail.com> wrote in message? > news:1165253096.904222.244400@l12g2000cwl.googlegroups.com...  > > Malcolm, > >  > > could you please do a  > >  > > SDA> READ/EXEC/NOLOG > > E > > before the EXA, then the routine name inside SYS$PEDRIVER will be  > > symbolized as well.  > >  > & > SDA> exam 81939680+VCRP$A_END_ACTION6 > FFFFFFFF.81939848:  00000000.8800FA70   "p=FA......"( > SDA> exa @(81939680+VCRP$A_END_ACTION)0 > PEM$IDLE_SBUF:  00000000.00003008   ".0......" > SDA>    F FWIW ... last June we added a dual NIC DS20e to our existing Vax-AlphaG cluster, and started seeing flapping (switching back and forth) between F all the dual NICs of the active SCS connection.  Eventually found some@ DECUS and C.O.V. postings (I think) that suggested that manuallyE adjusting the priority of one to be higher than the other was a "good  thing".   & Hence we now have in our startup file:  F    $ if node .eqs. "VAXA" then  mcr scacp set lan_device exa/prior=3D5F    $ if node .eqs. "ALPHA" then mcr scacp set lan_device eia/prior=3D5F    $ if node .eqs. "BETA" then  mcr scacp set lan_device eia/prior=3D5  D This solved our flapping issue yet maintained the live redundant SCS paths.  4 p=2Es. each of these systems also has a CI interface  
 Verne Britton  WVNET    ------------------------------  % Date: Mon, 04 Dec 2006 18:14:55 -0500 8 From: Stephen Hoffman <Hoff@HoffmanLabs-RemoveThis-.Org>= Subject: Re: Is the DCL SEARCH utility callable from FORTRAN? ( Message-ID: <el2a5g$d3q$1@pyrite.mv.net>  
 VMSguy wrote: M > Looking for a callable routine to invoke SEARCH from a FORTRAN application.  > G > I want to search text files for a string (much like SEARCH) within a   > FORTRAN applicaiton. > # > Or do I need to build one myself?     K    As Mr Deininger has mentioned, there is no API for SEARCH -- just SPAWN.   H    As for other approaches, callable versions of regex are around.  GNV D has one that HP was supporting, IIRC.  There are others around, too.  I    As for calling a C subroutine from Fortran, that's possible -- but it  H can be equally easy to call from Fortran into a local C wrapper, and to D then have the C wrapper call into regex or into other C subroutines.   ------------------------------   Date: 4 Dec 2006 15:16:45 -0800 ( From: "Carl Karcher" <wolf351@gmail.com>/ Subject: Moving privileged from VMS 7.2 to 8.x? @ Message-ID: <1165274204.864575.8450@79g2000cws.googlegroups.com>  B I'm pretty certain this can't be done, but to eliminate any doubt:  E I have a privileged image (MPRIV.EXE from Intersystems ISM V6.4) that E is currently running on Alpha VMS 7.3-2. I need to upgrade to VMS 8.3 C but there's no upgrade for that product available - they want me to D migrate to the current product Cache which would require developmentE (that the customer isn't going to pay for). There are no OBJ's on the . distribution CD so relinking is not an option.  C Am I correct in thinking that no amount of tinkering will make that @ existing 7.3 image work on 8.3? E.g. defining sys$base_image andC sys$public_vectors logicals in exec mode to point at V7.3 copies of A those images. Those are the only sharable images in the list from  ANALYZE/IMAGE.     --G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madison ; --                      karcher.n0m0respam@waisman.wisc.edu    ------------------------------  # Date: Tue, 05 Dec 2006 03:16:56 GMT / From: "Jeff Goodwin" <jgoodwin@maine.rrr-r.com> 3 Subject: Re: Moving privileged from VMS 7.2 to 8.x? 4 Message-ID: <Iw5dh.176$nq5.152@twister.nyroc.rr.com>  4 "Carl Karcher" <wolf351@gmail.com> wrote in message : news:1165274204.864575.8450@79g2000cws.googlegroups.com...D > I'm pretty certain this can't be done, but to eliminate any doubt: > G > I have a privileged image (MPRIV.EXE from Intersystems ISM V6.4) that G > is currently running on Alpha VMS 7.3-2. I need to upgrade to VMS 8.3 E > but there's no upgrade for that product available - they want me to F > migrate to the current product Cache which would require developmentG > (that the customer isn't going to pay for). There are no OBJ's on the 0 > distribution CD so relinking is not an option. > E > Am I correct in thinking that no amount of tinkering will make that B > existing 7.3 image work on 8.3? E.g. defining sys$base_image andE > sys$public_vectors logicals in exec mode to point at V7.3 copies of C > those images. Those are the only sharable images in the list from  > ANALYZE/IMAGE. >   @ Perhaps "@sys$update:register_privileged_image help" might help.   -Jeff    >  > --I > -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madison = > --                      karcher.n0m0respam@waisman.wisc.edu  >    ------------------------------  $ Date: Mon, 4 Dec 2006 20:29:03 -0600+ From: "Forster, Michael" <mforster@mcw.edu> 3 Subject: RE: Moving privileged from VMS 7.2 to 8.x? 8 Message-ID: <031a01c71815$27cd4040$46146a8d@mcwcorp.net>  F Knowing Intersystems and ISM and Cache, how would you support such a =J configuration and why? Is the database and data of no importance? Why do =C you need 8.3? Alpha systems are still around and 7.3-2 goes prior = E version support at the end of December and ISM is still supported.=20    -----Original Message-----( From: "Carl Karcher" <wolf351@gmail.com>3 To: "Info-VAX@Mvb.Saic.Com" <Info-VAX@Mvb.Saic.Com>  Sent: 12/04/06 5:23 PM/ Subject: Moving privileged from VMS 7.2 to 8.x?   B I'm pretty certain this can't be done, but to eliminate any doubt:  E I have a privileged image (MPRIV.EXE from Intersystems ISM V6.4) that E is currently running on Alpha VMS 7.3-2. I need to upgrade to VMS 8.3 C but there's no upgrade for that product available - they want me to D migrate to the current product Cache which would require developmentE (that the customer isn't going to pay for). There are no OBJ's on the . distribution CD so relinking is not an option.  C Am I correct in thinking that no amount of tinkering will make that @ existing 7.3 image work on 8.3? E.g. defining sys$base_image andC sys$public_vectors logicals in exec mode to point at V7.3 copies of A those images. Those are the only sharable images in the list from  ANALYZE/IMAGE.     --G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madison ; --                      karcher.n0m0respam@waisman.wisc.edu    ------------------------------  $ Date: Mon, 4 Dec 2006 20:38:07 -0600+ From: "Forster, Michael" <mforster@mcw.edu> 3 Subject: RE: Moving privileged from VMS 7.2 to 8.x? 8 Message-ID: <031b01c71816$673cf679$46146a8d@mcwcorp.net>  G Also are many ISM $ZU calls or other ISM specific MUMPS code/routines = F used? If not and you need to be on 8.3, why would migration to cache =3 other than licensing and testing be a show stopper?    -----Original Message-----( From: "Carl Karcher" <wolf351@gmail.com>3 To: "Info-VAX@Mvb.Saic.Com" <Info-VAX@Mvb.Saic.Com>  Sent: 12/04/06 5:23 PM/ Subject: Moving privileged from VMS 7.2 to 8.x?   B I'm pretty certain this can't be done, but to eliminate any doubt:  E I have a privileged image (MPRIV.EXE from Intersystems ISM V6.4) that E is currently running on Alpha VMS 7.3-2. I need to upgrade to VMS 8.3 C but there's no upgrade for that product available - they want me to D migrate to the current product Cache which would require developmentE (that the customer isn't going to pay for). There are no OBJ's on the . distribution CD so relinking is not an option.  C Am I correct in thinking that no amount of tinkering will make that @ existing 7.3 image work on 8.3? E.g. defining sys$base_image andC sys$public_vectors logicals in exec mode to point at V7.3 copies of A those images. Those are the only sharable images in the list from  ANALYZE/IMAGE.     --G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madison ; --                      karcher.n0m0respam@waisman.wisc.edu    ------------------------------  $ Date: Mon, 4 Dec 2006 14:43:46 -0500 From: norm.raphael@metso.com" Subject: Re: recursive copy in VMSQ Message-ID: <OFDB9E4FAC.E6871395-ON8525723A.006BB508-8525723A.006C60EF@metso.com>   ? "AEF" <spamsink2001@yahoo.com> wrote on 12/04/2006 12:02:35 PM:    >  > JF Mezei wrote:  > > AEF wrote: > > = > > > Normally only page- and swap-files have this attribute.  > >  > > J > > Due to the disk space growing at a faster rate than tape drives (in my caseK > > anyways),  I have often employed the SET FILE/NOBACKUP for files that I < > > know are copies of stuff than can be obtained elsewhere. > > I > > For instance, when I backup a mac to my vax, all those files are then J > > marked /NOBACKUP so that they don't take up huge amounts of space on a tapeG > > (and slow the backup process). Similarly, large patches that can be 6 > > downloadeed from the HP site get marked /NOBACKUP. > $ > But you cut too much from my post: > 
 > I wrote: > G > Normally only page- and swap-files have this attribute. And since you D > DON'T want to back up their contents, there is usually no need forH > /IGNORE=BACKUP. Other files would have had to have been explicitly setG > with the nobackup attribute, in which case you didn't want their data D > backed up anyway! So I think it would be a rare occasion for which > /IGNORE=BACKUP is needed.  >  > OK > H > Now, if you've marked these files NOBACKUP because you don't want themF > backed up, why all of a sudden do you want to use /IGNORE=BACKUP? Do! > you want them backed up or not?  > H > One problem with this approach is that when you restore from tape, allH > the NOBACKUP files will be restored, but without their data. They willI > take up space on the disk and you will have to delete them all. This is H > exactly what you want for pagefiles and swapfiles, but is not what youG > want for the files you mark NOBACKUP. Depending on how many and which G > files you mark NOBACKUP, you might be left with quite a mess to clean  > up!  >   @ Well, yes, of course.  What you are saying is, when you act as aC system manager, you must needs continue to act as a system manager. @ If you alter the attributes of files, you need to understand theB consequences of those altered attributes and pay attention in thisC case to backup/restore behavior resulting from your actions.  While C I can see a potential problem "down the line" with the next manager D not knowing of your actions and hence of their consequences, I stillC think it okay to think once and a while and do what you find solves D your concerns.  If you do not get it, of course, you should ask hereD first, and this discussion will let you make an informed decision as- to policy, procedure, and (gasp) documenting.   E Actually, if you are not careful, the restorer may never realize that 3 those files are for all intents and purposes empty.    Thanks for the value-add.      > AEF  >    ------------------------------   Date: 4 Dec 2006 18:18:09 -0800 $ From: "AEF" <spamsink2001@yahoo.com>" Subject: Re: recursive copy in VMSB Message-ID: <1165285089.921115.173050@f1g2000cwa.googlegroups.com>   norm.raphael@metso.com wrote: A > "AEF" <spamsink2001@yahoo.com> wrote on 12/04/2006 12:02:35 PM:  >  > >  > > JF Mezei wrote:  > > > AEF wrote: > > > ? > > > > Normally only page- and swap-files have this attribute.  > > >  > > > L > > > Due to the disk space growing at a faster rate than tape drives (in my > caseM > > > anyways),  I have often employed the SET FILE/NOBACKUP for files that I > > > > know are copies of stuff than can be obtained elsewhere. > > > K > > > For instance, when I backup a mac to my vax, all those files are then L > > > marked /NOBACKUP so that they don't take up huge amounts of space on a > tapeI > > > (and slow the backup process). Similarly, large patches that can be 8 > > > downloadeed from the HP site get marked /NOBACKUP. > > & > > But you cut too much from my post: > >  > > I wrote: > > I > > Normally only page- and swap-files have this attribute. And since you F > > DON'T want to back up their contents, there is usually no need forJ > > /IGNORE=BACKUP. Other files would have had to have been explicitly setI > > with the nobackup attribute, in which case you didn't want their data F > > backed up anyway! So I think it would be a rare occasion for which > > /IGNORE=BACKUP is needed.  > >  > > OK > > J > > Now, if you've marked these files NOBACKUP because you don't want themH > > backed up, why all of a sudden do you want to use /IGNORE=BACKUP? Do# > > you want them backed up or not?  > > J > > One problem with this approach is that when you restore from tape, allJ > > the NOBACKUP files will be restored, but without their data. They willK > > take up space on the disk and you will have to delete them all. This is J > > exactly what you want for pagefiles and swapfiles, but is not what youI > > want for the files you mark NOBACKUP. Depending on how many and which I > > files you mark NOBACKUP, you might be left with quite a mess to clean  > > up!  > >  > B > Well, yes, of course.  What you are saying is, when you act as aE > system manager, you must needs continue to act as a system manager. B > If you alter the attributes of files, you need to understand theD > consequences of those altered attributes and pay attention in thisE > case to backup/restore behavior resulting from your actions.  While E > I can see a potential problem "down the line" with the next manager F > not knowing of your actions and hence of their consequences, I stillE > think it okay to think once and a while and do what you find solves F > your concerns.  If you do not get it, of course, you should ask hereF > first, and this discussion will let you make an informed decision as/ > to policy, procedure, and (gasp) documenting.  > G > Actually, if you are not careful, the restorer may never realize that 5 > those files are for all intents and purposes empty.  >  > Thanks for the value-add.  >  >  > > AEF  > >   G Well, I speak from experience. When I was a graduate student I became a E new VMS user. The research group I was in had its own VAXcluster. The E system manager set up a scratch area on a disk and marked files on it C "NOBACKUP" so he could back up the disk and "only save files in the F non-scratch area". Well, one day he did a restore (failed disk, defragF -- I don't remember) and to everyone's surprise all the NOBACKUP filesE came back and were filled with random bytes! Not a happy moment, very A confusing (the system manager was a one of the researchers, not a F professional VMS system manager) and a real mess to clean up. But thisE is EXACTLY the behavior you want for pagefiles and swapfiles, which I D realized only much later. So I wouldn't be surprised if that is what this attribute was added for.    AEF    ------------------------------   Date: 4 Dec 2006 18:29:50 -0800 $ From: "AEF" <spamsink2001@yahoo.com>" Subject: Re: recursive copy in VMSB Message-ID: <1165285790.132985.233190@16g2000cwy.googlegroups.com>   JF Mezei wrote:  > AEF wrote:J > > Now, if you've marked these files NOBACKUP because you don't want themH > > backed up, why all of a sudden do you want to use /IGNORE=BACKUP? Do# > > you want them backed up or not?  >  > M > I want the /IGNORE=NOBACKUP to do disk to disk copies. (for instance moving M > the data to a new generation disk/system). I was the NOBACKUP attribute for  > when I do regular backups.  < Well, then it's like I said earlier: It's rare that you needC /IGNORE=NOBACKUP. How often are you copying directory trees disk to  disk?   C Anyway, the OP's OP didn't give any clear indication of what he was C trying to accomplish. He used SYS$LOGIN in a way that couldn't have E worked and wanted to restore to a top directory called [TEST]. He did G ask about recursive copy, not disk-to-disk copy, but who knows what his  true motivation is.    OK   AEF    ------------------------------   End of INFO-VAX 2006.668 ************************