0 INFO-VAX	Sat, 07 Jan 2006	Volume 2006 : Issue 14      Contents:$ Re: HSZTerm under newer VMS versionsF RE: MCI is not more. Curly gets 40 million to kill yet another companyF Re: MCI is not more. Curly gets 40 million to kill yet another company more lbr$flush problems  Round internal SCSI cables ??   Re: VLC & SCSI Drive Help Needed  Re: VLC & SCSI Drive Help Needed+ Re: WVNETcluster uptime reaches 10 years...   F ----------------------------------------------------------------------   Date: 7 Jan 2006 09:57:04 -0700 1 From: nothome@spammers.are.scum (Malcolm Dunnett) - Subject: Re: HSZTerm under newer VMS versions , Message-ID: <UPnKtDufps2n@malvm9.mala.bc.ca>  - In article <hBM4iwTBlFIE@cuebid.zko.hp.com>,  3    brooks@cuebid.zko.hp.nospam (Rob Brooks) writes:   K >> Rich, you bet we're still using HSZterm under VMS 7.3-2 (I haven't tried G >> it yet on 8.2).  Couldn't live without it.  Never ever had a problem * >> with it despite it being "unsupported". > K > It still works correctly (for me) on an early version of what will become  > V8.3.   G    Is there a place I can get HSZterm from? Does it work with an HSG80?   @    What is the "supported" method of remotely managing an HSG80?  F    I'm using SWCC from a WinXP box currently, but that doesn't work ifA the HSG80 is fully connected ( ie two FC apapters in the host and C a dual-controller HSG80 ) as there seems to be a bug in the startup E script that bombs with a DCL symbol too long error or something. SWCC ( doesn't appear to work at all on VMS 8.2  P    I've seen references to a "Storageworks Command Scripter" product but all theL web documentation I can find is several years old. Is this still a supportedG product? Will it continue to be? ( I'd hate to buy it and find it stops  working a year later )  L    I'm thinking of going the route others have suggested and hooking up someN old LAT terminal servers to the console ports on the HGS80s. In some ways thatM seems like the best idea as it would give me an out-of-band management access N to the arrays even if the host systems were down ( as long as I had a terminal availble).     ------------------------------  * Date: Sat, 7 Jan 2006 11:48:34 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)O Subject: RE: MCI is not more. Curly gets 40 million to kill yet another company $ Message-ID: <dpo9qi$2rp$1@online.de>  
 In articleG <FD827B33AB0D9C4E92EACEEFEE2BA2FB773AB5@tayexc19.americas.cpqcorp.net>, * "Main, Kerry" <Kerry.Main@hp.com> writes:   H > Well, as much as I do not like seeing industry exec's get big payoff'sJ > for little to show accomplishments, you have to admit that taking over aG > *bankrupt* company and then turning it around a short time later such < > that shareholders receive $6.7B is a very impressive task. > @ > I bet that most shareholders feel that he did a fantastic job.  < Of course, there are always two points of view: is the CEO'sH compensation justifiable from the point of view of the company and is itD justifiable from a "moral" point of view.  Basically, from the firstG point of view, if the savings incurred by hiring a cheaper CEO are more H than offset by the lower profitability of the company, the expensive CEOF is worth his salary, whatever it is---end of discussion.  In practice,B of course, it is difficult to answer the question "what would haveG happened if we had hired a CEO for half the salary?".  If one believes  C that the market takes care of everything, not only with respect to  H economics but with respect to society, then the second point of view is A irrelevant.  If one takes a different point of view, the easiest  I solution seems to be to have a rather steep progressive income tax (and,  G of course, make sure that all the compensation the CEO gets is taxable  H income and that it is actually taxed): the more the CEO earns, the more D he is contributing to society which would NOT be contributed if his ; compensation were, say, split up among 10 or 100 employees.   I Of course, someone decides how much to pay the CEO.  My point of view is  I that a success-dependent compensation is the best solution.  It means he  I gets a lot more if the company does well, but in that case he is "worth"  H it.  If the company doesn't do well, then he gets essentially nothing.  D Any CEO who rejects such a solution on the grounds that it would be I unfair for him to get essentially nothing if the company doesn't do well  H admits that he is not the main factor in determining the success of the H company (which might be true), in which case a substantial compensation C (whether or not dependent on the success of the company) should be  $ rejected by him for the same reason.  ; It's down to whoever sets the CEO's salary, essentially the A shareholders.  Many, of not most, shareholders are interested in  I increasing their own profit within their lifetimes and are not concerned  F with the long-term profitability of the company.  There is an obvious G conflict of interest here.  However, keep in mind that no-one forces a  H company to go public.  In the old days, it might have been necessary to D raise capital to buy a factory or whatever, but certainly with "new G economy" companies who don't need to make substantial investments, the  E motivation is greed on the part of the company, so they can't really  D complain when the company is bought up by hedge funds interested in A short-term profit, even if it means the end of the company.  The  H furniture company IKEA, for example, is not a public company, so no-one D has to worry about the share price.  I think Porsche is also a good I example of how to run a company well.  It is the smallest car company in  F Germany (apart from specialist literally "garage companies" who might F hand-build 2 or 3 cars per year) in terms of the number of cars sold, G but it is by far the most profitable, not just per car but in absolute  D terms.  Porsche recently bought 20% of VW with cash.  While it is a F public company on paper, in practice the only shares which are traded H are those which have no voting rights.  The other shares are owned by 8 F people: all of them grandchildren of Ferdinand Porsche.  (4 are named G Porsche and 4, because they are descendants of a daughter of Ferdinand  H Porsche, are named Pich.)  The CEO is a technical guy who used to work ' among the rank and file of the company.    ------------------------------  * Date: Sat, 7 Jan 2006 11:52:13 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)O Subject: Re: MCI is not more. Curly gets 40 million to kill yet another company $ Message-ID: <dpoa1c$2rp$2@online.de>  5 In article <43BF1A74.8D91BB61@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:     > "Main, Kerry" wrote:J > > Well, as much as I do not like seeing industry exec's get big payoff'sL > > for little to show accomplishments, you have to admit that taking over aI > > *bankrupt* company and then turning it around a short time later such > > > that shareholders receive $6.7B is a very impressive task. > I > Impressive task ? All he did was take the job and salary and change the I > name to MCI to make the company more attractive to an investor and then + > hired a M&A banker to find a good buyer.    I If it's not worth $40 million, then why did those who decided to pay him  + that much waste their money in this manner?    ------------------------------   Date: 6 Jan 2006 11:48:36 -0800 3 From: "Joshua Lehrer" <usenet_vms@lehrerfamily.com>   Subject: more lbr$flush problemsC Message-ID: <1136576916.454905.276650@g49g2000cwa.googlegroups.com>   ? I work with little 32-bit virtual address space to spare.  When E creating and save a module for a text library, I have discovered that G lbr$put_record consumes 32-bit virtual address space, cacheing the data G that was passed in.  It isn't until the module is ended, or the library E is closed, that the address space is cleared.  Thus, I can not create D an entry in a text library which is larger than the amount of 32-bitF virtual address space available.  This can easily be demonstrated with
 a program.  F My thought was to periodically call lbr$flush.  The documentation saysF "The LBR$FLUSH routine writes modified blocks back to the library fileG and frees the virtual memory the blocks had been using."  I could write , 10 megs, flush, wirte 10 more, flush, etc...  B However, I have discovered that calling lbr$flush between calls to6 lbr$put_record causes the library to become corrupted.  B To demonstrate this, I include a sample program and sample output:   $ dir *.tlb ! %DIRECT-W-NOFILES, no files found  $ main $ dir *.tlb   	 Directory   
 TEMP.TLB;1   Total of 1 file.   $ lib/list temp.tlb 6 Directory of TEXT library **** on  6-JAN-2006 14:43:46D Creation date:   6-JAN-2006 14:43:34      Creator:  Librarian A09-23? Revision date:   6-JAN-2006 14:43:34      Library format:   3.0 > Number of modules:      1                 Max. key length:  64D Other entries:          0                 Preallocated index blocks:   1 B Recoverable deleted blocks:      0        Total index blocks used:   1 B Max. Number history records:       0      Library history records:   0    TEMP  - $ lib temp.tlb/output=sys$output/extract=temp  ABCDEFGHIJKLMNOPQRSTUVWXYZ ABCDEFGHIJKLMNOPQRSTUVWXYZ    ) %LIBRAR-E-READERR, error reading temp.tlb  -LBR-E-INVRFA, invalid RFA    F Does anyone have any ideas that can help?  Is there another way to getD the librarian to consume less precious 32-bit virtual address space?F Am I using lbr$flush incorrectly?  I am actively working on freeing upG more 32-bit address space, so you don't need to recommend that, thanks.  :)  7 Here is the sample mainprog to demonstrate the problem:    #include <lbrdef.h>  #include <descrip.h> #include <credef.h>  #include <lbrdef.h>  #include <lbr$routines.h>    #include <string.h>   B struct dsc$descriptor_s *strconv(struct dsc$descriptor_s *desc,	/* place to put descriptor */1 				 const char *string)		/* string to convert */  { (   if (string) {			/* if string exists */=     desc->dsc$w_length=strlen(string); /* length of string */ >     desc->dsc$a_pointer=(char*)string; /* pointer to string */=     desc->dsc$b_class=DSC$K_CLASS_S;   /* descriptor class */ D     desc->dsc$b_dtype=DSC$K_DTYPE_T;   /* Data type: ASCII string */;     return desc;		       /* return pointer to descriptor */    }    else
     return 0;  }      int main() {=   int create_opts[]={		        	/* Array of create options */ (     LBR$C_TYP_TXT,				/* Text library */%     64,						/* Maximum key length */ 2     3,   	       				/* Initial file allocation */%     1,						/* Num library indexes */ +     0,						/* Num of add'l bytes in hdr */ .     1,					        /* Num entries preallocd */&     0,						/* Num update hist recs */'     CRE$C_VMSV3,				/* VMS version 3 */ 5     CRE$C_HLPCASING,				/* Character case handling */      0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0,						/* Reserved */     0						/* Reserved */      };     int lib_index=-1;    int type = LBR$C_TYP_TXT;    int flag = LBR$C_CREATE;   struct dsc$descriptor_s desc;    int rfa[2] = {0,0};   *   lbr$ini_control(&lib_index,&flag,&type);  =   lbr$open(&lib_index,strconv(&desc,"temp.tlb"),create_opts);     K lbr$put_record(&lib_index,strconv(&desc,"ABCDEFGHIJKLMNOPQRSTUVWXYZ"),rfa); 7   lbr$insert_key(&lib_index,strconv(&desc,"TEMP"),rfa);   K lbr$put_record(&lib_index,strconv(&desc,"ABCDEFGHIJKLMNOPQRSTUVWXYZ"),rfa); '   lbr$flush(&lib_index,LBR$C_FLUSHALL);   K lbr$put_record(&lib_index,strconv(&desc,"ABCDEFGHIJKLMNOPQRSTUVWXYZ"),rfa);      lbr$put_end(&lib_index);     lbr$close(&lib_index); }     
 joshua lehrer    ------------------------------  $ Date: Sat, 7 Jan 2006 13:15:55 -0500 From: "Jilly" <jilly@hp.com>& Subject: Round internal SCSI cables ??* Message-ID: <43c0055d@usenet01.boi.hp.com>  E Anyone ever seen round internal SCSI cables al what you can find for  M internal IDE cables?  I need to replace the 17-04474-01 & 17-04009-01 cables  E on my PWS 600au so that I can fit a few more options inside the case.    Jilly    ------------------------------   Date: 7 Jan 2006 17:52:22 +0100 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)) Subject: Re: VLC & SCSI Drive Help Needed , Message-ID: <43bfffd6$1@news.langstoeger.at>  Z In article <11rvroqnha2oq12@corp.supernews.com>, Glenn Everhart <Everhart@gce.com> writes:F >Normally I would expect any decent SCSI drive to work ok on a Vax VLC  D if it is only 1" high and "cool enough" and has no SCA but 50pin etc  G >with the caveat that VMS prior to V6 (or was it 6.1?) had a filesystem K >size limit of something like 2^24 blocks. Should a SCSI drive of over 2^31 I >blocks ever appear, it will cause problems. One could use a partitioning I >strategy for slightly larger disks, but that will quickly become hard to L >sustain. Drive vendors aren't quite there yet though (but very very close).  I and as the LD driver is now (since VMS V8.2) full part of the opsys it is L (for a non-system disk) not that problematic as it was in the 'earlier days'  H         An example can be to partition a disk without any file structure         on it:  =           $ LD CONNECT DKA0: LDA1:/LBN=(START=0,COUNT=100000) B           $ LD CONNECT DKA0: LDA2:/LBN=(START=100000,COUNT=100000)B           $ LD CONNECT DKA0: LDA3:/LBN=(START=200000,COUNT=100000)B           $ LD CONNECT DKA0: LDA4:/LBN=(START=300000,COUNT=100000)  L btw LD was there as a freeware LP, so you aren't SOL on earlier VMS versions   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Sat, 07 Jan 2006 11:36:28 -0500 ' From: Glenn Everhart <Everhart@gce.com> ) Subject: Re: VLC & SCSI Drive Help Needed 0 Message-ID: <11rvroqnha2oq12@corp.supernews.com>   Jim Mehlhop wrote: > Chris Scheers wrote: >  >> VLC User wrote: >>H >>> I believe the RZ24-L drive in my VLC is going to die soon (I've beenI >>> getting mount verification messages and the error count has been 2 or F >>> more lately when I do a SHO DEV), so I'd like to replace it with a/ >>> spare Seagate ST32171N I have sitting here.  >>>   >>> So, I have two questions ... >>> 1 >>> A) Will a Seagate ST32171N work in a VLC, and J >>> B) If so, what jumper settings on the Seagate ST32171N do I need to be* >>> aware of (besides SCSI ID, of course)? >> >> >>H >> I'm not familiar with this particular drive, but in general you need J >> to set the SCSI ID (obviously), enable parity and disable termination. > >> Those are probably all the jumpers you need to worry about. >>I >> Other concerns come into play depending on the version of VMS.  There  H >> are various SCSI mode page settings having to do with error recovery H >> which are an issue with VMS 5.5-2.  Generally, a drive coming out of H >> the PC world will have the error recovery options turned off and VMS C >> needs them turned on.  The result is a "fatal drive error" when   >> mounting the drive. >>2 >> I think the parameters are: RC, ARRE, and ARWE. >>G >> AFAIK, (but haven't tested) this restriction goes away with VMS 6.2.  >>
 >> Good luck!  >> >  > Thanks > H > That explains a lot of my problem with 5.5-2 (when I have to boot it)  > with a number of my drives.  >  > Jim J Well, normally RC (read continuous) will NOT be turned on for a PC either,J since it allows a drive to fabricate fake data to keep data rate up. (ThatH wouldn't work very well with any codec output either; can't imagine many' cases where RC would be used nowadays.)   ? VMS SCSI drivers should and as I recall do attempt to set error G parameters right, and even if that cannot be done will tend to tolerate 5 no AWRE/ARRE (but bad blk recovery won't be as good).   E Normally I would expect any decent SCSI drive to work ok on a Vax VLC F with the caveat that VMS prior to V6 (or was it 6.1?) had a filesystemJ size limit of something like 2^24 blocks. Should a SCSI drive of over 2^31H blocks ever appear, it will cause problems. One could use a partitioningH strategy for slightly larger disks, but that will quickly become hard toK sustain. Drive vendors aren't quite there yet though (but very very close).    Glenn Everhart   ------------------------------   Date: 7 Jan 2006 08:01:02 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen) 4 Subject: Re: WVNETcluster uptime reaches 10 years...3 Message-ID: <eWHVHo9INQti@eisner.encompasserve.org>   \ In article <43BF5649.51D600A6@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: > Kenneth Farmer wrote: L >> I'm sure it's correct.  I'm wondering if other OS fanatics could argue it >> can be rigged.  > * > This rigging isn't really the question.  > G > Comparing apples to oranges is. Is it fair to compare the uptime of a 4 > VMS cluster  against that of individual machines ?  G The purpose of computers is not to keep a particular green light on the E machine lit, but rather to provide some service to humans.  If VMS is G able to provide that service on a continuous basis that is what counts.   F The involvement of multiple CPUs, threads of execution, power suppliesD etc. is just so much geeky technical trivia not germane to the issue. of whether the service was provided to humans.  J > If you have 2 Solaris machines that provide HTTP/WEB servers and using aI > router to distribute traffic and stop sending traffic to a node that is @ > down, the uptimes project won't report a "cluster" up time butG > individual nodes uptime even though functionally, those solaris boxes 7 > would offer about the same uptime than a VMS cluster.   D That service provided would not be the same if the web site involvedB updating.  For read-only applications I have an even more reliable* technology pre-dating VMS called a "book".   ------------------------------   End of INFO-VAX 2006.014 ************************