1 INFO-VAX	Tue, 02 Nov 2004	Volume 2004 : Issue 608       Contents: Re: 1000a boot failure Re: 1000a boot failure+ ComputerWorld: OpenVMS Survives and Thrives / Re: ComputerWorld: OpenVMS Survives and Thrives + Re: MSL5026 Tape lib with OpenVMS & Windows  OpenVMS thriving! & Re: OT: New Vulnerabilities in Solaris( Re: Samba 2.2.8 for OpenVMS, new release( Re: Samba 2.2.8 for OpenVMS, new release( Re: Samba 2.2.8 for OpenVMS, new release( Re: Samba 2.2.8 for OpenVMS, new release( Re: Samba 2.2.8 for OpenVMS, new release Sr VMS administrator- PA Re: Suggestion for TPU: revert" Re: TCPIP NTP process NOT resident" Re: TCPIP NTP process NOT resident Re: Time Change  Re: Time Change  Re: Time Change  Re: VMS for Bush ? Re: VMS for Bush ?8 [OT]: No publication tests and writes about VMS this way  F ----------------------------------------------------------------------  $ Date: Mon, 1 Nov 2004 22:05:02 +0100, From: "Hans Vlems" <hvlems.dotweg@zonnet.nl> Subject: Re: 1000a boot failure + Message-ID: <2unmn5F2b0cjfU1@uni-berlin.de>   B "Ransom Fitch earthlnk" <rlf_vms@earthlink.net> schreef in bericht= news:l3uhd.14769$KJ6.4816@newsread1.news.pas.earthlink.net... 3 > I have a DEC Alpha 1000a that is failing to boot. = > Without being too tedious the various startup messages are:  >  > -------------------------  > probing hose 0, PCI " > probing PCI-ti-EISA bridge, bus1. > EISA Data in non-volatile storage is corrupt >  > EISA Configuration Error$ > Run the EISA Configuration Utility >  > ...  > no other errors reported for > bus, memory, or disks  >  > last message > ...  > " > Change to Normal Operating Mode. > >>>  > ------------------ >  > > > At this point the system does not respond to keyboard input.? > I have a ECU diskette, but have no idea how to run it without  > the keyboard.  >  > D > 2nd question: Is it possible to make a backup of the ECU diskette?@ > If so how? This seems to be a very critical piece of software! >  > 	 > Thanks,  > Ransom Fitch > rlf_vms  at earthlink dot net  > ! > apologies if this is a 2nd post  >  > H Did you try and connect a terminal to its serial ports (DB9 connector) ?C May be the system console is not set to graphics but set to serial.   > About the ECU diskette, no idea actually just two suggestions:& - diskcopy on a Windows or DOS systems5 - use the dd command available on unix/linus systems.    Hans   ------------------------------  $ Date: Mon, 1 Nov 2004 20:13:56 -0600* From: "ssj152" <ssj152 AT charter DOT net> Subject: Re: 1000a boot failure 0 Message-ID: <10odrdhsvgu4m90@corp.supernews.com>  @ "Ransom Fitch earthlnk" <rlf_vms@earthlink.net> wrote in message= news:l3uhd.14769$KJ6.4816@newsread1.news.pas.earthlink.net... 3 > I have a DEC Alpha 1000a that is failing to boot. = > Without being too tedious the various startup messages are:  >  > -------------------------  > probing hose 0, PCI " > probing PCI-ti-EISA bridge, bus1. > EISA Data in non-volatile storage is corrupt >SNIP< > D > 2nd question: Is it possible to make a backup of the ECU diskette?@ > If so how? This seems to be a very critical piece of software! >  > 	 > Thanks,  > Ransom Fitch > rlf_vms  at earthlink dot net  > ! > apologies if this is a 2nd post  >  >    Ransom,   L Diskcopy under windows should work fine for making a backup of the ECU disk.L If all else fails when (if) you lose yours, post online & someone might makeA you a copy. They are not frequently downloadable because they are @ copyrighted software, furnished by the disk controller vendor to DEC/Compaq/HP.   Stuart   ------------------------------  # Date: Mon, 01 Nov 2004 20:10:24 GMT 6 From: "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com>4 Subject: ComputerWorld: OpenVMS Survives and Thrives= Message-ID: <QKwhd.37972$Jb.1080502@twister.southeast.rr.com>   + ComputerWorld: OpenVMS Survives and Thrives  News Story by Drew Robb P http://www.computerworld.com/softwaretopics/software/story/0,10801,97032,00.html   :)  J We as a community owe Drew Robb a collective thanks.  He's gotten several L articles about VMS in the trade press over the course of 2004.  If everyone ' has a minute please drop him a message.    drewrobb (AT) sbcglobal.net   I Here's a list of Editorial Contacts at ComputerWorld.  Let them know you  ! appreciate them covering OpenVMS.   7 http://www.computerworld.com/aboutus/contacts?type=edit      Ken    OpenVMS.org % _____________________________________  Kenneth R. Farmer <>< % SpyderByte: http://www.SpyderByte.com    ------------------------------  $ Date: Mon, 1 Nov 2004 15:49:46 -0500* From: "Bill Todd" <billtodd@metrocast.net>8 Subject: Re: ComputerWorld: OpenVMS Survives and Thrives= Message-ID: <_JqdncHihYD6PBvcRVn-jg@metrocastcablevision.com>   A "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com> wrote in message 7 news:QKwhd.37972$Jb.1080502@twister.southeast.rr.com... - > ComputerWorld: OpenVMS Survives and Thrives  > News Story by Drew Robb  > L http://www.computerworld.com/softwaretopics/software/story/0,10801,97032,00. html  K Gee, VMS revenues now just 'over $2 billion', including services?  Seems to K suggest the possibility of a bit of a drop from Keith's "$2.5 - $3 billion" D a while ago - and a very definite drop from the $4 billion figure of
 2000/2001.  H And since about half of that $4 billion 3 - 4 years ago was in services,L which tend to be far more resistant to sudden drops than sales revenues are,J one might still suspect that VMS sales were pretty much in the toilet.  OfK cousre, examination of HP's profit figures could also help lead one to that  conclusion.   K But it's reassuring, in a perverse sort of way, to see those good old 'over I 400,000 systems' and '10 million users world-wide' figures still floating J around:  they'll probably be quoting them long after VMS is dead and gone,J without noting the date that they were generated.  And it seems that TerryG is touting another VMS 'revival', which the author has accepted without 9 question as fact:  can we stand the excitement of it all?   I It was a nice article, errors and omissions excepted.  I wish HP deserved  the support.   - bill   ------------------------------  $ Date: Mon, 1 Nov 2004 18:44:16 -0600 From: <ksrich@bellsouth.net>4 Subject: Re: MSL5026 Tape lib with OpenVMS & Windows8 Message-ID: <pJAhd.9170$Om6.4986@bignews5.bellsouth.net>  . "Dan Moore" <dmoore@sosu.edu> wrote in message8 news:rcthd.81$_l2.35@newsfe10.lga.highwinds-media.com... > Greetings! >  >  > H > We have a MSL5026 Tape library we would like to share with our windows	 > server:  > I >   a.. Drive 0 of the library is direct attached SCSI to the VMS system. L >   b.. Drive 1 (and robot controller) is direct SCSI to the windows system. >   c.. Drive 2,3 empty.> >   d.. We are close to starting our first SAN implementation. >   e.. SDLT 320 Drives L > The VMS systems have used the library in "SEQENTIAL SPLIT" mode just fine.H > The drive acts like a loader, so mount/dismount loads the next tape in line > in a predictable fashion.  >  >  > D > Our dilemma is the MODE setting on the library. The windows backup	 software, K > Backup Exec, requires the mode to be "RANDOM". We switched to random mode J > and discovered that VMS mount/dismounts commands will no longer load the > next tape in line on Drive 0.  >  >  > G > I'm reluctant to use special client/server software to backup our VMS K > systems. VMS Backup has worked well for years; however, we have plenty of L > capacity on this library. Currently, the VMS backups fit on one tape usingK > compaction. The windows backups require about 4 tapes for a full, and one  > tape for daily incrementals. >  >  > K > Any ideas on how to make this work? Solutions that use SAN components are 
 > welcome. >  >  >   G You can use MRU which should have come with the Library.  Add the robot  commandsK in the backup command procedure to load and unload the correct tape.  I use  it all the time.     Shael    ------------------------------   Date: 1 Nov 2004 18:10:49 -0800 ( From: bob@instantwhip.com (Bob Ceculski) Subject: OpenVMS thriving!< Message-ID: <d7791aa1.0411011810.bc9b1bb@posting.google.com>   take that Andrew! :)  P http://www.computerworld.com/softwaretopics/software/story/0,10801,97032,00.html   ------------------------------   Date: 1 Nov 2004 13:09:28 -0600 4 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)/ Subject: Re: OT: New Vulnerabilities in Solaris 3 Message-ID: <ZJyJqpMHuvhb@eisner.encompasserve.org>   h In article <d7791aa1.0411011018.1fdf9203@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes:< > Are you a member of the patch of the week club?  Join now!; > Just port your platform to windoze, linux or slowaris and , > you are automatically registered for FREE!  K Mnay years ago, at a previous place of employment, we had VMS and Tru64 and L HPUX side by side by side. The guy who sat in the cube in front of me seemedJ to spend ALL of his time installing HP-UX patches. Every weekend he was inI the office installing a bunch of patches. More than I installed on my VMS J system in 4 years. He used to say that HP stood for Hourly Patches. I used; to say that the purpose of HP was to make DEC look good :-(   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  1 	26-October, 2001: A day that will live in infamy 4 	Support Freedom: http://www.indefenseoffreedom.org/  ? 	Voting for "the lesser of two evils" is still voting for evil.    ------------------------------  % Date: Mon, 01 Nov 2004 21:59:06 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>1 Subject: Re: Samba 2.2.8 for OpenVMS, new release + Message-ID: <41870609.222DA0CF@comcast.net>    "John E. Malmberg" wrote:  >  > David J Dachtera wrote:  > > "John E. Malmberg" wrote:  > <snip>K > > Well, I tend to take my example from Hunter Goatley's InfoZip kits: the J > > kit contains object libraries and builds the .EXE on the target systemL > > (sorry, but sites with C compilers remain the exception in VMSland where* > > in UN*Xland they tend to be the rule). > J > Any way that it is done it is likely to require a number of object kits.I >   The object kit must be compiled on the lowest version of OpenVMS that ' > the resulting executable will run on.   E Please elucidate. How does the o.s. version effect compilation? (Lest A anyone forget, the process many people think of as "compiling" is E actually a two-stage process: Source code gets "translated" to object 9 code. Object code, while architecture-specific, should be D o.s.-version-agnostic. Language RTL elements may be language-versionH dependent, but should be o.s.-version-agnostic for the most part, exceptG where system-service dependencies exist. Object code then gets "linked" D (IBM I think still calls it the "Linkage Editor") into an executableA image: all symbolic references are resolved to either absolute or  relative addresses, and so on.)   H Linking, I can understand. Various elements, including the system image,G system symbol table and such will be version-dependent. Again, language = RTL elements may be language-version dependent, but should be D o.s.-version-agnostic for the most part, except where system-serviceC dependencies exist. In such case, the absence or presence of system < services and/or features in the system RTL will be an issue.  H > There have been significant improvements in OpenVMS compatibility withA > UNIX calls in the various 7.x and 8.x releases, and some of the 1 > improvements have been delivered with ECO kits.   E Have the C (and other langauge compilers) acquired such dependencies? 4 ... or is upward-compatibility no longer maintained?  G > It will take a bit of work to determine the optimum SAMBA for OpenVMS 5 > binary that would match a specific OpenVMS release.   H Optimum is probably not an issue. The "optimum" version is any that willH run on a specific version/architecture. Performance, etc. issues will beH on a "can do" basis, within the limits of what the operating environment	 provides.   J > While in some cases you can link on an older version of OpenVMS than theG > objects were compiled on and get away with it, the rule is that it is  > unsupported.    ) How so? ...and when did this come about?     References:   E Process Software Multinet is distributed in object form and linked on G the target system; it usually must be re-installed (no relink procedure G is provided) if the host o.s. is upgraded, but this is sometimes not an  absolute requirement.   F All-in-1 was a bit sensitive to the host o.s. version, usually needingE relinking after an upgrade (a procedure was provided by the install).   < > And it would take a bit of work to determine what cases it3 > could be supported on prior versions of OpenVMS,.   @ I should think any version that supports the required C-RTL, no?  K > >>As for SAMBA3/4 anything can happen.  My guess is that the target would > > >>be a PCSI installation for Alpha/I64 OpenVMS distribution. > > I > > YECCCCHHH!!! Only as a dead-last resort goto .PCSI, unless there is a K > > valid reason with a great deal of weight behind it. Very messy, and you I > > tend to be at the mercy of bugs in PCSI (much harder to dissect a kit I > > and over-ride an installation programming error than with VMSINSTAL), % > > errors in the PCSI database, etc.  > I > HP has produced in their roadmaps that they are investigating SAMBA for , > use as a product on the OpenVMS platform.    Bravo!!! Excellent move!!!  < Now, if only they understood marketing... (*HEARTFELT SIGH*)   > As an HP employee in theG > OpenVMS group, I have a small part in that investigation.  While I do J > not control the result, I would expect that HP would use their supportedE > distribution method if they decide to pursue a SAMBA based product.  > ) > You can see that I have a bias on this.   C Even so, what (HP) will support and what the open-source world will   tolerate may be quite disparate.  % For (HP) Samba, .PCSI may make sense.   1 For open-source Samba, it serves no good purpose.   H > > What I did was take Jean-Yves's kit and amssage so that it links theK > > .EXE on the target system directly from the .OLBs in the kit on CD-ROM, ; > > and does *NOT* install the object or source by default.  > A > As stated before, it is likely that many kits will be needed as H > different binaries will be needed for different VMS versions, if it isG > intended to provide binaries that will run on anything other than the  > current version of OpenVMS.   F "Current", in that context, is undefined, or defined by context as theH version on which the required C-RTL elements can be installed/supported.  J > With each release of VMS, better UNIX compatibility has been implemented@ > so object code built older releases may not provide as good ofF > functionality for later releases.  It would take a bit of testing to > determine the optimum result.   F Again, let's keep the cart behind the horse. Functional is by FAR moreF important than optimization. Any specific version of Samba may performH better in some environments than in others. So be it. *THAT* it performsF is, at this point, *FAR* more important than how well (within reason -! let's not get pedantic, people!).   L > > The thinking is that we need to get more acceptance of freeware and openG > > source, and a good way to do that is to make it as user-friendly as L > > possible without destroying its (forgive me) "geek" appeal. Not everyoneH > > who has need of it will be nuts-and-bolts coders - we need to enable$ > > usage without requiring hacking. > I > It would seem that making the installation of freeware to work the same E > as products supplied by HP would be the best way to meet that goal.  > I > It may not be as you view the optimal way to install a product, but you % > can see the logic behind that idea.   ; Optimal? Depends. Actually the "Install = PCSI, Configure = F post-install" paradigm is quite appropriate to Samba, although certain> assumptions must be made at install time (like does the personH installing understand that some decisions will be "cast in stone" by the= install, and cannot be changed later at configuration time?).   F > >>There is no shortage of work to be done on either new versions, orG > >>refinements to the old, including assistance on documentation, code # > >>review or all sorts of testing.  > >> > >>Any skill level can help.  > > @ > > I'd be willing to donate my repackaging code along with someG > > documentation (mostly caveats). Just tell me how/where. I'm already  > > subscribed to the list.  > I > Introduce your self to the mailing list, and tell them what you want to & > do, and see if you get any feedback.   Will do.  J > With open source, who ever controls the web/ftp space for a distributionH > controls the rules of what goes in it.  They can take suggestions, butD > they have the final control.  It is up to the users to decide what' > distribution that they choose to use.   @ Understood from the outset. I'm just here to offer alternatives.   > [snip]H > Now your input is valuable because you are focused on different issues > than I am.  D Yes - I'm looking to get a usable item onto the target system with aB minimum of "rocket science", and without installing a lot of filesH (read: source) that the user will never even so much as look at, in manyH cases. To that end, I intentionally choose to remain ignorant of certainD issues in the belief that "if a user needs to understand this, we'reG doing something that would be better done another, more automated way". F This is also in keeping with my philosophy that "the machine works for me, not the other way around".  B > I am focused mainly on build and testing issues.  To my mindset,= > packaging is part of the build, but one of the end phases.     Well, yes and no.   H To me, the build should do exactly that: Build the most elemental binary> pieces possible (read: object code). The package, not the hostF environment, must provide any dependencies not guaranteed to be served4 by the default installation of the host environment.   > First you B > need to be able to convert the source code into usable binaries.  B The result should be an architecture-specific object package to beB linked in the target environment. Again, the package, not the hostF environment, must provide any dependencies not guaranteed to be served4 by the default installation of the host environment.  J > If you do not have the ability to build the source into a usable binary,  > the packaging is meaningless.   G The folks at DEC and Process Software might (have) argue(d) that point. G Unless a C-to-Macro/32 translator is included with the package (to this @ day, *EVERY* (Open)VMS system ships with a Macro/32 assembler orF compiler; to my knowledge, there has never been a default installationF that provided a C compiler), the C-compiler requirement is, in effect,A causing the product to commit hara-kiri simply by distributing it D without at least object code: the presence of a C compiler cannot beA guaranteed. (Perhaps this a suggests a possible cause for Samba's , seemingly low acceptabnce in the VMS world.)  H That said, if you're aware of any movement to distribute C/C++ compilersD with OpenVMS by default, even if you cannot divulge, that would help= bring VMS closer in-line with the UN*X/open-systems paradigm.   ) > But that is just my opinion, so until I I > am able to produce a SAMBA 4.x binary, I am not going to concern myself ' > on how to produce a distribution kit.   ? Build the object libraries - leave the rest to the likes of me.    > [snip]H > As for now, I am only going to distribute source code on EncompasserveI > because I do not have the means to do full quality testing on what I am F > posting, so anyone that uses it needs to have a least the ability toL > compile it, and hopefully can also diagnose problems in their environment.  D Well, one would assume that source code released into the wild wouldE have a fair chance to compile clean, so long as compiler versions and : such are clearly stated in the accompanying README's, etc.  F If you need someone to produce object libraries, I might be willing to take that on, provided:   E o No requirement for any elements of DECset (CMS, MMS, LSE, etc.), or E other build environment elements that are not provided by the default  OpenVMS installation options.   E o Clearly state the C version (minimum required) for Alpha and VAX (I G can probably use the HP testdrives for I64, for now). The C version may F require a minimum OpenVMS version; however, the software itself should2 be agnostic of such outside of C version features.   > [snip]E > Right now, as far as I can determine from postings in the SAMBA-VMS B > mailing list or in comp.os.vms, there are only a couple of dozen > SAMBA-VMS sites.    That's what I'm trying to "fix".  H > You may find the DOCBOOK issues to also be something that fits in withG > your talents.  The products to implement it appear to be available on F > OpenVMS, but setting them up for use by people who do not want to beG > open source document processing experts does not appear to be trivial  > the way they distributed now.   G Well, o.k., but understand that "DOCBOOK for Dummies" will be a project , unto istelf. For example, what is "DOCBOOK"?   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Mon, 01 Nov 2004 23:10:41 -0500 - From: "John E. Malmberg" <wb8tyw@qsl.network> 1 Subject: Re: Samba 2.2.8 for OpenVMS, new release 1 Message-ID: <HKGdnZxIU9pclRrcRVn-qQ@adelphia.com>    JF Mezei wrote:  > "John E. Malmberg" wrote:  > J >>Any way that it is done it is likely to require a number of object kits.I >>  The object kit must be compiled on the lowest version of OpenVMS that ' >>the resulting executable will run on.  >  > $ > Can you please elaborate on this ? > M > If you are using standard calls whose arguments have not changed over time, 7 > why would version of VMS at time of compile matter ?    F Because the compiler may generate different code based on reading the F external definition files on version it is running on.  Many of these 2 files are tailored to the specific version of VMS.  N > If you have source code that make use of system/rtl calls that did not existN > in previous version of VMS, then it doesn't matter if you distribute source,5 > object or executable binaries since none will work.  > O > But if all the routines existed in previous versions, why would an object not ( > work when linked against any version ?  H The arguments in the routines may have changed in the newer version and > the compiler knows to take advantage of that for optimization.  D On an optimizing compiler when you call a standard library routine, B sometimes the compiler will replace that call with more efficient E routine calls.  Some of those routines may not be available on older  H versions.  Many issues are possible as it has not been tested, so it is I impossible to give a generic answer of what you can get away with or not.   I It is possible to make it work, but it is not something that HP wants to   support or to recommend.  + Hoff has posted previously on this subject.    > I >>HP has produced in their roadmaps that they are investigating SAMBA for + >>use as a product on the OpenVMS platform.  >   N > Please, you either go in full blast, or you stay away from it.  Because openM > source tends to be more fluid with changes coming in faster, sometimes very L > important changes, one cannot afford to have the owner of VMS turn an openP > source product into a proprietary one where VMS customers have no way to applyT > updates/patches and become dependant on the owner of VMS providing timely patches.  F SAMBA is GPL licensed, so the source must be provided.  Currently the D only support for SAMBA on VMS is to post questions on the SAMBA-VMS 5 mailing list or here and hope that they are answered.   I The HP-UX distribution of SAMBA is a special one that the the HP-UX team  H has done qualifications on.  Installations can use that and get support B from HP, or they build one of the SAMBA distributions and support  themselves.   N > The owner of VMS should take open source stuff and keep it open source. ThisP > means that when/if the owner of VMS cuts development funds, then VMS customers@ > can still move the product ahead and keep up with the Jones'.   - The SAMBA product would still be open source.   4 > Consider Motif which has fallen way behind on VMS.  I I was not aware that Motif is an Open Source product.  And in this case,  C my focus is on SAMBA.  I am not in a position to address the other   issues that you bring up.    [snip] > A >>As stated before, it is likely that many kits will be needed as H >>different binaries will be needed for different VMS versions, if it isG >>intended to provide binaries that will run on anything other than the  >>current version of OpenVMS.  > P > You know very well that there wouldn't be funding to create kits for differentO > versions of VMS. Heck, you don't even have funding to release one version for  > VAX anymore.  I Currently SAMBA is only available as provided by volunteers that are not  H getting any funding that I am aware of.  As to what HP does or does not D support, that will depend on the results of the investigation being 
 conducted.   > I >>It would seem that making the installation of freeware to work the same E >>as products supplied by HP would be the best way to meet that goal.  >   H > No. There is one important consideration:while you can generally trustH > proprietary software and trust that the vendor has a certain amount ofN > responability and quality assurance, that is not the case with freeware. AndL > many companies won't accept freeware unless they also have the source codeN > they can analyse, and more importantly, when bugs are deiscovered, they wantP > the ability to apply them and not have to rely on some 3rd party company whichT > many take years to release a new proprioetary version that iumplements that patch.  G If someone is not going to accept freeware, then the distribution does  G not matter.  But if they are, there are some that do not want to use a  F different installation method, and different directory tree structure - for every freeware product that they install.   H What is involved in the productization of an Open Source project is the / testing and certifying a specific distribution.   H The available ability of alternate distributions depends totally on the A the what independent developers will take the time to work on it.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------  % Date: Tue, 02 Nov 2004 00:25:28 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 1 Subject: Re: Samba 2.2.8 for OpenVMS, new release , Message-ID: <41871A2F.DB95C2A5@teksavvy.com>   David J Dachtera wrote: G > Have the C (and other langauge compilers) acquired such dependencies? 6 > ... or is upward-compatibility no longer maintained?  F Where problems arise is when you have software written for more recentN versions of the "system" and when you try to link it against an older version,M the linker does not find some of the routines that were used in the software.   J For instance, Motif has a routine used to tell the main event loop to stopN looping. This routine does not exist on VMS because VMS has a very old versionN of Motif. So when you try to link such software, you find out XtAppSetExitFlagC doesn't exist on VMS, so you have to write your own and not use the 2 XtAddMainLoop and write your own (simple routine).  N Since VMS is moving to add more Unix compatibility, those who port software toL VMS will need to make fewer and fewer adaptations because VMS will have moreJ and more of the routines present on Unix but trditionally not available on VMS.    M However, as you pointed out, I still do not understand why objects code in an + by itself would be so OS version dependant.    ------------------------------  % Date: Tue, 02 Nov 2004 00:30:52 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 1 Subject: Re: Samba 2.2.8 for OpenVMS, new release , Message-ID: <41871B72.65C3C50F@teksavvy.com>   David J Dachtera wrote: H > All-in-1 was a bit sensitive to the host o.s. version, usually needingG > relinking after an upgrade (a procedure was provided by the install).   ! There were many reasons for this:   L one was the customization capabvilities of ALL-In-1 that allows customers toK add their own code to the A1 main image (new functions etc), as well as add J interfaces to packages such as datatrieve which were linked into All-in-1.  M Also, at installation time, you have the opportunity to choose which language J variant, and you could change that later on (for instance, real english or( american english spelling dictionaries).   ------------------------------  % Date: Tue, 02 Nov 2004 01:15:20 -0500 - From: "John E. Malmberg" <wb8tyw@qsl.network> 1 Subject: Re: Samba 2.2.8 for OpenVMS, new release 1 Message-ID: <zd2dnQiZ4YFluBrcRVn-jA@adelphia.com>    David J Dachtera wrote:  > "John E. Malmberg" wrote:   J >>Any way that it is done it is likely to require a number of object kits.I >>  The object kit must be compiled on the lowest version of OpenVMS that ' >>the resulting executable will run on.  >   A > Please elucidate. How does the o.s. version effect compilation?   A The compiler uses information obtained from the operating system  : supplied data files to determine how to compile a program.  F With the C compiler, these are known as the header files, and are not = actually supplied as individual files, but in a text library.   E Other compilers have their own definition libraries, including Macro.   I Sometimes a new entry point is introduced for improved operation, or new  A parameters are activated.  These can be hidden in the definition  I libraries.  Old objects would not specify these codes, and when linked a   default behavior would occur.   E New objects will may set extra flags which may not be handled by the  ? older shared images.  And this may not be visible to a program.   H These are all hypothetical issues, as such configurations have not been H tested, there is no way to predict what will work and what will not, or   how visible such issues will be.  H >>There have been significant improvements in OpenVMS compatibility withA >>UNIX calls in the various 7.x and 8.x releases, and some of the 1 >>improvements have been delivered with ECO kits.  > G > Have the C (and other langauge compilers) acquired such dependencies? 6 > ... or is upward-compatibility no longer maintained?  D Upward compatible means that a non-privileged object compiled on an I older version of OpenVMS can link and run on a newer one.  And generally  / a best effort is attempted for privileged code.   G It does not mean that code compiled on a newer version of OpenVMS will  B run correctly or predictably on an older version.  As that is not : tested, there is no listing of what will or will not work.   > G >>It will take a bit of work to determine the optimum SAMBA for OpenVMS 5 >>binary that would match a specific OpenVMS release.  >   J > Optimum is probably not an issue. The "optimum" version is any that willJ > run on a specific version/architecture. Performance, etc. issues will beJ > on a "can do" basis, within the limits of what the operating environment > provides.   B Would you want to limit the performance of a program on a current G version of VMS so that it can share the same binary as a older version?    > J >>While in some cases you can link on an older version of OpenVMS than theG >>objects were compiled on and get away with it, the rule is that it is  >>unsupported.   >   + > How so? ...and when did this come about?     It has always been the case.   > References:   I Hoff has posted about this many times before.  In one case before I took  H my current job, I was proposing to compile on current versions and link  on older versions.  I It is a matter of risk management.  Compiling on older, linking on newer  & is supported for upward compatibility.  G Compiling on newer, linking on older is not supported and has not been N tested.o  G > Process Software Multinet is distributed in object form and linked onoI > the target system; it usually must be re-installed (no relink procedure:I > is provided) if the host o.s. is upgraded, but this is sometimes not an  > absolute requirement.   E I have no idea what Process Software is doing for their build, but I aH would assume that they do quality testing on all their supported target D versions.  In this case Process Software is the one responsible for $ making sure that what they do works.  H > All-in-1 was a bit sensitive to the host o.s. version, usually needingG > relinking after an upgrade (a procedure was provided by the install).a  5 Again, compiling on old, linking on new is supported.u  < >>And it would take a bit of work to determine what cases it3 >>could be supported on prior versions of OpenVMS,.- > B > I should think any version that supports the required C-RTL, no?  B The DEC/COMPAQ/HP CRTL and definition files are supplied with the I operating system since the 6.X release, and the shared image is specific v to that operating system.n  A Some updates have been distributed with the compiler in the past..   <snip>  E > Even so, what (HP) will support and what the open-source world willu" > tolerate may be quite disparate. > ' > For (HP) Samba, .PCSI may make sense.  > 3 > For open-source Samba, it serves no good purpose.D  E There are two methods that are dominant in the UNIX world that I can  F detect.  One is a tarball and a configure script, the other is to use  the Red Hat Package manager.  H I doubt that many will find the tarball method to be user friendly, and H I have not yet been able to get the Red Hat Package manager to build on G OpenVMS.  It requires a large number of components, including Berkeley r; DB and other scripting tools.  Python may even be involved.i  H Even if the Red Hat Package Manger was available, it effectively is the G same functionality as PCSI with a user interface that only LINUX users h would recognize.   >   H > "Current", in that context, is undefined, or defined by context as theJ > version on which the required C-RTL elements can be installed/supported.  A Which are supplied with the Operating system or ECO kit, not the iG compiler, except in the cases of older releases of OpenVMS.  The DEC C rI RTL is not supplied with OpenVMS 5.5-2.  There is a special kit and ECOs s	 for that.-  J >>With each release of VMS, better UNIX compatibility has been implemented@ >>so object code built older releases may not provide as good ofF >>functionality for later releases.  It would take a bit of testing to >>determine the optimum result.n >  aH > Again, let's keep the cart behind the horse. Functional is by FAR moreH > important than optimization. Any specific version of Samba may performJ > better in some environments than in others. So be it. *THAT* it performsH > is, at this point, *FAR* more important than how well (within reason -# > let's not get pedantic, people!).   I In the case of SAMBA, speed and scalability is a major usability factor. fF   Anything that can be done to speed up the file server on OpenVMS is E needed.  And some of that speed up may depend on version of the host g( operating system that it is targeted to.  C Also anything that get improve the number of sessions that a given t& platform can handle is also important.  G Right now, a lot of SAMBA functionality is disabled on the VMS port of t the 2.2.x stream.d  G The SAMBA 2.2.x release in the UNIX world is going end of life.  There l* may not be any more releases after 2.2.12.  H The current SAMBA UNIX release is 3.0.7, with 3.0.8 betas available.  A 3 code drop for the 4.x development may show up soon.j   <snip> >  r >>[snip]H >>As for now, I am only going to distribute source code on EncompasserveI >>because I do not have the means to do full quality testing on what I am F >>posting, so anyone that uses it needs to have a least the ability toL >>compile it, and hopefully can also diagnose problems in their environment. >   F > Well, one would assume that source code released into the wild wouldG > have a fair chance to compile clean, so long as compiler versions andr< > such are clearly stated in the accompanying README's, etc. > H > If you need someone to produce object libraries, I might be willing to > take that on, provided:d > G > o No requirement for any elements of DECset (CMS, MMS, LSE, etc.), or_G > other build environment elements that are not provided by the defaulte > OpenVMS installation options.-  E MMK is available free even for non-hobby users.  It will process the lD same files.  GNV is available in binary form for ALPHA.  HP has not " created a VAX binary distribution.  : Future builds may end up using GNU make for compatibility.  A There are other tools that are needed to make it practical to be 0F building OpenVMS test versions of open source projects that are up to J date with the mainstream UNIX versions.  RSYNC and SUBVERSION for example.  E RSYNC is a remote synchronization product.  It basically is a client  F server application where the client keeps a directory tree up to date G with a server, and only transfers the changed components of a file.  I  C do not have it working well enough on VMS to be of general use yet.   H SUBVERSION is a replacement for CVS.  I have not even looked at a port, 6 but have discovered that Netbeans has a plugin for it.  G > o Clearly state the C version (minimum required) for Alpha and VAX (IcI > can probably use the HP testdrives for I64, for now). The C version may H > require a minimum OpenVMS version; however, the software itself should4 > be agnostic of such outside of C version features.  H Until OpenVMS 8.2 ships, the minimum C version is still a moving target.J For Alpha 6.2-002 is what I am using.  My VAXen are currently powered off.  H >>You may find the DOCBOOK issues to also be something that fits in withG >>your talents.  The products to implement it appear to be available onaF >>OpenVMS, but setting them up for use by people who do not want to beG >>open source document processing experts does not appear to be trivial  >>the way they distributed now.- >   I > Well, o.k., but understand that "DOCBOOK for Dummies" will be a projectc. > unto istelf. For example, what is "DOCBOOK"?  = Use GOOGLE to quickly find out as much as I know or more. :-)M  G As near as I can tell it is a set of preset templates and instructions .= on how to use them that is used to feed document translators.S  G It is built on SGML and XML which are implemented in TeX and LaTeX and   other open source packages.y  F There seems to be a wide varieties of ways to implement it, including C editing the source which will look quite a bit like HTML, and then mJ compiling the result, to tools that will do GUI WYSIWYG type environments.  F Most of the explanations that I have found is that you run the source G document through several build procedures that vary depending on which   outputs that you want.   -Johnn wb8tyw@qsl.net Personal Opinion Only    ------------------------------   Date: 1 Nov 2004 11:05:56 -0800o> From: bruce.x.mackenzie@questdiagnostics.com (Bruce MacKenzie)! Subject: Sr VMS administrator- PAv= Message-ID: <f1a48fa9.0411011105.458e9b19@posting.google.com>   E For any member who may know of someone interested, I am looking for aiF strong senior VMS administrator for our data center here in the Valley8 Forge PA area.  Not contract, and I'm not a third party.   Responsibilities include:l@ Design, manage and maintain the system and applications softwareC environment across a range of VMS clusters and standalone systems. y? Install and configure operating system and layered products and F applications software as required.  Management: must be able to manage@ multiple projects simultaneously and adhere to strict deadlines.B Problem Management: must be able to resolve complex system/network! problems within tight timeframes.m  C Looking for minimum of 7 years experience in production VMS cluster:C and DSM DB application environments. Knowledge of high availabilityV@ VMS systems management. Knowledge of DSM (MUMPS) and Cache based( laboratory systems setup and management.  B This is a f/t position and we are offering an attractive salary as; well as excellent benefits.  Write or call if you need more: information. www.questdiagnostics.com  	 Thank you   
 Best Regards,j   Bruce MacKenziel Quest Diagnosticsa 400 Egypt Road West Norriton, PA 19403c& bruce.x.mackenzie@questdiagnostics.com 610-650-2527   ------------------------------  % Date: Mon, 01 Nov 2004 22:05:53 -0600r2 From: David J Dachtera <djesys.nospam@comcast.net>' Subject: Re: Suggestion for TPU: reverte+ Message-ID: <418707A1.6418E117@comcast.net>a   "Keith A. Lewis" wrote:  >  > JF Mezei <jfmezei.spamnot@teksavvy.com> writes in article <41803BD6.7482EE31@teksavvy.com> dated Wed, 27 Oct 2004 20:22:55 -0400:rO > >In the event someone is still collecting suggestions for improvements on VMSdM > >applications, adding a  REVERT command (and in the decwindows interfacem am, > >FILE->REVERT menu) would be very welcome. > I > Just a guess, but my take on why it isn't there is because with version M > numbers it could be ambiguous.  When a user types <DO>revert, does he mean:  > F > 1.  Load the same version which was used when the buffer was created: > 2.  Load the last version written with the WRITE commandD > 3.  Load the most recent version even if it's from another process  8 Simple solution: Prompt (ask the question) for it. Next?   -- i David J Dachtera dba DJE Systemso http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:-" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Mon, 01 Nov 2004 14:26:21 -06002( From: brandon@dalsemi.com (John Brandon)+ Subject: Re: TCPIP NTP process NOT resident 1 Message-ID: <04110114262168@dscis6-0.dalsemi.com>e   Mike Rechtman wrote:F > IIRC SYS$STARTUP:TCPIP$NTP_STARTUP checks for a logical name in the  > TCPIP$STARTUP_TABLE.J > If for some reason a previous NTP startup (or shutdown) bombed and left H > the logical name wrongly defined, you would see the "already started" 
 > message.   This makes sense.y  L However NTP seems to be shutting itself down - I can not find any stack dumpJ foot prints.  I do find that NTP log files are being generated frequently:  < TCPIP$NTP_RUN.LOG;2354     3/18       1-NOV-2004 08:02:16.63< TCPIP$NTP_RUN.LOG;2353     1/18       1-NOV-2004 08:02:16.30< TCPIP$NTP_RUN.LOG;2352     2/18       1-NOV-2004 06:37:16.68< TCPIP$NTP_RUN.LOG;2351     1/18       1-NOV-2004 06:37:16.40< TCPIP$NTP_RUN.LOG;2350     1/18       1-NOV-2004 06:35:39.86< TCPIP$NTP_RUN.LOG;2349     1/18       1-NOV-2004 06:35:39.52< TCPIP$NTP_RUN.LOG;2348     1/18      31-OCT-2004 07:48:04.45< TCPIP$NTP_RUN.LOG;2347     1/18      31-OCT-2004 07:47:58.27< TCPIP$NTP_RUN.LOG;2346     6/18      31-OCT-2004 00:59:23.05< TCPIP$NTP_RUN.LOG;2345     9/18      30-OCT-2004 01:58:14.01< TCPIP$NTP_RUN.LOG;2344     6/18      29-OCT-2004 01:56:26.81< TCPIP$NTP_RUN.LOG;2343     4/18      28-OCT-2004 01:54:43.43  L It seems that NTP was being automatically started (28th thru 30th) at around
 1:58 or so.  e  ! So back to my (refined) question..  B Does V5.1 TCPIP NTP always have a resident process - as does V5.0?         John "REBOOT" Brandon  VMS Systems Administrators* firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  % Date: Tue, 02 Nov 2004 05:42:41 +0200p* From: Mike Rechtman <rechtman@tzora.co.il>+ Subject: Re: TCPIP NTP process NOT residenth+ Message-ID: <2uoe1nF2bs5qkU1@uni-berlin.de>C   John Brandon wrote:( > Mike Rechtman wrote: > F >>IIRC SYS$STARTUP:TCPIP$NTP_STARTUP checks for a logical name in the  >>TCPIP$STARTUP_TABLE.J >>If for some reason a previous NTP startup (or shutdown) bombed and left H >>the logical name wrongly defined, you would see the "already started" 
 >>message. >  >  > This makes sense.  > N > However NTP seems to be shutting itself down - I can not find any stack dumpL > foot prints.  I do find that NTP log files are being generated frequently: > > > TCPIP$NTP_RUN.LOG;2354     3/18       1-NOV-2004 08:02:16.63> > TCPIP$NTP_RUN.LOG;2353     1/18       1-NOV-2004 08:02:16.30> > TCPIP$NTP_RUN.LOG;2352     2/18       1-NOV-2004 06:37:16.68> > TCPIP$NTP_RUN.LOG;2351     1/18       1-NOV-2004 06:37:16.40> > TCPIP$NTP_RUN.LOG;2350     1/18       1-NOV-2004 06:35:39.86> > TCPIP$NTP_RUN.LOG;2349     1/18       1-NOV-2004 06:35:39.52> > TCPIP$NTP_RUN.LOG;2348     1/18      31-OCT-2004 07:48:04.45> > TCPIP$NTP_RUN.LOG;2347     1/18      31-OCT-2004 07:47:58.27> > TCPIP$NTP_RUN.LOG;2346     6/18      31-OCT-2004 00:59:23.05> > TCPIP$NTP_RUN.LOG;2345     9/18      30-OCT-2004 01:58:14.01> > TCPIP$NTP_RUN.LOG;2344     6/18      29-OCT-2004 01:56:26.81> > TCPIP$NTP_RUN.LOG;2343     4/18      28-OCT-2004 01:54:43.43 > N > It seems that NTP was being automatically started (28th thru 30th) at around > 1:58 or so.  t > # > So back to my (refined) question.a > D > Does V5.1 TCPIP NTP always have a resident process - as does V5.0? >  >  >  >  > John "REBOOT" Brandone > VMS Systems Administrator , > firstname.lastname.spam.me.not@dalsemi.com  @ As far as I know TCPIP NTP starts a new log file every 24 hours.- This does NOT signify restarting the service.t   Mike.a   -- rJ New to Usenet? read http://eisner.encompasserve.org/~rechtman/post_hlp.htmE --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that.!D Mike Rechtman                                 *rechtman@tzora.co.il*C    "20% of a job takes 80% of the time, the rest takes another 80%"hE ---------------------------------------------------------------------t   ------------------------------   Date: 1 Nov 2004 13:04:14 -0600i4 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) Subject: Re: Time Change3 Message-ID: <KAzn5P+73+Xo@eisner.encompasserve.org>o  q In article <WACwvISVvhaU@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: J >    If you load and start almost any time service that will take care of ) >    it for you.  We use Multinet's XNTP.E  I NTP under TCPIP will only handle small changes, not errors as big as 3600  seconds.  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"s& 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfeL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  1 	26-October, 2001: A day that will live in infamye4 	Support Freedom: http://www.indefenseoffreedom.org/  ? 	Voting for "the lesser of two evils" is still voting for evil.t   ------------------------------  % Date: Mon, 01 Nov 2004 12:28:31 -0700a+ From: "Barry Treahy, Jr." <Treahy@MMaz.com>d Subject: Re: Time Change' Message-ID: <41868E5F.8040902@MMaz.com>e   Bob Kaplow wrote:e  r >In article <WACwvISVvhaU@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: >    >eJ >>   If you load and start almost any time service that will take care of ) >>   it for you.  We use Multinet's XNTP.o >>     >> >iJ >NTP under TCPIP will only handle small changes, not errors as big as 3600	 >seconds.' >    > @ Process's TCPware didn't miss a beat, it handled it just fine...   Barrya   -- f  < Barry Treahy, Jr                     E-mail: Treahy@MMaz.com< Midwest Microwave, Inc.                  Phone: 480/314-1320< Vice President & CIO                      FAX:  480/661-7028  I                        ... but it's a DRY HEAT!                          u   ------------------------------  % Date: Mon, 01 Nov 2004 22:08:12 -0600e2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: Time Change+ Message-ID: <4187082C.E9EDA817@comcast.net>    "Barry Treahy, Jr." wrote: >  > Bob Kaplow wrote:e > t > >In article <WACwvISVvhaU@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: > >  > >pK > >>   If you load and start almost any time service that will take care ofi+ > >>   it for you.  We use Multinet's XNTP.t > >> > >> > >iL > >NTP under TCPIP will only handle small changes, not errors as big as 3600 > >seconds.  > >  > >rB > Process's TCPware didn't miss a beat, it handled it just fine...  G I know. Really screws you up if you don't have hings timed quite right.e  C I've been in touch with the CSC about this. Seems there are still a D number of issues surrounding this, UTC, etc. Watch for related ECOs,$ that's about all we can do, I guess.   --   David J Dachtera dba DJE Systemss http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:a" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/r   ------------------------------  * Date: Mon, 1 Nov 2004 19:48:11 +0000 (UTC), From: lewis@PROBE.MITRE.ORG (Keith A. Lewis) Subject: Re: VMS for Bush ?t. Message-ID: <cm63tr$pn8$1@newslocal.mitre.org>   "Doc." <doc.cypher@openvms-rocks.com> writes in article <Xns9591D5848B045dcovmsrox@212.100.160.123> dated 29 Oct 2004 18:58:20 GMT:bJ >%NEWS-I-NEWMSG, David Svensson wrote in news:734da31c.0410291032.316f6826 >@posting.google.com >mD >> OpenVMS.org: www.openvms.org (and SKHPC: www.shannonknowshpc.com); >> advertise for George W Bush: http://www.georgewbush.com.l0 >> (I cannot see it because I am not in the US.) >> t* >> And I thought VMS people were wise ? ;)  G >I hope Terry isn't expecting too warm a welcome here in Belgium after i> >this, he'd do well to remember that this country brought war I >crimes/genocide charges against members of GWB's government.  Promoting c5 >them immediately prior to a visit might not be wise.m  G Cut Terry some slack on this.  Under certain circumstances US media aresI *required* to sell advertising space to political candidates who want it.o  F At least the Republicans chose to spend their money supporting the VMSH world... I was going to say "rather than Windows" but there are probably tons of ads on msn/msnbc too.n  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------   Date: 1 Nov 2004 20:08:06 -0600 4 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) Subject: Re: VMS for Bush ? 3 Message-ID: <eztm1fF1f1iz@eisner.encompasserve.org>r  ] In article <cm63tr$pn8$1@newslocal.mitre.org>, lewis@PROBE.MITRE.ORG (Keith A. Lewis) writes:oI > Cut Terry some slack on this.  Under certain circumstances US media areeK > *required* to sell advertising space to political candidates who want it.f  J At HPworld, Terry had a "VietNam Vet against Kerry" or something like that8 button. Don't think I got into politics with Ken Farmer.  E There's three things that shouldn't be discussed in public; Religion,t  Politics, and the Great Pumpkin.  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfcL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  1 	26-October, 2001: A day that will live in infamyc4 	Support Freedom: http://www.indefenseoffreedom.org/  ? 	Voting for "the lesser of two evils" is still voting for evil.    ------------------------------  $ Date: Mon, 1 Nov 2004 17:31:13 -0500# From: "John Smith" <a@nonymous.com>iA Subject: [OT]: No publication tests and writes about VMS this wayo, Message-ID: <ApKdnYoxO6KvJBvcRVn-uQ@igs.net>  L Even as shallow an article like this about VMS would help...as it appears to# be impartial even if it may not be. ) The article is about 6 weeks old but.....o    ? http://www.thechannelinsider.com/article2/0,1759,1647671,00.aspv  " Solaris 10 Shines in Early Testing By Jason Brooks, eWEEK September 20, 2004  L The increasing prominence of freely licensed Linux has prompted many to viewA operating systems in general as a commodity. With Solaris 10, SuntF Microsystems hopes to demonstrate that a company's choice of operatingF system does matter and that the level of innovation Sun has built intoB Solaris 10 can deliver benefits across a company's infrastructure.  H eWEEK Labs has been testing early-release versions of Solaris 10 throughH Sun's Software Express Program, in which beta versions of Solaris 10 are available for free download.  L Based on the time that we've spent with Solaris 10, we believe the system isH a much more attractive option for enterprises than previous versions-notG only because of the operating system's increased functionality but alsouK because of Sun's more inclusive platform strategy: At launch, Sun will shipuJ Solaris 10 for the x86, SPARC and Advanced Micro Devices' AMD64 platforms.  K DTrace is one of the best examples of how Solaris 10 takes advantage of therE central position of the operating system in an enterprise's computingtL infrastructure. DTrace is a diagnostic tool that opens a window to the innerK workings of Solaris 10 and any application running on it-in a finer-grainedoK manner than any other utility of which we're aware. What's more, DTrace can G be used safely on running systems; this allows system administrators to D locate performance bottlenecks in production settings, something not+ generally possible with tools of this type.t  J Also impressive in our Solaris 10 testing so far is the operating system'sA N1 Grid Container feature, which lets administrators run multipleaH applications, each in its own sandbox, on a single system. This feature,K similar to the "jails" found in FreeBSD, is good for isolating applications-H for security reasons or when applications require conflicting libraries.  K During testing, we've found N1 Grid Containers easy to set up and certainlyDE much faster to reboot and update than full systems or virtual machine.
 instances.   ------------------------------   End of INFO-VAX 2004.608 ************************