1 INFO-VAX	Sat, 15 Apr 2006	Volume 2006 : Issue 208       Contents: Re: DSPP and OpenVMS media Re: DSPP and OpenVMS media Re: DSPP and OpenVMS mediaC Re: Getting device/directory info for an arbitrary user (F$GETUAI?) C Re: Getting device/directory info for an arbitrary user (F$GETUAI?) C Re: Getting device/directory info for an arbitrary user (F$GETUAI?) E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E RE: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). E Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64). D Re: Last VAX release (WAS: How does a fix become a published patch?)* Re: manually resetting file revision dates4 Re: set file empty (dfu set <file>/eblock=1/ebyte=0) SFTP oddity when PUTting files$ Re: SoyMail & insufficient privilege  F ----------------------------------------------------------------------  # Date: Sat, 15 Apr 2006 00:32:05 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com># Subject: Re: DSPP and OpenVMS media 2 Message-ID: <9aX%f.6398$OA6.1165@news.cpqcorp.net>   Mark Schafer wrote:   J > If VMS Engineering provided o/s images, I would consider providing that F > download also. (Hoff: call me if I'm missing out on an opportunity.)  E    There have been informal engineering discussions around providing  F block-based disk image copies of the distribution disks (eg: ISO, raw H disk, or binary copies, depending on your preferred terminology), but I H am presently aware of no particular demand for and accordingly no plans  to provide these images.  G    There are a couple of issues that would have to be resolved as part  E of generating and providing this version of these images, of course.  I (Nothing in technology is ever quite as easy as it looks (nor as easy as  I the marketeers involved might lead you to believe; but I digress), and I  F already know where a couple of the problem areas here will lurk.  And E generating these images can involve more than a straight copy, in at   least one case.  :-)  F    But again, if this particular distribution mechanism is desired or L desirable, then it is certainly something that should be formally requested.   ------------------------------  + Date: Fri, 14 Apr 2006 19:52:20 -0500 (CDT) * From: sms@antinode.org (Steven M. Schweda)# Subject: Re: DSPP and OpenVMS media 2 Message-ID: <06041419522018_202002C3@antinode.org>  , From: Hoff Hoffman <hoff-remove-this@hp.com>   > Mark Schafer wrote:  > L > > If VMS Engineering provided o/s images, I would consider providing that H > > download also. (Hoff: call me if I'm missing out on an opportunity.) > G >    There have been informal engineering discussions around providing  H > block-based disk image copies of the distribution disks (eg: ISO, raw J > disk, or binary copies, depending on your preferred terminology), but I J > am presently aware of no particular demand for and accordingly no plans  > to provide these images.  D    There may be no demand from the _paying_ customers, but I suspectG that a lot of us parasites would be pleased to be able to suck down VMS 4 CD-ROM images.  You know, like the ones for Solaris:  1       http://www.sun.com/software/solaris/get.jsp   F They seem to use Zip-compressed CD-ROM images (although in some cases,G the compression doesn't buy much).  (The file names say "ISO", but it's  a lie.)   I >    There are a couple of issues that would have to be resolved as part  G > of generating and providing this version of these images, of course.  K > (Nothing in technology is ever quite as easy as it looks (nor as easy as  K > the marketeers involved might lead you to believe; but I digress), and I  H > already know where a couple of the problem areas here will lurk.  And G > generating these images can involve more than a straight copy, in at   > least one case.  :-)  2    I'm sure that the details would be interesting.  H >    But again, if this particular distribution mechanism is desired or N > desirable, then it is certainly something that should be formally requested.  F    Would adding "If you please, Sir" provide the kind of formality you seek?   H ------------------------------------------------------------------------  3    Steven M. Schweda               sms@antinode-org 4    382 South Warwick Street        (+1) 651-699-9818    Saint Paul  MN  55105-2547    ------------------------------  % Date: Fri, 14 Apr 2006 21:44:50 -0400 % From: BRAD <bradhamilton@comcast.net> # Subject: Re: DSPP and OpenVMS media ( Message-ID: <44405012.50607@comcast.net>   Steven M. Schweda wrote:. > From: Hoff Hoffman <hoff-remove-this@hp.com> [...] H >>   But again, if this particular distribution mechanism is desired or N >>desirable, then it is certainly something that should be formally requested. > H >    Would adding "If you please, Sir" provide the kind of formality you > seek?   F I suspect (but I could be wrong) that such a request would be akin to I the impetus for the original Hobbyist program, proposed (I think) in the  1 mid/late 1990s at a DECUS symposium (or several?)   I As a Hobbyist, I would be glad to make a formal proposal at the upcoming  G Symposium this fall, but alas, I no longer make a living with VMS, and  6 attendance at future Symposia is problematic, at best.  E Perhaps someone who is a Hobbyist, customer, articulate, and will be  2 attending the Symposium could make the effort?	:-)   ------------------------------  # Date: Fri, 14 Apr 2006 18:28:45 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>L Subject: Re: Getting device/directory info for an arbitrary user (F$GETUAI?)2 Message-ID: <xRR%f.6354$Mq6.2769@news.cpqcorp.net>   Dave Froble wrote:  C > It seems to me that reading the SYSUAF record to get the data is  J > elegant, accurate, and proper.  That's what you would do with any other 0 > data file.  For your purposes, it's just data.  F    There are UAI-based tools around that allow access into the SYSUAF G and related files from DCL, but the supported means is via sys$getuai.  G (The unfortunate problem here is the lack of a wildcard in sys$getuai.)   H    Going direct after SYSUAF might be easier in some ways, but it risks I running into any changes that might be propagated through in the current  F or future releases.  The lack of the wildcard then tosses you back at  direct access.  G    One brute-force means is a variation of the hybrid access I tend to  I use for this -- the code looks at  and adapts to the keys in SYSUAF, and  G then uses a key search to make sys$getuai calls.  (Remember to quickly  I unlock the records as you traverse the key, or to entirely avoid locking  H them.)  From there, for a case such as this, you can write out the data E with formatting as needed into a file, and process that intermediate   file from your DCL code.   ------------------------------  % Date: Fri, 14 Apr 2006 15:45:30 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> L Subject: Re: Getting device/directory info for an arbitrary user (F$GETUAI?), Message-ID: <443FFBD9.ACF77FBB@teksavvy.com>   A simple way to get it:    $write sys$output "Hello World"  $theuser = "JDOE"  $  $allin1/noinit oa$ini_initialize F get log$userdict = UAI$.DEFDEV[CLI$theuser] + UAI$.DEFDIR[CLI$theuser] exit $ P $write sys$output "The directory for user ''theuser' is ''f$trnlnm("userdict")'"   ------------------------------  % Date: Fri, 14 Apr 2006 21:41:01 -0400 / From: "William Webb" <william.w.webb@gmail.com> L Subject: Re: Getting device/directory info for an arbitrary user (F$GETUAI?)I Message-ID: <8660a3a10604141841y15f8f5adn25d0ddbdcd6a5f22@mail.gmail.com>   4 On 4/13/06, Dave Froble <davef@tsoft-inc.com> wrote:. > Alan Winston - SSRL Central Computing wrote:L > > So my actual problem is that I'm reworking my delete-a-user-and-backup-= all-his L > > -files routine.  The version I -was- using trawled through my big user = diskJ > > and did a backup/delete of every file owned by that user, which caughtI > > whatever he owned whether it was in his directories or not, but makes L > > _enormous_ save sets because it saves the whole directory structure of = the L > > disk.  I need to keep those savesets around for a year, so when they ru= n L > > 250mb for 10 mb of actual files, it makes a difference.  And now I have=  _two_L > > big user disks, so if I do the trawl for each disk, I'll end up with _t= wo_ 5 > > gigantic savesets, one of the guaranteed useless.  > > H > > Usage patterns have changed, in any case, since I wrote this routineJ > > originally, and I think now it would be fine to just backup/delete the > > user's directory tree. > > L > > What's the recommended way for DCL to get information for a particular = user out- > > of the SYSUAF?  I want his /dev and /dir.  > > L > > (It seems a little excessive to do a keyed read of SYSUAF and then F$EX= TRACT L > > the relevant bytes out of the record.  I've got a Datatrieve record def= inition L > > for SYSUAF and I could call Datatrieve and have it FN$SET_LOGICAL for m= e,< > > but again, this seems inelegant. Also inelegant would be > >   > > $ AUTHORIZE LIST 'USER /FULL > >  > > and parsing the result.) > > L > > I'd like something that returns a distinct status if the user isn't in = the UAF, > > while I'm asking.  > > 7 > > So is there a generally accepted way of doing this?  > > - > > (7.3-2, usually, but could do it on 8.2).  > >  > > Thanks,  > >  > > -- Alan  > B > It seems to me that reading the SYSUAF record to get the data isI > elegant, accurate, and proper.  That's what you would do with any other 0 > data file.  For your purposes, it's just data. >  > --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  > B GETUAI sufficeth for most folks, except, perhaps for those who areD bound and determined to roll their own/reinvent the wheel, whatever.  D That's all good and well if it's really your heart's desire, just as( long as you remember to open it /SHARED.  B I was a witness to a situation where a programmer (who should haveD known better) opened SYSUAF.DAT and forgot that particular flag on aF system with several thousand users and hundreds of batch jobs running.  A It wasn't pretty, to say the least, and the Tier 1 helpdesk phone % lines lit up like a Christmas tree...    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: 14 Apr 2006 12:20:38 -0700- From: "Doug Phillips" <dphill46@netscape.net> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1145042438.233569.189250@e56g2000cwe.googlegroups.com>    JF Mezei wrote:  > Doug Phillips wrote:E > > (BTW: I propose a contest with the Grand Prize of "Fleeting Fame" J > > awarded to the person who can get to the first mention of OpenVMS with( > > the fewest mouse-clicks from Cerner, > H > Hey, I though I would be smart and find VMS with the "SEARCH" functionJ > of the Cerner web site. Unfortunatly, even with javascript enabled, thatI > doesn't work. It seems that the form contains only hidden fields, so it F > is impossible to enter the search text before clicking on the submit
 > button ! >   G Rule violation alert! Web-page Search feature is a search engine and is " excluded by original contest rule:  7 "access from Google or other search engine not allowed"    <snip>   >  > OK,  interest of fair game:  >  > > http://www.cerner.com/public/search/default.aspx?ip_text=OpenVMS&ip_server=schqryprd.northamerica.cerner.net:8500&ip_searchWithin=off  >     5 Rule violation alert! URL access is excluded by rule:   > "fewest mouse-clicks from Cerner, Oracle and EMC's home pages"    G > So someone clicking on the above can get to VMS in a single click :-) A > (Looking for "VMS" instead of "OpenVMS" yields a few more hits.  > J > Most of the hits are simply references to patches and documents aimed at > existing customers.     B Entry rejected! "Fleeting Fame" not awarded! Please try again;-)))  C Seriously, this lack of OpenVMS mention doesn't surprise me. Cerner F sell solutions and customers buy solutions and the "stuff" beneath theG application isn't usually all that important to the customer --- except D for cost. Organizations Cerner sells to don't usually go out and buyD hardware & OS, then go shopping for an application that runs on thatG platform. They buy the package that best fits their need and budget. (I + think that was a quote from Mr. Obvious!:-)   E The fact that Cerner stuff isn't cluster-aware is surprising, though,  given their history.   ------------------------------    Date: 14 Apr 2006 13:08:32 -0700- From: "Doug Phillips" <dphill46@netscape.net> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1145045312.528726.311740@i40g2000cwc.googlegroups.com>    David J. Dachtera wrote: > Doug Phillips wrote: > > ' > >  <snip some questions about Cerner>  > ( > <snip some answers> (Thank you, David) > H > > Is it a licensing/cost issue, footprint consideration, or is there aE > > technical reason why a single 64cpu system is more desirable than H > > multiple "lesser" clustered systems in a data intensive environment? > B > The biggest problem is the lack of cluster awareness in both theJ > application and Oracle. Oracle can't run "across nodes" (it can't handleG > bound volume sets either). In this regard, Oracle is "monolithic". So 9 > that one "database node" has to pack one helluva punch!  > ? > Add in the inherent inefficiencies of healthcare "three-tier" I > applications, and the dominoes begin to fall rather quickly from there.  >   C So, I'm amazed that Cerner, given that they've been with VMS for so D long, aren't *leaders* in cluster utilization! Really! It seems like? they're doing exactly the kind of thing for which clusters were 	 invented!   G FWIW, a guy named Jim Cramer** has been thumbs-up-bullish on Cerner for E months. Sees big growth in Health Care automation; believes Cerner is > "best of breed" and that they'll continue to ride that growth.  - It will be a shame if VMS doesn't ride along.    --------D **(www.thestreet.com; "Mad Money", CNBC,...) No wars, please. If youD know who Camer is, you probably either love him or hate him. I don't= care. BTW,  he has an interesting take on AMD vs. Intel, too.    ------------------------------    Date: 14 Apr 2006 13:36:57 -0700- From: "Doug Phillips" <dphill46@netscape.net> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1145047017.507706.113330@g10g2000cwb.googlegroups.com>    Bill Gunshannon wrote:E > In article <1145042438.233569.189250@e56g2000cwe.googlegroups.com>, 2 > 	"Doug Phillips" <dphill46@netscape.net> writes: > > I > > The fact that Cerner stuff isn't cluster-aware is surprising, though,  > > given their history. > E > Not really.  They do AIX as well and someone here already mentioned I > the need to keep different implementations as close as possible meaning ; > special features of on particular platform won't be used.  >  > bill >   D Well, okay. That only makes sense if they hadn't started out so intoE VMS. Clustering is such an obvious fit with their application. And it C should be obvious that other platforms would eventually gain (real)  clustering.   D I understand the all-the-eggs-in-one-basket thing, but at what levelD does the application need to be "cluster aware"? Not at the API. WhyF should the application care. That's a job done in deeper places, isn'tG it? Applications --- meaning, most of the code base --- just read stuff F in, crunch some numbers, twiddle some strings and write stuff out. LetC the RTL's & drivers and all those other black-box things handle the @ mechanics. Cluster awareness should be an installable additional option. JMO.   ------------------------------  % Date: Fri, 14 Apr 2006 16:52:25 -0400 ' From: "Main, Kerry" <Kerry.Main@hp.com> N Subject: RE: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).T Message-ID: <FA60F2C4B72A584DBFC6091F6A2B86840125B096@tayexc19.americas.cpqcorp.net>   > -----Original Message-----J > From: David J. Dachtera [mailto:djesys.nospam@NeOaSrPtAhMlNiOnWk.net]=20 > Sent: April 13, 2006 11:07 PM  > To: Info-VAX@Mvb.Saic.Com : > Subject: Re: Is OpenVMS certified yet for 64-way Hard=20 > partitions (GS1280 M64). >=20  	 [snip ..]     >=20 H > > Given the medical environment is extremely conservative and cautious@ > > (with good reason), making a major change like this would=20 > only be doneA > > if there was serious dissatisfaction or major wall issues.=20  > If a majorB > > wall issue was perceived, then the amount of $'s to certify=20
 > VMS 64cpu's F > > would be a fraction of what it would take to make a major platform
 > > change=20  >=20@ > Cost is not the issue. The issue that no future is seen for=20
 > OpenVMS, at , > least not at the needed scale/performance. >=20( > >- think about staffing changes alone. >=20 > Say, "retrain".  >=20  B Nope - think re-hire.  Careful what you are pushing for as most ITF Managers I have seen would rather re-hire admin type people at cheaperE rates than existing staff than re-train existing staff on an entirely 
 new platform.    > > Also HA and DR means 2 x8 > > number of server replacements. Twice x 64cpu server? >=20D > DR =3D 2x whatever. Not really a consideration. Either you have=20 > DR or you  > don't. >=20  F Perhaps the difference between a 64cpu system and a 32CPU cost has not> been presented to you. Keep in mind that Oracle licenses theirC enterprise DB at $40K/CPU (not system) list. Of course, you can get E discounts, but even at 25% off, that is a huge $ for one system - let 2 alone the two you need to have for DR purposes.=20  B > > > Consider also that we're on the threshold of V8.3 and Cerner > > > has yet to@ > > > certify V8.2 (which will be two versions back, allowing=20 > for V8.2-1). > > >  > >=20B > > First, the recommended version is V8.2-1 on IA64, so Cust's=20 > would not do
 > > 8.2 first  >=20@ > ...unless they were staying on Alpha until Itanic matures a=20 > bit more and' > achieve better performance figures...  >=20= > > and then do 8.2-1. From V7.3-2, they would go straight to  > > V8.2-1.  >=20 > Maybe. >=20  C Absolutely no reason to do 8.2 vs 8.2-1, so I do not know what your 	 point is.   F > > Second - Major app vendors do not certify their app on every majorE > > release of a vendors OS release. These re-cert's can be extremely @ > > expensive. There has to be good reason and if the current=20 > OS version is > > > perceived to be working well, then they will wait for a=20 > good business 4 > > reason to re-cert their app on a new OS version. >=20* > Ever dealt with healthcare very much?=20 >=20  E Yep, and like manufacturing, are very cautious about what they do and H when. Both typically are behind the curve when it comes to being current on new versions.=20   = > > > Consider that the last sale dates for Alpha are just=20  > around the corner ( > > > and Cerner has yet to certify I64. > > >   A Again, do you think Cerner might need a DB being available on the F platform before it certs its application or should it go ahead without it?    :-) :-)      > >=20H > > I wonder if the availability of Oracle 10G on IA64 had any impact on	 > > that?  >=20E > Arguable, but unlikely. Cerner's disenchantment with VMS/HP is fast  > becoming the stuff of legend.  >=20@ > > And before anyone jumps on Oracle, keep in mind that they=20
 > also had to 3 > > get the 10G Alpha release out the door as well.  > >=20A > > Here is a question for you - has Cerner cert'ed Oracle 10G=20  > yet on anyH > > platform? What about RAC? [Perhaps they have, but I am interested to > > confirm] >=209 > 9i-RAC, yes. We're puuting that in soon. 10g, doubtful.  >=20  G So, if putting in Oracle 9i RAC, and assuming it scales even reasonably D well, is this not a better solution to your problem that putting all your egg's in one large basket?   B > > > I know you're an intelligent guy, Dave. I'm sure you can put > > > 2.0 and 2.0 G > > > together come with something approximating 4.0, even allowing for  > > > floating-point errors. > > > C > > > As someone in my office put it, the "marriage" between Cerner  > > > and VMS is > > > seriously on the rocks.  > > >  > >=20
 > > David, > >=20? > > I am not involved, so have no idea of what the Cerner-HP=20  > relationship is A > > right now. I do know that a significant number of both new=20  > and GS1280@ > > upgrades's are continuing to be sold to Cerner sites as I=20
 > see the win 	 > > info.  > >=20B > > Keep in mind the generic pro's-n-con's of scale up vs scale=20 > out. There> > > are obviously places for both. If the app can not scale=20 > out, then the : > > only solution is scale up, but what about HA and DR=20 > requirements now as 	 > > well?  >=20G > As noted earlier, the issue is not the app., its the database and its # > limited of cluster compatibility.  >=20  E RAC is how you do Oracle with active-active OpenVMS clusters and many B Customers use this today, so I am not sure what your point is here= either. Here is a pointer to a Oracle 9i RAC mission critical F testimonial on Alpha OpenVMS active-active cluster: (multi-site btw as well)   ; Oracle 9i RAC/OpenVMS ES45 testimonial and video: (Fraport) E http://h71000.www7.hp.com/openvms/brochures/fraport/fraport.pdf (PDF) B http://h71000.www7.hp.com/openvms/brochures/fraport/fraportag.html (Video)   ! Other Cerner testimonials are at: . http://h71000.www7.hp.com/success-stories.html    H > > You now need 2 64 CPU servers [+ big memory which can cost more than > > CPU's] instead of one. >=200 > Can't do it unless you go active/active (RAC). >=20  G The issue is not Oracle Server or OpenVMS as both support active-active B clusters and both have supported for a very long time - i.e. since Oracle 7, so what is the point?   3 > > And if the second server can not be effectively E > > used, then that is one large pile of HW sitting around doing very = > > little. Does that impact overall cost? Of course it does.  >=20
 > Cost of DR.  >=20H > > As others have stated, the 64CPU cert on OpenVMS Alpha can likely beC > > done, but it will require some firm business justifications and # > > commitment from real Customers.  >=20; > How 'bout, "staying in business"? Is that considered a=20  > "justification"? >=20 > --=20   G Absolutely. However, doing expensive testing when you have no Customers F or ISV's willing to commit to a purchase is a great way to hasten that	 point.=20   G Most vendors (not just HP) are well beyond the techie concept of "build @ it and they will come" attitudes - that went out many years ago.   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: Fri, 14 Apr 2006 16:50:51 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64)., Message-ID: <44400B24.DF29DC07@teksavvy.com>   Doug Phillips wrote:F > I understand the all-the-eggs-in-one-basket thing, but at what level2 > does the application need to be "cluster aware"?  E In its design. Consider a "server" process that receives requests and D queues them for processing by itself and assumed that it is the onlyG process accessing the database/files because all requests are funnelled  through it.   F Making this work in a cluster requires substantial changes because you? would now have 2 or more processes accessing the same database.     D VMS clustering was "transparent" in the days of interactive sessionsH with users at the $ sign all day and running TPU/Compilers. But when youB get into server type applications, the plot thickens. distributingB transactions to spread the load amongst the available nodes is notF trivial unless you can be sure that no client casches IPs adresses andL continuously do DNS requests to make use of DNS load balancing for instance.    C It is definitely doable. But it isn't "out of the box" that way for  every type of application.   ------------------------------    Date: 14 Apr 2006 14:28:55 -0700- From: "Doug Phillips" <dphill46@netscape.net> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).C Message-ID: <1145050135.216888.124780@e56g2000cwe.googlegroups.com>    JF Mezei wrote:  > Doug Phillips wrote:H > > I understand the all-the-eggs-in-one-basket thing, but at what level4 > > does the application need to be "cluster aware"? > G > In its design. Consider a "server" process that receives requests and F > queues them for processing by itself and assumed that it is the onlyI > process accessing the database/files because all requests are funnelled 
 > through it.  >   F Right. Sorry you clipped the next part of my post. Say a program readsB something from someplace (screen, keyboard, machine,...), builds aF data-record key and requests a data-record read or write. It then gets$ an indication of success or failure.  A Where the record came from, or what sent the error status isn't a A matter of concern at that level. Process the record or handle the B error. And, at that level is the majority of the code base and the substance of the application.   H > Making this work in a cluster requires substantial changes because youA > would now have 2 or more processes accessing the same database.  >   % But, that's what database servers do.    > F > VMS clustering was "transparent" in the days of interactive sessionsJ > with users at the $ sign all day and running TPU/Compilers. But when youD > get into server type applications, the plot thickens. distributingD > transactions to spread the load amongst the available nodes is notH > trivial unless you can be sure that no client casches IPs adresses andN > continuously do DNS requests to make use of DNS load balancing for instance. >   D And that's all too deep for the API to even care about. VMS clustersG would need different RTL's & such than Veritas or whatever, but the API  should be the same.    > E > It is definitely doable. But it isn't "out of the box" that way for  > every type of application.  
 You're right.    ------------------------------  % Date: Fri, 14 Apr 2006 22:28:09 -0400 ' From: Dave Froble <davef@tsoft-inc.com> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).9 Message-ID: <mp6dnVfB7oUzxN3ZnZ2dnUVZ_tCdnZ2d@libcom.com>    Doug Phillips wrote: > JF Mezei wrote:  >  >>Doug Phillips wrote: >>D >>>(BTW: I propose a contest with the Grand Prize of "Fleeting Fame"I >>>awarded to the person who can get to the first mention of OpenVMS with ' >>>the fewest mouse-clicks from Cerner,  >>H >>Hey, I though I would be smart and find VMS with the "SEARCH" functionJ >>of the Cerner web site. Unfortunatly, even with javascript enabled, thatI >>doesn't work. It seems that the form contains only hidden fields, so it F >>is impossible to enter the search text before clicking on the submit
 >>button ! >> >  > I > Rule violation alert! Web-page Search feature is a search engine and is $ > excluded by original contest rule: > 9 > "access from Google or other search engine not allowed"  >  > <snip> >  >>OK,  interest of fair game:  >> >> >>>http://www.cerner.com/public/search/default.aspx?ip_text=OpenVMS&ip_server=schqryprd.northamerica.cerner.net:8500&ip_searchWithin=off >> >  > 7 > Rule violation alert! URL access is excluded by rule:  > @ > "fewest mouse-clicks from Cerner, Oracle and EMC's home pages" >  >  > G >>So someone clicking on the above can get to VMS in a single click :-) A >>(Looking for "VMS" instead of "OpenVMS" yields a few more hits.  >>J >>Most of the hits are simply references to patches and documents aimed at >>existing customers.  >  >  > D > Entry rejected! "Fleeting Fame" not awarded! Please try again;-))) > E > Seriously, this lack of OpenVMS mention doesn't surprise me. Cerner H > sell solutions and customers buy solutions and the "stuff" beneath theI > application isn't usually all that important to the customer --- except F > for cost. Organizations Cerner sells to don't usually go out and buyF > hardware & OS, then go shopping for an application that runs on thatI > platform. They buy the package that best fits their need and budget. (I - > think that was a quote from Mr. Obvious!:-)  > G > The fact that Cerner stuff isn't cluster-aware is surprising, though,  > given their history. >   H A good friend works for UPMC (University of Pittsburgh Medical Center). I   Some years ago they were looking for applications.  There was a windoz  F bigot making lots of noise.  My friend pointed out "The software that H best fits the task runs on the 2 operating systems that we get the best 7 results and least problems from.  What's your problem?"   F  From more than a few I've gotten the impression that Cerner is 'the' G application for hospitals.  That's what hospitals buy, the application.   G The only problem now for them is that some have a particular supported  6 platform, and the cost of change will be rather large.  I The problem for HP is losing one of their 'targeted' markets.  (Targeted  F in this sense is those that existed when the list of targeted markets 
 was defined.)    --  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: Fri, 14 Apr 2006 22:30:33 -0400 ' From: Dave Froble <davef@tsoft-inc.com> N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).9 Message-ID: <mp6dnVbB7oWgx93ZnZ2dnUVZ_tCdnZ2d@libcom.com>    Doug Phillips wrote: > Bill Gunshannon wrote: > E >>In article <1145042438.233569.189250@e56g2000cwe.googlegroups.com>, 2 >>	"Doug Phillips" <dphill46@netscape.net> writes: >>H >>>The fact that Cerner stuff isn't cluster-aware is surprising, though, >>>given their history.  >>E >>Not really.  They do AIX as well and someone here already mentioned I >>the need to keep different implementations as close as possible meaning ; >>special features of on particular platform won't be used.  >> >>bill >> >  > F > Well, okay. That only makes sense if they hadn't started out so intoG > VMS. Clustering is such an obvious fit with their application. And it E > should be obvious that other platforms would eventually gain (real) 
 > clustering.  > F > I understand the all-the-eggs-in-one-basket thing, but at what levelF > does the application need to be "cluster aware"? Not at the API. WhyH > should the application care. That's a job done in deeper places, isn'tI > it? Applications --- meaning, most of the code base --- just read stuff H > in, crunch some numbers, twiddle some strings and write stuff out. LetE > the RTL's & drivers and all those other black-box things handle the B > mechanics. Cluster awareness should be an installable additional > option. JMO. >   C Not making any accusations here, but what makes you think that the  I people developing the application were smart enough to take advantage of  G clusters?  They were smart enough to pick a good market.  Sort of like  < grocery stores and funeral homes.  Always going to be there.   --  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: Fri, 14 Apr 2006 22:31:35 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).6 Message-ID: <44406917.21AB3957@NeOaSrPtAhMlNiOnWk.net>   Richard Maher wrote: >  > Hi David,  > $ > >  Oracle can't run "across nodes" > ? > Now I'm definitely interested! Do tell - more details please.   E Not much to tell. Oracle has evolved from its VMS roots to a UN*Xland = item. It runs within its own "realm", oblivious to O.S.-level H clustering. Oracle on another node is entirely another database. It evenB has its own lock-manager, based (I'm told) on lock manager code itG purchased from DEC, and has revised over the years to be o.s.-agnostic.   E > Does Cerner have some sort of strangle-hold on the health industry?  > [snip]   Of course not.  E ...but then, neither has VMS a "strangle-hold" on Cerner-land. Hence, ' this sub-thread and the others like it.    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  % Date: Fri, 14 Apr 2006 22:35:28 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).6 Message-ID: <44406A00.35794AB4@NeOaSrPtAhMlNiOnWk.net>   Dave Froble wrote: >  > John Reagan wrote:E > > I just thought I'd mention that with the last sale date for Alpha J > > systems coming later this year, it is very reasonable for customers toL > > evaluate their current system load and try to predict their future AlphaI > > requirements.  For instance, do you buy another M32 or do you inquire 5 > > about an M32 to M64 upgrade with OpenVMS support.  > > J > > I've heard that while the announced date is October, you might be ableK > > to buy new Alpha system afterwards just as long as you give HP a 'heads J > > up' about your future system needs by October.  All the more reason toH > > do evaluations/predictions today and pass them along to your account > > person/reseller. > >  > H > I'm thinking that some of the problem is that HP announced a last saleG > date while the itanic just isn't there yet in the large system arena.    Now you're catching on!   B > Burning bridges behind you, while the road ahead is still in the( > planning stages, just ain't too smart.  . ...yet it continues to be HP's business model.  4 > Pissing away such customers just isn't too smart.   . ...yet it continues to be HP's business model.   > These are the type of I > customers who continue to spend big bucks.  They are not they type that I > change systems easily, and getting them back in the future would entail  > purchasing IBM.  > " > Still holding onto my IBM stock.  
 Good move.  * ...and get your money outta that mattress!   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  % Date: Fri, 14 Apr 2006 22:50:38 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>N Subject: Re: Is OpenVMS certified yet for 64-way Hard partitions (GS1280 M64).6 Message-ID: <44406D8E.A6FAD5BF@NeOaSrPtAhMlNiOnWk.net>   "Main, Kerry" wrote: >  > > -----Original Message-----I > > From: David J. Dachtera [mailto:djesys.nospam@NeOaSrPtAhMlNiOnWk.net] ! > > Sent: April 13, 2006 11:07 PM  > > To: Info-VAX@Mvb.Saic.Com 9 > > Subject: Re: Is OpenVMS certified yet for 64-way Hard  > > partitions (GS1280 M64). > >  >  > [snip ..]  >  >  >J > > > Given the medical environment is extremely conservative and cautious? > > > (with good reason), making a major change like this would  > > only be done@ > > > if there was serious dissatisfaction or major wall issues. > > If a majorA > > > wall issue was perceived, then the amount of $'s to certify  > > VMS 64cpu's H > > > would be a fraction of what it would take to make a major platform > > > change > > ? > > Cost is not the issue. The issue that no future is seen for  > > OpenVMS, at . > > least not at the needed scale/performance. > > * > > >- think about staffing changes alone. > >  > > Say, "retrain".  > >  >  > Nope - think re-hire.   H Nope - too much of gamble. O.s. proficiency is only a very small part of@ what's needed in healthcare. Can't "buy" that in the job market.  - > Careful what you are pushing for as most IT H > Managers I have seen would rather re-hire admin type people at cheaperG > rates than existing staff than re-train existing staff on an entirely  > new platform.   D You're thinking for-profit. Most healthcare concerns are non-profit. Entirely different motivation.   > > > Also HA and DR means 2 x: > > > number of server replacements. Twice x 64cpu server? > > A > > DR = 2x whatever. Not really a consideration. Either you have 
 > > DR or you 
 > > don't. > >  > H > Perhaps the difference between a 64cpu system and a 32CPU cost has not > been presented to you.  1 Actually, it has, as noted in another sub-thread.   ) > Keep in mind that Oracle licenses their E > enterprise DB at $40K/CPU (not system) list. Of course, you can get G > discounts, but even at 25% off, that is a huge $ for one system - let 1 > alone the two you need to have for DR purposes.   D Exactly the point. Big bux for 32->64, even bigger bux for 2x64, and buku bux for Oracle licensing.   Starting to get the picture?  ? Gee - I almost wonder if Oracle didn't sabotage the database so @ performance falters as scaling/workload exceeds certain metrics.  D > > > > Consider also that we're on the threshold of V8.3 and Cerner > > > > has yet to? > > > > certify V8.2 (which will be two versions back, allowing  > > for V8.2-1). > > > >  > > > A > > > First, the recommended version is V8.2-1 on IA64, so Cust's  > > would not do > > > 8.2 first  > > ? > > ...unless they were staying on Alpha until Itanic matures a  > > bit more and) > > achieve better performance figures...  > > ? > > > and then do 8.2-1. From V7.3-2, they would go straight to 
 > > > V8.2-1.  > > 
 > > Maybe. > >  > E > Absolutely no reason to do 8.2 vs 8.2-1, so I do not know what your  > point is.    No V8.2-1 for Alpha, only I64.  H > > > Second - Major app vendors do not certify their app on every majorG > > > release of a vendors OS release. These re-cert's can be extremely ? > > > expensive. There has to be good reason and if the current  > > OS version is = > > > perceived to be working well, then they will wait for a  > > good business 6 > > > reason to re-cert their app on a new OS version. > > ) > > Ever dealt with healthcare very much?  > >  > G > Yep, and like manufacturing, are very cautious about what they do and J > when. Both typically are behind the curve when it comes to being current > on new versions.  D ...except that due to gov't certification requirements, and the timeB required to acieve them, healthcare tends to be the furthest back.  < > > > > Consider that the last sale dates for Alpha are just > > around the corner * > > > > and Cerner has yet to certify I64. > > > >  > C > Again, do you think Cerner might need a DB being available on the H > platform before it certs its application or should it go ahead without > it?    Say, "middleware".   > > > J > > > I wonder if the availability of Oracle 10G on IA64 had any impact on > > > that?  > > G > > Arguable, but unlikely. Cerner's disenchantment with VMS/HP is fast ! > > becoming the stuff of legend.  > > ? > > > And before anyone jumps on Oracle, keep in mind that they  > > also had to 5 > > > get the 10G Alpha release out the door as well.  > > > @ > > > Here is a question for you - has Cerner cert'ed Oracle 10G > > yet on anyJ > > > platform? What about RAC? [Perhaps they have, but I am interested to > > > confirm] > > ; > > 9i-RAC, yes. We're puuting that in soon. 10g, doubtful.  > >  > I > So, if putting in Oracle 9i RAC, and assuming it scales even reasonably F > well, is this not a better solution to your problem that putting all! > your egg's in one large basket?   1 No - see the prior discussion of licensing costs.   D > > > > I know you're an intelligent guy, Dave. I'm sure you can put > > > > 2.0 and 2.0 I > > > > together come with something approximating 4.0, even allowing for  > > > > floating-point errors. > > > > E > > > > As someone in my office put it, the "marriage" between Cerner  > > > > and VMS is > > > > seriously on the rocks.  > > > >  > > >  > > > David, > > > > > > > I am not involved, so have no idea of what the Cerner-HP > > relationship is @ > > > right now. I do know that a significant number of both new > > and GS1280? > > > upgrades's are continuing to be sold to Cerner sites as I  > > see the win  > > > info.  > > > A > > > Keep in mind the generic pro's-n-con's of scale up vs scale  > > out. There= > > > are obviously places for both. If the app can not scale  > > out, then the 9 > > > only solution is scale up, but what about HA and DR  > > requirements now as  > > > well?  > > I > > As noted earlier, the issue is not the app., its the database and its % > > limited of cluster compatibility.  > >  > G > RAC is how you do Oracle with active-active OpenVMS clusters and many D > Customers use this today, so I am not sure what your point is here	 > either.   , At one point, Cerner refused to certify RAC.  7 > Here is a pointer to a Oracle 9i RAC mission critical H > testimonial on Alpha OpenVMS active-active cluster: (multi-site btw as > well)  > = > Oracle 9i RAC/OpenVMS ES45 testimonial and video: (Fraport) G > http://h71000.www7.hp.com/openvms/brochures/fraport/fraport.pdf (PDF) D > http://h71000.www7.hp.com/openvms/brochures/fraport/fraportag.html	 > (Video)  > # > Other Cerner testimonials are at: 0 > http://h71000.www7.hp.com/success-stories.html  ! Salesdroid stuff. Not impressive.   J > > > You now need 2 64 CPU servers [+ big memory which can cost more than > > > CPU's] instead of one. > > 2 > > Can't do it unless you go active/active (RAC). > >  > I > The issue is not Oracle Server or OpenVMS as both support active-active D > clusters and both have supported for a very long time - i.e. since! > Oracle 7, so what is the point?   , See the prior discussion of licensing costs.  5 > > > And if the second server can not be effectively G > > > used, then that is one large pile of HW sitting around doing very ? > > > little. Does that impact overall cost? Of course it does.  > >  > > Cost of DR.  > > J > > > As others have stated, the 64CPU cert on OpenVMS Alpha can likely beE > > > done, but it will require some firm business justifications and % > > > commitment from real Customers.  > > : > > How 'bout, "staying in business"? Is that considered a > > "justification"? > >  > > -- > I > Absolutely. However, doing expensive testing when you have no Customers H > or ISV's willing to commit to a purchase is a great way to hasten that > point.  H ...unless your competition is already doing it which case you either get$ with the program or GTFO of the way.  I > Most vendors (not just HP) are well beyond the techie concept of "build B > it and they will come" attitudes - that went out many years ago.  H ...along with the platforms that were sold that way during their heyday.H When the business/sales model changed and put the cart before the horse,G the market evaporated, as we have seen - and continue to see - with VMS 
 and Alpha.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------  # Date: Fri, 14 Apr 2006 18:14:01 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>M Subject: Re: Last VAX release (WAS: How does a fix become a published patch?) 1 Message-ID: <JDR%f.6352$Nm6.997@news.cpqcorp.net>    Tom Linden wrote: 3 > On Wed, 12 Apr 2006 14:43:20 -0700, Hoff Hoffman  " > <hoff-remove-this@hp.com> wrote: >  >> Tom Linden wrote: >>* >>> BTW,  is 7.3 the last release for VAX? >>0 >>    It is the current release for OpenVMS VAX. > 4 > Sorry, that was ambiguous, what I meant is 7.3 theA > last to be released for the VAX?  Haven't followed this closely 2 > but I thought that 8.2 was going to be the last.  @    OpenVMS VAX V7.3 is the most current release for VAX systems.  @    OpenVMS I64 V8.2-1 is the most current for Integrity systems.  ;    OpenVMS Alpha V8.2 is the most recent for Alpha systems.   @    OpenVMS F8.3 is the current field test release for Alpha and I Integrity systems, for those sites in the field test or that are running   the SDK.  G    For the official statements on what the next release might be for a  H particular platform, the OpenVMS roadmap is the document -- and I'm not H aware of a statement around the plans for or the version number for the  next release of OpenVMS VAX.  C    (The roadmaps are public statements of the plans for OpenVMS as  G currently established, and these plans are subject to change; they are  E not intended as and not provided as formal commitments.  The wording  F within in the roadmap documents has traditionally strove to make this  status quite clear.)  F    As for the OpenVMS version numbers that might be chosen, there's a A reason why release code names are used.  There have been various  E internal discussions within this area, with the two salient releases  I that immediately occur to me being "V6.4" and "V2.0".  These two numbers  G didn't make it out... I know where these slotted in, but both of these  4 release ended up shipping out with another number...   ------------------------------  # Date: Fri, 14 Apr 2006 18:50:35 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>3 Subject: Re: manually resetting file revision dates 2 Message-ID: <%9S%f.6356$sq6.5514@news.cpqcorp.net>  G Add the following two .psect lines shown below into your local copy of   the code, and off you go.    ---- ...           $fabdef          $fibdef          $atrdef ;     .psect data,noexe,noshr,wrt ;  ; File Attributes Block  ...  ---- ...  ; "          .psect code,nowrt,shr,exe ;           .entry  begin,^M<>  ;  ; open the file  ;  ...    ------------------------------  % Date: Fri, 14 Apr 2006 22:53:18 -0500 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>= Subject: Re: set file empty (dfu set <file>/eblock=1/ebyte=0) 6 Message-ID: <44406E2E.AE4C8D89@NeOaSrPtAhMlNiOnWk.net>   "Steven M. Schweda" wrote: > B > From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net> >  > > mckinneyj@saic.com wrote:  > > >  > > > $ dir/siz=all x.x  > > > $ > > > Directory DISK$CIR:[MCKINNEYJ] > > > & > > > X.X;1                      15/18 > > > $ > > > Total of 1 file, 15/18 blocks. > > > $ set file/attr=ebk:0 x.x  > > > $ set file/trunc x.x > > ) > > Leave that out. It serves no purpose.  > 5 >    Leave what out?  No purpose?  What am I missing?  > = > alp $ dire /size = all xxx.pdf  ! (Output edited slightly.)   F If your goal is to re-use a file header and/or keep a specific version number, than nothing is missed.   G If the goal is to re-use an existing allocation, then what is missed is = that SET FILE/TRUNCATE will release the existing allocation.    8 All depends on what it being sought for whatever reason.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  & Unofficial OpenVMS Marketing Home Page! http://www.djesys.com/vms/market/   ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/   ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/   ------------------------------    Date: 14 Apr 2006 18:17:34 -05007 From: "Gareth V. Williams" <graff@cfa0.cfa.harvard.edu> ' Subject: SFTP oddity when PUTting files . Message-ID: <44401f7e@cfanews.cfa.harvard.edu>  A We have a problem with out VMS 8.2 TCPIP V5.5 ECO 1 systems using A SFTP within command files to copy HTML files generated on our VMS @ cluster over to the Unix-based webservers in use our Computation> Facility.  These command files are fairly simple, just "cd" to@ a specific remote directory and then one or more "put"s to place= files.  Some of these command files have more than 1000 "put" B statements.  The files being copied are typically small, a few 10s= of KB.  The problem is that after copying just a small number B of files (92 in one case, but this number varies for other command& files) error messages start appearing:  G open: J99CD1X.html (src): unspecified failure (server msg: 'syserr: i/o  error, file: J99CD1X.html')   J If I rerun the command procedure, it will fail at the same "put" statment.D If I try running the command procedure on another cluster member, it+ will also fail at the same "put" statement.   F If I allow it to run after the errors start to appear, then eventually the error message changes to:   J open: /data/wdocs/iau/www-docs/Ephemerides/Distant/./J99O04C_1.html (dst):D unspecified failure (server msg: 'syserr: Too many open files, file:? /data/wdocs/iau/www-docs/Ephemerides/Distant/./J99O04C_1.html')   G Am I reading the error message correctly?  Is the "too many open files"   error coming from the Unix side?  J Is there any way around this problem other than splitting the copying over multiple SFTP sessions?      Gareth     --  H ------------------------------------------------------------------------H Gareth V. Williams, MS 18, 60 Garden Street, Cambridge, MA 02138, U.S.A.+ Associate Director, IAU Minor Planet Center H gwilliams@cfa.harvard.edu        http://cfa-www.harvard.edu/iau/mpc.html7 OpenVMS & RISC OS: refined choices in operating systems    ------------------------------  + Date: Fri, 14 Apr 2006 20:20:57 +0000 (UTC) < From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann)- Subject: Re: SoyMail & insufficient privilege ) Message-ID: <e1p072$53s$2@news.BelWue.DE>   h In article <1145032438.499413.64620@e56g2000cwe.googlegroups.com>, "chaos" <dciobanu@rogers.com> writes: >Hello,  > ? >I'm getting the same error: Fatal soyMAIL Error:  Insufficient F >privilege or object protection violation. when trying to log private.! >I'm using Apache 1.3 on VMS 8.2.  > , >Alias /soymail/-/ "/apache$common/soymail/" > % ><Directory ~ "^/cgi-bin/soymail/\~">  >    AuthType Basic & >    AuthName "OpenVMS authentication" >    AuthUserOpenVMS On  >    require valid-user 
 ></Directory>  > 
 >soymail.conf  >  >[private-access]  >*/*/* > & >[logout-realm] OpenVMS authentication > ' >I can login public but I want private.  >  >What am I missing ?  K What URL do you enter in your browser? I got this error when the URL didn't L contain a tilde. So "http://www.server/mail" gave the error although the URLA was mapped correctly. "http://www.server/mail/~" did work though.    Regards,    Christoph Gartmann    --  E  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de   D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------   End of INFO-VAX 2006.208 ************************