1 INFO-VAX	Fri, 25 Mar 2005	Volume 2005 : Issue 168       Contents:@ Re: Announcing Availability of HP OpenVMS Migration Software for@ Re: Announcing Availability of HP OpenVMS Migration Software for@ RE: Announcing Availability of HP OpenVMS Migration Software for@ Re: Announcing Availability of HP OpenVMS Migration Software forP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP RE: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP RE: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to IntegrP Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr$ Re: Carly's Way - article retracted?$ RE: Carly's Way - article retracted?$ re: Carly's Way - article retracted?" CLD symbol table object on Itanium& Re: CLD symbol table object on Itanium& Re: CLD symbol table object on Itanium& Re: CLD symbol table object on Itanium7 Re: How to load a shared img without defining a symbol? 7 Re: How to load a shared img without defining a symbol? 7 Re: How to load a shared img without defining a symbol? 7 Re: How to load a shared img without defining a symbol? P Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteraliasP Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteraliasP RE: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteraliasP Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteralias5 Re: TCPIP: why can an MX record not point to a CNAME?   F ----------------------------------------------------------------------    Date: 25 Mar 2005 06:50:32 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) I Subject: Re: Announcing Availability of HP OpenVMS Migration Software for 3 Message-ID: <6DtgSxR1+ssB@eisner.encompasserve.org>   k In article <4243d058$0$67260$157c6196@dreader2.cybercity.dk>, Karsten Nyblad <nospam@nospam.nospam> writes:  > Larry Kilgallen wrote:Z >> In article <3aharnF69bs4pU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >>   >>  E >>>There is nothing controversial about it at all.  He stated it as a H >>>fact and the majority will accept it as such and go on to make futureF >>>decisions based on that information.  Note, that one of the reasonsJ >>>given for this was that VMS never gets mentioned in the trade journals.F >>>Whre have we heard this before?  And who is the only one capable of >>>doing something about it? >>   >>   >> VMS customers.  > K > n any other forum than this, people would be laughing at the notion that  6 > it is the customers that should promote the product.  / I gather you have never heard of the Macintosh.   I When Forbes does stories centered on customer loyalty, it says something.    ------------------------------    Date: 25 Mar 2005 07:19:15 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) I Subject: Re: Announcing Availability of HP OpenVMS Migration Software for 3 Message-ID: <F6mmHfaADoRz@eisner.encompasserve.org>   k In article <4243d058$0$67260$157c6196@dreader2.cybercity.dk>, Karsten Nyblad <nospam@nospam.nospam> writes:    > In military it is common to G > use computer systems 10-20-30 years.  How many of you are willing to  > > claim that VMS will be around and supported in say 10 years?      You can start counting at 1.    ------------------------------  % Date: Fri, 25 Mar 2005 09:22:46 -0500 # From: "Dan Allen" <dallen@nist.gov> I Subject: RE: Announcing Availability of HP OpenVMS Migration Software for : Message-ID: <JFEPKAPBPMDFDBOIANGDIEHBGDAA.dallen@nist.gov>  H Leads me to ask - "What's the typical lifespan of NASA gear and how well& embedded is VMS in that world today?".   Dan    > -----Original Message-----D > From: Bob Koehler [mailto:koehler@eisner.nospam.encompasserve.org]& > Sent: Friday, March 25, 2005 8:19 AM > To: Info-VAX@Mvb.Saic.Com G > Subject: Re: Announcing Availability of HP OpenVMS Migration Software  > for  >  > G > In article <4243d058$0$67260$157c6196@dreader2.cybercity.dk>, Karsten ' > Nyblad <nospam@nospam.nospam> writes:  >  > > In military it is common to H > > use computer systems 10-20-30 years.  How many of you are willing to@ > > claim that VMS will be around and supported in say 10 years? > ! >    You can start counting at 1.  >  >  >    ------------------------------  % Date: Fri, 25 Mar 2005 06:48:31 -0800 # From: "Tom Linden" <tom@kednos.com> I Subject: Re: Announcing Availability of HP OpenVMS Migration Software for ( Message-ID: <opsn67e5fxzgicya@hyrrokkin>  , On 25 Mar 2005 07:19:15 -0600, Bob Koehler  0 <koehler@eisner.nospam.encompasserve.org> wrote:  I > In article <4243d058$0$67260$157c6196@dreader2.cybercity.dk>, Karsten   ' > Nyblad <nospam@nospam.nospam> writes:  >  >> In military it is common toG >> use computer systems 10-20-30 years.  How many of you are willing to ? >> claim that VMS will be around and supported in say 10 years?  > ! >    You can start counting at 1.  > J I am aware (because of PL/I) of a space application running on biggest VAX8 which will not be decommissioned until 202n, where 1<n<4   ------------------------------  % Date: Fri, 25 Mar 2005 06:57:30 -0800 # From: "Tom Linden" <tom@kednos.com> Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr ( Message-ID: <opsn67t4iyzgicya@hyrrokkin>  K On Fri, 25 Mar 2005 09:48:09 +0100, Karsten Nyblad <nospam@nospam.nospam>    wrote:   > Larry Kilgallen wrote:H >> In article <3aharnF69bs4pU1@individual.net>, bill@cs.uofs.edu (Bill   >> Gunshannon) writes: >>F >>> There is nothing controversial about it at all.  He stated it as aI >>> fact and the majority will accept it as such and go on to make future G >>> decisions based on that information.  Note, that one of the reasons K >>> given for this was that VMS never gets mentioned in the trade journals. G >>> Whre have we heard this before?  And who is the only one capable of  >>> doing something about it?  >>   VMS customers.  > L > n any other forum than this, people would be laughing at the notion that  6 > it is the customers that should promote the product.  I I have voiced the same view in the past and posed the not so rhetorical   	 question, H why?  Particularly given the financial performance, you would think thatH sound business practice would prevail.  But that is not the case.  Where does the buck stop?  > F > As for the statement about VMS being dead from the instructor.  It  J > depends very much on how you define dead.  In military it is common to  H > use computer systems 10-20-30 years.  How many of you are willing to  J > claim that VMS will be around and supported in say 10 years?  VMS will  L > be a tough sale to a military without a commitment from HP that VMS will   > be around for +10 years.   ------------------------------  # Date: Fri, 25 Mar 2005 16:47:04 GMT % From: "Bob Lail" <Robert.Lail@hp.com> Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr 2 Message-ID: <cgX0e.2268$242.2251@news.cpqcorp.net>  : "Michael D. Ober" <mdo.@.wakeassoc..com> wrote in message ) news:4dH0e.40$j27.1029@news.uswest.net... K > Putting this type of porting in plus HPs apparent lack of support for the G > Alpha processor leads me to believe Alpha is a dead end chip.  We're  	 > running J > OpenVMS 7.3-2 on our Alpha Server 1200 and have a software contract that= > includes new versions of VMS, yet haven't received VMS 8.2.  >  > Mike.  >  > E > "Peter 'EPLAN' LANGSTOEGER" <peter@langstoeger.at> wrote in message ( > news:424338f9$1@NEWS.LANGSTOEGER.AT...@ >> In article <uNE0e.2216$yq1.1658@news.cpqcorp.net>, "Bob Lail" > <Robert.Lail@hp.com> writes:J >> >  Why do you say that?  OpenVMS V8.2 for AlphaServers is available andL >> >shipping. The AlphaServer 1200 is one of the supported platforms. FutureG >> >versions will continue to support AlphaServers until at least 2010.  >>L >> Maybe because he might not order it, because his boss never heard of VMS,$ >> or heard it died many years ago ? >>C >> Why do you think, he himself believes that VMS (Alpha) is dead ?  >> >> --  >> Peter "EPLAN" LANGSTOEGER( >> Network and OpenVMS system specialist >> E-mail  peter@langstoeger.at I >> A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist  >> >  >  >    Mke   G I can't argue the assumption, only the perception that OpenVMS is dead  I because the Alpha Microprocessor is a dead end chip. That isn't news, in  H case you have been in a box for FOUR years. Alpha's imminent demise was M announced in 2001, by Compaq, and no I don't believe the conspiracy theories  K about HP/Compaq/INTEL collusion to kill Alpha. Compaq/HP has and continued  J the roadmap for OpenVMS, yes with changes, to get to Itanium. In the mean K time HP will continue to sell AlphaServers for at least another year.  For  L many customers EV68 and EV7 based AlphaServers are the long term, 5-10 year I solution. They will continue to see how HP does with Itanium II and move  L when they feel comfortable. If they don't ever reach that point then HP has C failed them. If that occurs HP will probably loose them a customer.   F   Itanium II is here and performing quite well. OpenVMS available for J Itanium II and will continue to be enhanced. INTEL has a plan for Itanium J IIs future. HP has a plan for OpenVMS's future on BOTH Itanium and Alpha. E Porting is NOT a dirty word, it is a reality of modern day computing.   L      You want to abandon the best OS in the industry because HP changed the @ underlying technology? It must be true, perception is reality !!  H NOTE: How many technologies have IBM's OS's gone through and they still  exist and sell today.   	 \Bob Lail    --  
 Robert G Lail  Pre-Sales Solution Consultant  Corporate Accounts Hewlett-Packard Company 
 Merrimack, NH    ------------------------------    Date: 25 Mar 2005 09:50:31 -0800 From: bob@instantwhip.com Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr B Message-ID: <1111773031.915714.98280@g14g2000cwa.googlegroups.com>  A no, it is the students place to walk out of the class because the  instuctor is a moron ...   ------------------------------  % Date: Fri, 25 Mar 2005 09:48:09 +0100 + From: Karsten Nyblad <nospam@nospam.nospam> Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr = Message-ID: <4243d058$0$67260$157c6196@dreader2.cybercity.dk>    Larry Kilgallen wrote:Y > In article <3aharnF69bs4pU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >  > D >>There is nothing controversial about it at all.  He stated it as aG >>fact and the majority will accept it as such and go on to make future E >>decisions based on that information.  Note, that one of the reasons I >>given for this was that VMS never gets mentioned in the trade journals. E >>Whre have we heard this before?  And who is the only one capable of  >>doing something about it?  >  >  > VMS customers.  I n any other forum than this, people would be laughing at the notion that  4 it is the customers that should promote the product.  C As for the statement about VMS being dead from the instructor.  It  G depends very much on how you define dead.  In military it is common to  E use computer systems 10-20-30 years.  How many of you are willing to  G claim that VMS will be around and supported in say 10 years?  VMS will  I be a tough sale to a military without a commitment from HP that VMS will   be around for +10 years.   ------------------------------  % Date: Fri, 25 Mar 2005 06:49:35 -0500 ' From: "Main, Kerry" <kerry.main@hp.com> Y Subject: RE: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB5953A8@tayexc19.americas.cpqcorp.net>   > -----Original Message-----7 > From: Karsten Nyblad [mailto:nospam@nospam.nospam]=20  > Sent: March 25, 2005 3:48 AM > To: Info-VAX@Mvb.Saic.Com A > Subject: Re: Announcing Availability of HP OpenVMS Migration=20 ? > Software for Alpha to Integrity Servers Alpha to Integrity=20 $ > Servers Alpha to Integrity Servers >=20 > Larry Kilgallen wrote:3 > > In article <3aharnF69bs4pU1@individual.net>,=20 , > bill@cs.uofs.edu (Bill Gunshannon) writes: > >=20 > >=20F > >>There is nothing controversial about it at all.  He stated it as a@ > >>fact and the majority will accept it as such and go on to=20
 > make future G > >>decisions based on that information.  Note, that one of the reasons > > >>given for this was that VMS never gets mentioned in the=20 > trade journals. G > >>Whre have we heard this before?  And who is the only one capable of  > >>doing something about it?  > >=20 > >=20 > > VMS customers. >=20A > n any other forum than this, people would be laughing at the=20  > notion that=206 > it is the customers that should promote the product. >=20  B That is not what was stated. It requires both the vendor and those' interested in the product to defend it.   G Case in point - stand up in front of a bunch of Windows/Linux folks and G say "Windows/Linux is history". Do you think those in the audience will C wait for that respective vendor to correct that obviously incorrect 
 statement?  G > As for the statement about VMS being dead from the instructor.  It=20 A > depends very much on how you define dead.  In military it is=20  > common to=20I > use computer systems 10-20-30 years.  How many of you are willing to=20 B > claim that VMS will be around and supported in say 10 years? =20
 > VMS will=20 ? > be a tough sale to a military without a commitment from HP=20  > that VMS will=20 > be around for +10 years. >=20  C Ok, OpenVMS has a public roadmap on its web site. Can you provide a H pointer that shows where the long term (even a few years) public roadmap is for Solaris, AIX or Windows?   E Re: "How many of you are willing to claim that VMS will be around and  supported in say 10 years?"   G I could say the same about Solaris, AIX or Windows. Heck, with mergers, F buyouts, and government lawsuits, who knows what is going to happen in 10 years?=20  B In reality, I expect all of these to be around and supported in 10H years. I also expect both 32 bit and the more recent 64bit mainframes to( also still be doing well in 10 years.=20   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Fri, 25 Mar 2005 06:38:14 -0500 ' From: "Main, Kerry" <kerry.main@hp.com> Y Subject: RE: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB5953A7@tayexc19.americas.cpqcorp.net>   > -----Original Message-----, > From: nobody [mailto:nobody@nobody.org]=20 > Sent: March 24, 2005 11:55 PM  > To: Info-VAX@Mvb.Saic.Com A > Subject: Re: Announcing Availability of HP OpenVMS Migration=20 A > Software for Alpha to Integrity Servers to Integrity Servers=20  > to Integrity Servers >=20 > "Main, Kerry" wrote:F > > Bill - give me a break. Do you mean to tell me that a room full of> > > seasoned IT Pro's is going to be bashful about speaking=20 > their mind? Yes ' > > sir. No sir. You must be right sir.  >=20 >=20H > If you're in a room full of Windows Weenies, it is sometimes better toH > remain under the radar and not speak out to defend the myths about VMSF > since you will only succeed in discrediting yourself in the minds of; > those Weenies, and when you lose their respect, YOU lose.  >=20A > The VMS image is such that it is often not defensible by the=20  > loyalists.9 > And this is why visible marketing of VMS by HP is so=20  > important. It gives G > us , the loyalists, the amminition to allow us to stand up and defend E > VSM in such situations. Without it, you can't succeed in convincing " > poeple that the myths are false. >=20  D Well, perhaps I am getting to crusty in my passing years, but what IG tend to do now in places like this is that I would make statements like A "Every OS has its strengths and weaknesses. OpenVMS has security, G reliability and availability as strengths. Active-active 96 node (3,072 C cpu's) clustering up to 800km apart and dynamic sharing of CPU's in E different partitions on the same box are just a few examples. Windows < has a better GUI environment and more commercially available applications."  F And if they even hint that Windows has better security, I simply pointG out that the OS vendor itself recommends monthly reboots for bundles of = security patches that have become available since last month.   H The point to emphasize with those less informed is that there is no such2 thing as one OS which is better in every category.  B Again, I am not saying more could or should not be done to improve@ marketing. Perhaps I am getting a bit to crusty, but I think theD industry as a whole needs to start fighting back against trends thatG have gradually made sub-optimal IT practices "acceptable" because those C who know better are worried about being called "dinosaurs" in their = organization.  And yes, I know this is easier said than done.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Fri, 25 Mar 2005 07:02:15 -0800 # From: "Tom Linden" <tom@kednos.com> Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr ( Message-ID: <opsn6711vjzgicya@hyrrokkin>  J On Fri, 25 Mar 2005 06:38:14 -0500, Main, Kerry <kerry.main@hp.com> wrote:  D > Again, I am not saying more could or should not be done to improve > marketing.  & To improve presupposes that it exists.   ------------------------------  % Date: Fri, 25 Mar 2005 09:12:42 -0800 % From: DeanW <dean.woodward@gmail.com> Y Subject: Re: Announcing Availability of HP OpenVMS Migration Software for Alpha to Integr 7 Message-ID: <3f119ada050325091225c3b885@mail.gmail.com>   J On Fri, 25 Mar 2005 06:38:14 -0500, Main, Kerry <kerry.main@hp.com> wrote:F > Well, perhaps I am getting to crusty in my passing years, but what II > tend to do now in places like this is that I would make statements like C > "Every OS has its strengths and weaknesses. OpenVMS has security, I > reliability and availability as strengths. Active-active 96 node (3,072 E > cpu's) clustering up to 800km apart and dynamic sharing of CPU's in G > different partitions on the same box are just a few examples. Windows > > has a better GUI environment and more commercially available > applications."  E Again, you've missed the point. Your statements do nothing to counter 2 the point that the speaker believes VMS is "dead".   ------------------------------  % Date: Fri, 25 Mar 2005 09:41:14 +0100 & From: Paul Sture <paul.sture@decus.ch>- Subject: Re: Carly's Way - article retracted? , Message-ID: <3ai147F6c1lloU1@individual.net>   JF Mezei wrote:     H > I am not sure if HP has strict rules with regards to employees talking8 > to the media. But I woudln't be surprised if they did.  F In my experience employment contracts often carry a clause forbidding : employees from talking to the media about company affairs.  F I recall a few years ago a case of a fairly senior employee issuing a H statement to the press, which got misinterpreted in the report, and the D edict went out that all press releases were to be cleared by the PR L department, as they have the training to spot ambiguous wording, gaffes etc.   ------------------------------  % Date: Fri, 25 Mar 2005 06:21:10 -0500 ' From: "Main, Kerry" <kerry.main@hp.com> - Subject: RE: Carly's Way - article retracted? R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB5953A6@tayexc19.americas.cpqcorp.net>   > -----Original Message-----2 > From: Paul Sture [mailto:paul.sture@decus.ch]=20 > Sent: March 25, 2005 3:41 AM > To: Info-VAX@Mvb.Saic.Com / > Subject: Re: Carly's Way - article retracted?  >=20 > JF Mezei wrote:  >=20 >=20; > > I am not sure if HP has strict rules with regards to=20  > employees talking : > > to the media. But I woudln't be surprised if they did. >=20J > In my experience employment contracts often carry a clause forbidding=20< > employees from talking to the media about company affairs. >=20J > I recall a few years ago a case of a fairly senior employee issuing a=20< > statement to the press, which got misinterpreted in the=20 > report, and the=20H > edict went out that all press releases were to be cleared by the PR=20< > department, as they have the training to spot ambiguous=20 > wording, gaffes etc. >=20  F Yep - this mandate is usually in place for most companies - especially these days.   G Same goes for trade shows etc. That is why news folks sometimes get big > bright coloured badges at shows - to make them easier to spot.  E It is just way to easy for well meaning and/or innocent remarks to be E mis-interpreted, mis-quoted and twisted depending on what the authors  predefined agenda might be.=20  E Course, that is not to say this happens with all news folks, but most / companies prefer to err on the side of caution.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------    Date: 25 Mar 2005 07:52:16 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)- Subject: re: Carly's Way - article retracted? 3 Message-ID: <LblSjWZPlp+6@eisner.encompasserve.org>   s In article <OFFD9D36DA.1A18AF4B-ON85256FCD.0053EA53-85256FCD.0054D18A@cca-int.com>, Tym_Stegner@cca-int.com writes: L > Technology Review, Inc. cannot vouch for the accuracy of "Carly's Way," byH > Michelle Delio, published online on March 4, 2005. Nor can we stand byL > "Carly's Gone. HP Celebrates," also by Delio, published online on February- > 10, 2005. We regret publishing the stories.   H Did this article ever make it in the printed edition? If so, what issue?  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  B         They that can give up essential liberty to obtain a little?         temporary safety deserve neither liberty nor safety. -- A         Benjamin Franklin Historical Review of Pennsylvania. 1759    ------------------------------  # Date: Fri, 25 Mar 2005 16:19:51 GMT " From:   VAXman-  @SendSpamHere.ORG+ Subject: CLD symbol table object on Itanium 0 Message-ID: <00A414B5.06C91059@SendSpamHere.ORG>   OK,   H So how are you supposed to use CLD symbol tables on Itanium?  The linker! balks at them when I try to link.   O %ILINK-W-RELODIFTYPE, relocation requests the linker to build a function descri          symbol: CMDTBLB         relocation section: .rela$CODE$ (section header entry: 37)4         relocation type: RELA$K_R_IA_64_LTOFF_FPTR22         relocation entry: 51         module: CMD +         file: TOP$:[OBJ.IA64.V8_2]CMD.OBJ;2    --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Fri, 25 Mar 2005 17:05:25 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks) / Subject: Re: CLD symbol table object on Itanium - Message-ID: <yAchkdRkV0xn@cuebid.zko.dec.com>   " VAXman-  @SendSpamHere.ORG writes:J > So how are you supposed to use CLD symbol tables on Itanium?  The linker# > balks at them when I try to link.  > K > %ILINK-W-RELODIFTYPE, relocation requests the linker to build a function  
 >   descri >         symbol: CMDTBLD >         relocation section: .rela$CODE$ (section header entry: 37)6 >         relocation type: RELA$K_R_IA_64_LTOFF_FPTR22 >         relocation entry: 51 >         module: CMD - >         file: TOP$:[OBJ.IA64.V8_2]CMD.OBJ;2     What's the definition of CMDTBL?   --    M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com    ------------------------------  # Date: Fri, 25 Mar 2005 17:17:08 GMT " From:   VAXman-  @SendSpamHere.ORG/ Subject: Re: CLD symbol table object on Itanium 0 Message-ID: <00A414BD.078290D9@SendSpamHere.ORG>  _ In article <yAchkdRkV0xn@cuebid.zko.dec.com>, brooks@cuebid.zko.dec.nospam (Rob Brooks) writes: # >VAXman-  @SendSpamHere.ORG writes: K >> So how are you supposed to use CLD symbol tables on Itanium?  The linker $ >> balks at them when I try to link. >>  L >> %ILINK-W-RELODIFTYPE, relocation requests the linker to build a function  >>   descri  >>         symbol: CMDTBL E >>         relocation section: .rela$CODE$ (section header entry: 37) 7 >>         relocation type: RELA$K_R_IA_64_LTOFF_FPTR22  >>         relocation entry: 51  >>         module: CMD. >>         file: TOP$:[OBJ.IA64.V8_2]CMD.OBJ;2 > ! >What's the definition of CMDTBL?  >  >--  > N >Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com  	 ! CMD.CLD 
 Module CMDTBL      --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Fri, 25 Mar 2005 18:05:31 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks) / Subject: Re: CLD symbol table object on Itanium - Message-ID: <lhPBirN90sWC@cuebid.zko.dec.com>   " VAXman-  @SendSpamHere.ORG writes:3 > brooks@cuebid.zko.dec.nospam (Rob Brooks) writes: $ >>VAXman-  @SendSpamHere.ORG writes:L >>> So how are you supposed to use CLD symbol tables on Itanium?  The linker% >>> balks at them when I try to link.  >>> M >>> %ILINK-W-RELODIFTYPE, relocation requests the linker to build a function   >>>   descri >>>         symbol: CMDTBLF >>>         relocation section: .rela$CODE$ (section header entry: 37)8 >>>         relocation type: RELA$K_R_IA_64_LTOFF_FPTR22  >>>         relocation entry: 51 >>>         module: CMD / >>>         file: TOP$:[OBJ.IA64.V8_2]CMD.OBJ;2  >>" >>What's the definition of CMDTBL?  M Yeah, but how is it referenced in the calling program?  I seem to recall that K the I64 linker is a bit more strict about certain attributes than the Alpha  linker.    --    M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com    ------------------------------  + Date: Fri, 25 Mar 2005 08:18:56 +0000 (UTC) & From: Bernhard <lithium@nospam_gmx.at>@ Subject: Re: How to load a shared img without defining a symbol?> Message-ID: <Xns96245E3836D04lithiumnospamgmxat@212.33.32.148>  H "DL Phillips" <whohe@whoever.com> wrote in news:1111695301.396461.271580 @l41g2000cwc.googlegroups.com:    > Bernhard,  > G > Consider that even though you want your application to *run* platform ? > independent, you will need a platform *specific* installation F > procedure. VMS applications are usually installed by the priviledgedH > SYSTEM account and can involve adding commands to the systartup and/orE > sylogicals files which run at system startup time. This is where to G > define logicals you need to point to your application's library file, H > install shared images and otherwise define the environment required by > your application: i.e. > 5 > $ define/system/exec foo_root disk:[foo_directory.] @ > $ define/system/exec foo foo_root:[foo_library]foosharelib.exe	 > and/or, 1 > $ install add foo_root:[foo_library]foolib.exe-   >  /open/header/share/priv=share >  > or whatever. > G > Alternatively, you could place the define's in a command file used to  > run the application: > " > foo_root:[foo_exec]Start_foo.com4 > $ define foo foo_root:[foo_library]foosharelib.exe* > $ run foo_root:[foo_exec]foo_application > % > And, in sylogin.com you could have:  > / > $ DOFOO :== @foo_root:[foo_exec]Start_foo.com  > H > So (in this simplified example) the user would just enter DOFOO to run. > the app, and System.load("foo") should work. >  >  -Doug >  >    Hi Doug,  / Thanks for your infos. Some clarifications..... G First time install of the base application is not the problem! At that  F time doing whatever sort of system operations poses no problem and we D certainly have no restrictions here. This works already, even under  OpenVMS.  D The problem occurs during operation: We are dynamically loading new H "plugins", most of them are in pure Java, so there is no need to do any D OpenVMS related config. But some are using legacy C++ shares. These D shares are of course compiled and linked target platform confirmant.  Such a Java plugin JAR contains:F 1. the Java code to satisfy the interface requirements of our base app> 2. a JNI wrapper to access the native share functions/methods.H 3. a set of platform dependent shares: a DLL for Win, a lib.so for Unix H and a SHR for OpenVMS. The base app automatically detects, which system C it is deployed on and extracts the native share to the file system.   F We have the absolute requirement to load the plugins without stopping I the base and the rest of the plugins! So that's why making Defines gives   us troubles.  F Our base app is an implementation of OSGI. If you are interested, you ' may have a look at http://www.osgi.org.    THX+ Happy Easter  Bernhard   ------------------------------  % Date: Fri, 25 Mar 2005 06:39:19 -0800 # From: "Tom Linden" <tom@kednos.com> @ Subject: Re: How to load a shared img without defining a symbol?( Message-ID: <opsn66ztcvzgicya@hyrrokkin>  L On Fri, 25 Mar 2005 08:18:56 +0000 (UTC), Bernhard <lithium@nospam_gmx.at>   wrote:  G > We have the absolute requirement to load the plugins without stopping J > the base and the rest of the plugins! So that's why making Defines gives > us troubles.G > Our base app is an implementation of OSGI. If you are interested, you ) > may have a look at http://www.osgi.org.   K I wasn't aware how important OSGi was or may become, and I would think that J this should get the attention of VMS management to support you.  Contact   yourC VMS ambassador and have them escalate the visibility of the problem    ------------------------------  % Date: Fri, 25 Mar 2005 10:28:05 -0500 - From: "John E. Malmberg" <wb8tyw@qsl.network> @ Subject: Re: How to load a shared img without defining a symbol?1 Message-ID: <TZmdnX9WBbiYs9nfRVn-2Q@adelphia.com>    Bernhard wrote: = > dooleys@snowy.net.au wrote in news:1111637636.718209.325810   > @g14g2000cwa.googlegroups.com: > F > First option: I am afraid not, as this would involve another define K > operation by a user. This would be the same as the way suggested by HP -  " > define a symbol for the lib..... > I > Second: I am not sure, if I understand your suggestion here. We have a  I > shared lib, which represents the legacy code. We call that code fomr a  , > Java Application - so there is no linking. > We must use either8 > System.loadLibrary(..), which takes a symbolic name or= > System.load(..), which expects an absolute path to the lib.  >  > I tried the latter today: I > extracted the path, where the framework caches the native lib (aka the  J > shared img) to the local file system and issued a System.load call with  > it. ? > Works fine under Win and Unix/Linux, but not so under OpenVMS  > D > The problem seems to be that the HP-JavaVM does not map filenames  > consistently:  > ' > File.getCanonicalPath() returns e.g:  4 > /DISK$USER1/MYLOGINDIR/APP/...../SHAREDIMGFILE.EXEK > (but Java APIDOC says this call should return a system dependent form of  ) > the path! which it does under other OS) 3 > but System.load(...) seems to expect it this way: 4 > DISK$USER1:[MYLOGINDIR.APP......]SHAREDIMGFILE.EXE6 > (in fact the HP JavaVM should map this name for me?)  3 Have you filed a formal bug report with HP on this?    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------    Date: 25 Mar 2005 09:05:26 -0800' From: "DL Phillips" <whohe@whoever.com> @ Subject: Re: How to load a shared img without defining a symbol?B Message-ID: <1111770326.568513.13990@o13g2000cwo.googlegroups.com>   Bernhard wrote:aC > First time install of the base application is not the problem! Atf thatG > time doing whatever sort of system operations poses no problem and wen  E > certainly have no restrictions here. This works already, even undera
 > OpenVMS. >   G Okay, so this is where you could do a define to make sys$share a search ? list. Not the best solution, but unless HP comes up with a realiG solution that's what will work. Phil mentioned this before in his post:    ##' dool...@snowy.net.au 	  Mar 23, 8:13 pm 
 Could you notpE $ define/sys/exec sys$share disk_user:[java_lib],sys$sysroot:[syslib]a% (usual security considerations apply)A .=2E.l Phil ##  E You could instead do this definition when you start your base app andtE use the the /JOB or /GROUP table instead of the /SYS if you need moreA than just /PROC.    E > The problem occurs during operation: We are dynamically loading neweE > "plugins", most of them are in pure Java, so there is no need to dol anyrE > OpenVMS related config. But some are using legacy C++ shares. These F > shares are of course compiled and linked target platform confirmant." > Such a Java plugin JAR contains:D > 1. the Java code to satisfy the interface requirements of our base appo@ > 2. a JNI wrapper to access the native share functions/methods.D > 3. a set of platform dependent shares: a DLL for Win, a lib.so for UnixB > and a SHR for OpenVMS. The base app automatically detects, which systemE > it is deployed on and extracts the native share to the file system.( >sG > We have the absolute requirement to load the plugins without stoppingl  D > the base and the rest of the plugins! So that's why making Defines givesa > us troubles. > G > Our base app is an implementation of OSGI. If you are interested, you-  ) > may have a look at http://www.osgi.org.- >-   OSGi, cool.5D So, within that model, your base application must also keep track ofD the loaded modules and interdependencies like which modules might beD sharing which libraries? So you know when a module or library can be& removed. You've provided this service:  D =B7  Configuration Admin Service  - This service provides a flexible; and dynamic model to set and get configuration information..  F so, (if you don't make sys$share a search-list) once your base app hasB determined that it must load the VMS SHR, it needs to execute code@ specific to the VMS platform. Within that code you could use theE $CRELNM and $DELLNM system services to set and remove the environment F for each of your client modules. You wouldn't need to stop or start or interfere with anything else.b  F Since the VMS specific library & module parts were probably created byF people who are used to the VMS way of doing things, it wouldn't be tooG much to ask for a definition file to be packaged along with the SHR, orlA even a standardized call to the SHR that returns the definitions.n  C Just rambling. Sounds like a fun project. Hope you find a solution.s    -Doug   ------------------------------  + Date: Fri, 25 Mar 2005 09:01:02 +0000 (UTC)oP From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)Y Subject: Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteraliasr$ Message-ID: <d20k0d$kdo$1@online.de>  6 In article <00A41388.EFD94FFE@SSRL.SLAC.STANFORD.EDU>,F winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) writes:   K > Not sure if this will work, but have you tried defining the one necessaryo) > numeric ip localaddress as 127.0.0.1?  u  J > If that doesn't work, you need a logical name in your OSU startup.  It'sI > probably most convenient to put it in your HTTP_STARTUP file, where the:M > directory to be WWW_SYSTEM gets defined, so you don't have to pick it apartEE > completely from WWW_SYSTEM.  Or you can have it sit somewhere else.   H Thanks; I'll give both suggestions a try.  It is not obvious how to run A the server in a cluster, using the cluster-alias address AND use hI localaddress simultaneously.  I'm surprised more people haven't reported   problems with this.o   ------------------------------  + Date: Fri, 25 Mar 2005 09:07:29 +0000 (UTC)gP From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)Y Subject: Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteraliasi$ Message-ID: <d20kcg$kdo$3@online.de>  9 In article <d1vlcb$ngm$1@news01.intel.com>, Ken Fairfieldh! <my.full.name@intel.com> writes:    + > So you can use the inverted test like so:n > ? >       $ Pipe Mcr Tcpip$IfConFig -a | Search Sys$Pipe failSAFE2 > / > and if that is _true_, you're _not_ the "CI".e  & OK; the test I made was with 5.3.  :-|   ------------------------------  % Date: Fri, 25 Mar 2005 06:12:56 -0500@' From: "Main, Kerry" <kerry.main@hp.com>0Y Subject: RE: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteralias8R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB5953A5@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Phillip Helbig---remove CLOTHES to reply=20s- > [mailto:helbig@astro.multiCLOTHESvax.de]=20a > Sent: March 25, 2005 4:01 AM > To: Info-VAX@Mvb.Saic.Com ? > Subject: Re: TCPIP 5.4: obsolete information in HELP, what=20o! > are new defaults (clusteralias)g >=208 > In article <00A41388.EFD94FFE@SSRL.SLAC.STANFORD.EDU>,H > winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) > writes:=20 >=20B > > Not sure if this will work, but have you tried defining the=20 > one necessaryS- > > numeric ip localaddress as 127.0.0.1? =206 >=20@ > > If that doesn't work, you need a logical name in your OSU=20 > startup.  It's> > > probably most convenient to put it in your HTTP_STARTUP=20 > file, where thedA > > directory to be WWW_SYSTEM gets defined, so you don't have=20  > to pick it apartG > > completely from WWW_SYSTEM.  Or you can have it sit somewhere else.r >=20A > Thanks; I'll give both suggestions a try.  It is not obvious=20  > how to run=20oE > the server in a cluster, using the cluster-alias address AND use=20e< > localaddress simultaneously.  I'm surprised more people=20 > haven't reported=20o > problems with this.e >=20   Phillip,  G This whitepaper on using TCPIP in HA environments may have been already, posted, but just in case:t? http://h71000.www7.hp.com/openvms/journal/v2/articles/tcpip.pdf    Regards   
 Kerry Main Senior Consultante HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477a kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Fri, 25 Mar 2005 10:16:25 -0800i, From: Ken Fairfield <my.full.name@intel.com>Y Subject: Re: TCPIP 5.4: obsolete information in HELP, what are new defaults (clusteralias-+ Message-ID: <d21khv$jrd$1@news01.intel.com>   / Phillip Helbig---remove CLOTHES to reply wrote:s; > In article <d1vlcb$ngm$1@news01.intel.com>, Ken Fairfieldh# > <my.full.name@intel.com> writes: l >  > + >>So you can use the inverted test like so:- >>? >>      $ Pipe Mcr Tcpip$IfConFig -a | Search Sys$Pipe failSAFE< >>/ >>and if that is _true_, you're _not_ the "CI".i >  > ( > OK; the test I made was with 5.3.  :-|  =      I'm not sure what to make of your response.  The subjectM< of this thread is "TCPIP 5.4 ...".  Now you're talking about TCPIP 5.3. :-(  >      I do see that 5.3 does NOT include the same output as 5.4; and is missing the "failSAFE" text, etc.  I also don't knowe= how many alias addresses you need to support in your cluster.p; For the simple case of one "cluster impersonator", it looksr; like just counting the number of addresses assigned to eachd; interface tells you whether that interface is servicing theR> alias or not (IP address count > 1).  But if there are several? of these, you'd be back to hard-coding the various IP addressesc and searching for them.R  8      Hey, can you just skip 5.3 and upgrade to 5.4?  :-|   	-Kenf -- p6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfieldt! D1C Automation VMS System Support>" who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------  + Date: Fri, 25 Mar 2005 09:02:19 +0000 (UTC)aP From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)> Subject: Re: TCPIP: why can an MX record not point to a CNAME?$ Message-ID: <d20k2r$kdo$2@online.de>  5 In article <42427046.C5EED7F7@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes: m  1 > Phillip Helbig---remove CLOTHES to reply wrote:kL > > OK, so a temporary DNS problem could lead to a bounce, whereas (assumingJ > > the backup MX could be resolved) having a backup-MX server would avoid	 > > this.h > J > Nop. Your backup MX entry will be resolved by the same DNS server in the > same transaction.I > D > If the DNS query fails, then whether you have one or 10 MX entries > doesn't make a difference. > I > If then DNS query succeeds, then the SMTP server will know about your 1nH > or 10 tarhet for emails and will sequentially try all of them until itN > can succesfully connect, and if it still fails, it will requeue the message.  3 Yes, that occurred to me after I wrote my response.P  B So, again, what's the point of a backup-MX, as long as one is not / offline for longer than the maximum retry time?t   ------------------------------   End of INFO-VAX 2005.168 ************************