/ INFO-VAX	Tue, 03 Jan 2006	Volume 2006 : Issue 5       Contents: Re: HP's strategy explained :  Re: HP's strategy explained : 
 MX Problem* Optimum settings for RMS Buffers in SYSGEN. Re: Optimum settings for RMS Buffers in SYSGEN. Re: Optimum settings for RMS Buffers in SYSGEN. Re: Optimum settings for RMS Buffers in SYSGEN. Re: Optimum settings for RMS Buffers in SYSGEN Re: OT: Mozilla/Firefox D Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major certD Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major certD Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major certP Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert   ...)   ..I Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) I Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) I Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) I Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) I Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) P Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) ...)..P Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) ...)..& Re: Samba v3 on VMS and HP VMS Roadmap& Re: Samba v3 on VMS and HP VMS Roadmap& Re: Samba v3 on VMS and HP VMS Roadmap( TCPIP Services config files in a cluster, Re: TCPIP Services config files in a cluster, Re: TCPIP Services config files in a cluster, Re: TCPIP Services config files in a cluster, Re: TCPIP Services config files in a cluster, Re: TCPIP Services config files in a cluster Re: tcpip5.4 ?' Re: Using VT420 as Alpha System Console # Re: VMS MAIL: DIRECTORY strangeness # Re: VMS MAIL: DIRECTORY strangeness # Re: VMS MAIL: DIRECTORY strangeness   F ----------------------------------------------------------------------  # Date: Mon, 02 Jan 2006 22:43:34 GMT ( From: sdgross@att.net (Stephen Grossman)& Subject: Re: HP's strategy explained :T Message-ID: <sdgross-0201061743140001@173.cambridge-16rh16rt.ma.dial-access.att.net>  @ In article <aqggq1pqsq082p4dfj6n2b2tvi6ghrv0v6@4ax.com>, Charles! <ckraft@SPAMTRAP.west.net> wrote:   F > On Tue, 20 Dec 2005 16:10:27 GMT, sdgross@att.net (Stephen Grossman) > wrote: > E > >In article <1134600217.498950.61910@z14g2000cwz.googlegroups.com>,  > >mark_hpq@yahoo.com wrote: > > ! > >> Latest news on the layoffs :  > >>   > >> http://hpwfr.blogspot.com/  > > E > >Every time you buy x  instead of y, you are indirectly causing the  > >downsizing of  I > >companies which make X. Just think of all the lost jobs in the candle,  > >horse and buggy, D > >and typewriter companies when anti-social, unfeeling, plutocratic > >consumers starting G > >buying light bulbs, cars and computers! And what about the makers of 
 man-pushedJ > >plows!! And water buckets balanced on the head!! And spears!! Mud-brick makers!!( > >I weep for them all. Solidarity Now!! >  > , > I think you got your "X" and "Y" mixed up.   I wail and gnash my teeth!   --  4 ===================================================== Reason is the basic method of human survival.        AYN RAND U ------------------------------------------------------------------------------------- 8 Radically systematic, radical metaphysics: "Existence 2" http://home.att.net/~sdgrossU ------------------------------------------------------------------------------------- 9 Stephen Grossman    Fairhaven, MA, USA    sdgross@att.net    ------------------------------  # Date: Mon, 02 Jan 2006 23:16:45 GMT ( From: sdgross@att.net (Stephen Grossman)& Subject: Re: HP's strategy explained :T Message-ID: <sdgross-0201061816230001@173.cambridge-16rh16rt.ma.dial-access.att.net>  < In article <I6eqf.164475$Gd6.88730@pd7tw3no>, Arthur Entlich <e-printerhelp@mvps.org> wrote:   @ > Might be a nice excuse for making you feel better about being I > unsupportive to workers and people who are trying to make a living and  D > raise their families, but it doesn't convince me, and its a bogus  > argument in this case.  F Personal attacks are an illogical evasion of the need to judge ideas.   E > Companies are not necessarily making changes due to "progress" but  K > because they think they are evolving the way their stockholders wish, or  K > at least, that whatever they do will be perceived as "positive" by those  B > stockholders.  Unfortunately, often stockholders have the ethicsI > of sharks, in other words, none.  Many don't even know what it is they  3 > "own", only if it is making them a profit or not.   L Profits are made by producing the goods and services needed for human life. O This is both moral and practical since morality is a guide to life, not to the  O sacrifice of life; and moral, because profits, ie, values, are a need of life.  A A farmer whose work was unprofitable, ie, without personal value  N for him,  would starve to death. Capitalism requires mutually profitable tradeF and those firms which dont meet the market, go out of business unless E altruists like you vote for govt subsidies or laws which destroy the  
 competition.    H > Apparently, there is still a market for those buggy whips... they get D > used on the workers, but forcing them to work under lowest common G > denominator conditions or they will be replaced with workers who are  @ > even more desperate and will work for less under even harsher H > conditions.  This is the new "global" international workplace we live K > with, and it is creating a downward spiral, rather than lifting up those  I > who are worse off.  And it is indeed fueled by greedy stockholders who    > only care about their returns.  I Youre a commie liar. The economic history of the last several centuries,  F of growing capitalism, is radically different from the endless povertyA of all preceding history (partially excepting Greece and Rome).   C The spreading, growing wealth of the new capitalism in China, India ? and even Africa gives the lie to your absurd "downward spiral." 6 Capitalism is practical because it is the economics ofJ the mind's ability to guide the production of the values needed for life. B You are typing at a computer, not dragging out your days as a serf? pulling a plow w/your own body and living in a dirt-floored hut ? with your farm animals (and w/o room spray!). Capitalism ended  ! famines and tyranny and increased > life spans. Of course, if mindlessness is your ideal, then theD rationality needed for capitalism will be hateful. Im sorry but the H Garden of Eden (or Marx's communism) is impossible. Focusing your mind, E logically and thru the senses, onto the concrete, material universe,  C is a need of human life. The mind is not a game show for neurotic,  E disinterested, pseudo-intellectuals. Its mans main tool of survival.  A People in capitalist economies live better than was even imagined  centuries ago.    % FOCUS YOUR MIND! BAD IDEAS CAUSE ADD!     F > > Every time you buy x  instead of y, you are indirectly causing the > > downsizing of J > > companies which make X. Just think of all the lost jobs in the candle, > > horse and buggy,  E > > and typewriter companies when anti-social, unfeeling, plutocratic  > > consumers startingH > > buying light bulbs, cars and computers! And what about the makers of
 man-pushedA > > plows!! And water buckets balanced on the head!! And spears!!  Mud-brick makers!!) > > I weep for them all. Solidarity Now!!  > >    --  4 ===================================================== Reason is the basic method of human survival.        AYN RAND U ------------------------------------------------------------------------------------- 8 Radically systematic, radical metaphysics: "Existence 2" http://home.att.net/~sdgrossU ------------------------------------------------------------------------------------- 9 Stephen Grossman    Fairhaven, MA, USA    sdgross@att.net    ------------------------------  % Date: Mon, 02 Jan 2006 14:11:35 -0800 # From: "Tom Linden" <tom@kednos.com>  Subject: MX Problem ( Message-ID: <ops2rulluozgicya@hyrrokkin>  @ Have had MX5.4 running successfully on a 7.3 system with mx_rootB located on a shadowed shared scsi drive and decided for redundancy? reasons to configure agents on a 7.3-2 node.  Now I use Outlook C for my mail client.  After doing this incoming mail stopped working F on the first node and changing the server settings in Outlook to pointE incoming to the new node I was able to receive mail; however outgoing G mail was still to the old node.  The incoming and outgoing server names ; were the same, mail.kednos.com  so by chaning incoming to    newnode.kednos.comI Outlook was happy.  TCPIP5.4 is on both.  So the processes running on the  two are:   On Old) FREJA> pipe sho system | sear sys$pipe mx K 210007C6 MX FLQ Manager  HIB      6       20   0 00:00:00.05       145       154 K 210007C7 MX Router       HIB      6      533   0 00:00:00.20       170       176 K 210007C8 MX Local        HIB      6      418   0 00:00:00.05       202       203 K 210007C9 MX SMTP         HIB      6     1106   0 00:00:00.23       223       231 K 210007CA MX SMTP Server  HIB      6     7254   0 00:00:01.74       278       308   , FREJA> pipe sho system | sear sys$pipe tcpipK 2100011C TCPIP$INETACP   HIB     10    47878   0 00:00:09.37       288       211 K 2100011D TCPIP$ROUTED    LEF      6    94048   0 00:00:00.05       307       130  SK 2100011F TCPIP$NTP_1     LEF      9   721389   0 00:00:05.14       292       272  NJ 21000121 TCPIP$POP_1     HIB      9 10695967   0 00:47:34.24      7964     5759  N K 21000122 TCPIP$PWIP_ACP  HIB      9       66   0 00:00:00.06       133       156 K 2100012F TCPIP$SSH_BG119 LEF      8    61529   0 00:00:11.03       470       507  N     and on the New  K 2080B271 MX FLQ Manager  HIB      6     1936   0 00:00:00.37       160       194 K 2080B272 MX Router       HIB      6     1909   0 00:00:00.39       183       212 K 2080B273 MX Local        HIB      6     2876   0 00:00:00.76       289       324   K 2080B26A TCPIP$INETACP   HIB     10      201   0 00:00:00.09       295       224 K 2080B26D TCPIP$NTP_1     LEF     10    25721   0 00:00:00.52       283       297  NK 2080B26E TCPIP$POP_1     HIB      9    59950   0 00:00:15.81       592       606  NK 2080B276 TCPIP$SSH_BG80  LEF      8    14402   0 00:00:02.45       554       700  NK 2080B27C TCPIP$ROUTED    LEF      6     2167   0 00:00:00.02       429       136  S   so  - 1. why did incoming mail stop on the old, and 5 2. why didn't SMTP and SMTP Server start on the new ?    ------------------------------  % Date: Mon, 02 Jan 2006 16:54:49 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 3 Subject: Optimum settings for RMS Buffers in SYSGEN , Message-ID: <43B9A111.CD56CC87@teksavvy.com>  B While I am playing in modparams.dat, I figures it would be time to2 decide on whether to add settings for rms buffers.   This is on VAX.   > First, if RMS_DFMBC is set to 16 and all the other RMS buffersG parameters are set to 0, does this mean that all the others inherit the  value of "16" ?   ? Also, is there much of a cost in memory to increase the default F multibuffer count ?  Looking back at google, seems 132 was mentioned a lot. (would 128 be better ?)  B Also,  $SET RMS_DEFAULT  allows a /BLOCK parameter which specifies" blocks akllocated for each buffer.  > But in SYSGEN, there doesn't seem to be a corresponding set of0 parameter, there is only a network block count. D (equivalent of SET RMS/NETWORK/BLOCK=8  if I understand correctly).   F Is there a magical relationship between block count and buffer count ?  A for instance, if I have 2 buffers of 100 blocks, is that the same & peformance as 25 buffers of 8 blocks ?   ------------------------------   Date: 2 Jan 2006 14:28:38 -0800 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>7 Subject: Re: Optimum settings for RMS Buffers in SYSGEN C Message-ID: <1136240917.840991.143520@f14g2000cwb.googlegroups.com>   E JF, you seem to be confused between buffer size and buffer counts, or % perhaps your writing has me confused. E SYSGEN (7.2-1) is also confused/conusing labeling DFMBFSDK as blocks, % where it should be buffers... I think   D The DFMBC is all about Multi-Block-Count, just for sequential files.A Indexed and Relative files have their BUCKET SIZE and so there is C nothing to inherit there. For those you can really only choose then  number of buffers (RMS_DFMBF*)  ? For a VAX you often need to worry about memory pressure then IO D performance may be limited, so 8KB buffers may still be appropriate,G but I would lean towards a default of 32 for 16KB buffers. You may want A to conside a compromise by using SET RMS/SEQ/BLO=32 for batchjobs  through SYLOGIN.COM   I >> Is there a magical relationship between block count and buffer count ?   # memory = buffer_count * buffer_size   ? The total memory pressure is obtained by multiplying BLOCKS and G BUFFERS. That is 8 buffers of 16 blocks require as much memory (a touch : more due to control structures) as 4 buffers of 32 blocks.F In my experiments/experience I have rarely seen a benefit of more thanA 4 buffers for sequential files, so i would send my memory towards  larger buffers, not more.   C Where did you see 132 buffer (count) for RMS memtioned? That's just  silly!D Why would 128 be better? (actually it will be, because 132 is likely* very bad, and anything less is better :-).  C For multi block count, the absolute max is 127 and there you indeed C might want to tweak down to a multiple of 4 (120, for an EVA Raid-5 G quirk) or even down to a mutliple of 16 (112 or 96, to lign up with the  XFC)  5 > if I have 2 buffers of 100 blocks, is that the same & peformance as 25 buffers of 8 blocks ?  G Unlikely to be the same. 2 buffers of 100 blocks is likely to be better D for large sequential files read or writes: fewer trips down IO lane.F 3 buffers of 64 might be better if/when Read-ahead and/or Write-behind> is active (default for fortran and cobol applications). 1 or 2  outstanding IOs is often plenty.? The 8 buffers of 25 could be better if the application is doing D random-ish access over a small range in a sequential file which is aG rather unlikely scenario. Such larger number of (read-ahead) buffers my G make a read-ahead look  like a random IO and disable optimizations down 
 the IO stack.    When in doubt... experiment!   hth, Hein.    ------------------------------  % Date: Tue, 03 Jan 2006 00:15:35 +0100 & From: Paul Sture <paul.sture@decus.ch>7 Subject: Re: Optimum settings for RMS Buffers in SYSGEN , Message-ID: <41tqgoF1ggi5tU1@individual.net>   Hein RMS van den Heuvel wrote:G > JF, you seem to be confused between buffer size and buffer counts, or ' > perhaps your writing has me confused. G > SYSGEN (7.2-1) is also confused/conusing labeling DFMBFSDK as blocks, ' > where it should be buffers... I think  > F > The DFMBC is all about Multi-Block-Count, just for sequential files.C > Indexed and Relative files have their BUCKET SIZE and so there is E > nothing to inherit there. For those you can really only choose then   > number of buffers (RMS_DFMBF*) > A > For a VAX you often need to worry about memory pressure then IO F > performance may be limited, so 8KB buffers may still be appropriate,I > but I would lean towards a default of 32 for 16KB buffers. You may want C > to conside a compromise by using SET RMS/SEQ/BLO=32 for batchjobs  > through SYLOGIN.COM  >  > I >>>Is there a magical relationship between block count and buffer count ?  >  > % > memory = buffer_count * buffer_size  > A > The total memory pressure is obtained by multiplying BLOCKS and I > BUFFERS. That is 8 buffers of 16 blocks require as much memory (a touch < > more due to control structures) as 4 buffers of 32 blocks.H > In my experiments/experience I have rarely seen a benefit of more thanC > 4 buffers for sequential files, so i would send my memory towards  > larger buffers, not more.  > E > Where did you see 132 buffer (count) for RMS memtioned? That's just  > silly!F > Why would 128 be better? (actually it will be, because 132 is likely, > very bad, and anything less is better :-). > E > For multi block count, the absolute max is 127 and there you indeed E > might want to tweak down to a multiple of 4 (120, for an EVA Raid-5 I > quirk) or even down to a mutliple of 16 (112 or 96, to lign up with the  > XFC) >  > 5 >>if I have 2 buffers of 100 blocks, is that the same  > ( > peformance as 25 buffers of 8 blocks ? > I > Unlikely to be the same. 2 buffers of 100 blocks is likely to be better F > for large sequential files read or writes: fewer trips down IO lane.H > 3 buffers of 64 might be better if/when Read-ahead and/or Write-behind@ > is active (default for fortran and cobol applications). 1 or 2" > outstanding IOs is often plenty.A > The 8 buffers of 25 could be better if the application is doing F > random-ish access over a small range in a sequential file which is aI > rather unlikely scenario. Such larger number of (read-ahead) buffers my I > make a read-ahead look  like a random IO and disable optimizations down  > the IO stack.  >  > When in doubt... experiment! >   F Definitely experiment! One of my standard performance boosters in VAX ; days was to use large multibuffer counts (64 or 127 *) for  > reading/writing large files sequentially, though according to A application I'd use it programatically per file so that I wasn't  A applying it to random access files. As that implies I did it for  : specific (batch) jobs rather than applying it system wide.  I I never found a significant gain by increasing the multiblock counts, so  G concentrated on multibuffers instead. Of course this is a typical area   where YMMV.   G * high multibuffer values came out of NPAGEDYN back when I did it last  F on VAX (pre-V7.n). If hitting that too hard, reduce multibuffer count  and try again.   ------------------------------  % Date: Mon, 02 Jan 2006 19:44:35 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 7 Subject: Re: Optimum settings for RMS Buffers in SYSGEN , Message-ID: <43B9C8D0.E3CF18C8@teksavvy.com>   Hein RMS van den Heuvel wrote: > G > JF, you seem to be confused between buffer size and buffer counts, or ' > perhaps your writing has me confused.   D Yes, I was confused...  Actually, the help SYS RMS_DFMBC agrees withG your decription, I was the one confused because it specified sequential G files. But your explanation of block counts not applying to other files  clarifies this.   < > but I would lean towards a default of 32 for 16KB buffers.  A > The total memory pressure is obtained by multiplying BLOCKS and A > BUFFERS. That is 8 buffers of 16 blocks require as much memory    A Does this mean that for every file that is opened, the #buffers * ! #blocks-per-buffer is allocated ?   H > In my experiments/experience I have rarely seen a benefit of more thanC > 4 buffers for sequential files, so i would send my memory towards  > larger buffers, not more.   6 For sequential files, when are multiple buffers used ?  E For indexed files, I assume that the buffer count determines how many H buckets are maintained in memory by default, helping keep parts of indexH and data buckets, especially if there are multiple streams accessing the same file in the same program ?       E > Where did you see 132 buffer (count) for RMS memtioned? That's just  > silly!  2 It was a post on comp.os.vms about enhancing gzip:X http://groups.google.com/group/comp.os.vms/msg/e98991a8a7fd0aee?dmode=source&hl=xx-elmer   > When in doubt... experiment!  ; No ! When in doubt I come to the experts on comp.os.vms :-)    ------------------------------   Date: 2 Jan 2006 20:23:28 -0800 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>7 Subject: Re: Optimum settings for RMS Buffers in SYSGEN C Message-ID: <1136262208.096207.319540@z14g2000cwz.googlegroups.com>   G Hmm... My kids liked the Elmer language effect on that link. We had not  seen that before. :-).  D I do not entirely trust the details of the article pointed to, but I applaud the general direction.A These are the knobs to turns, the effects to monitor, and too few ! applications have paid attention. F I do not doubt the effect, but there are too many knobs turned at once to know which one did what.   " A few comments on that article....  G MBC=128 is simply invalid in RMS context. The C rtl maxes it out to 127   A And if you ask the C RTL for 'normal' IO it will not even use RMS : record mode, but does it's own blocking/de-blocking  usingG SYS$READ/SYS$WRITE albeit with buffers sizes and count suggested by the 
 RMS knobs.A Check with MONI MODE and/or SET FILE/STAT  + MONI RMS as desired.   F The file system does not null out initial allocations any more or less than subsequent extents.F Simple sequential writes, even with 'highwatermarking enabled', do not cause zeroing out.E But the 'contiguous-best-try' requested may well slow an overly large - initial allocation down on a fragmented disk.   G The GZIP suggestion of 132 buffers each 32KB allocates 4MB per file, if  not explicitly controlled.A This is fine for gzip and some batch jobs, but not fine for every 1 single process times every single defaulted file. @ It might not even be fine for gzip... in a multy user production environment.  G Think about what this unzip suggestion tells RMS: when ever a buffer is F filled, RMS launches the IO without waiting (wbh), and the applicationE can start filling he next buffers, for which the IO also gets queued. 7 Do you really think a vax can get 132 useful IOs going? @ How many MB/sec can a single VAX (un)zip without counting the IO effort? @ Just a handfull in the queue will be more than enough to sustain	 100mb/sec 5 Can your VAX IO system handle 100 MB/sec or 10MB/sec? A Do you really think your IO subsystem will gracefully swallow 132  concurrent IOs?   G Do you think the other users on the system like some gunzipper blasting E 132 IOs in front of the quick single data lookup IO they need to move > forwar, holding them up for a second when they could have have millisecond response? ) It's just not right as a general default!   G On the read side RMS simply ignores more than 8 buffers for read ahead. D SEARCH[/STAT], a not unlikely test tool, hardcodes that to 3 buffersC The other buffers take up useless virtual memory at best, and cause 4 pagefaults at worst, for data never to be revisited.> Look at the ANA/SYS ... SHOW PROC/RMS=BDBSUM for that example:   SDA>9 Process index: 0082   Name: HEIN   Extended PID: 0003B682 9 --------------------------------------------------------- 0                                 BDB/GBPB Summary0                                 ----------------D Address USERS    SIZE     NUMB     VBN    BLB_PTR   ADDR    VAL   ID FLGSD ------- -----    ----     ----     ---    -------   ----    ---   -- ----  E 7B088070   0  00004800 00004800 00002569 00000000 002C8800   0    BDB  VAL E 7B088350   0  00004800 00004800 00002665 00000000 002E3800   0    BDB  VAL E 7B088400   0  00004800 00004800 00002641 00000000 002E8000   0    BDB  VAL E 7B0882E0   0  00004800 00004800 0000261D 00000000 002DF000   0    BDB  VAL E 7B088270   0  00004800 00004800 000025F9 00000000 002DA800   0    BDB  VAL E 7B088200   0  00004800 00004800 000025D5 00000000 002D6000   0    BDB  VAL E 7B088150   0  00004800 00004800 000025B1 00000000 002D1800   0    BDB  VAL E 7B0880E0   0  00004800 00004800 0000258D 00000000 002CD000   0    BDB  VAL E 7B083F38   0  00004800 00000000 00000000 00000000 7B070000   0    BDB E 7B085300   0  00004800 00000000 00000000 00000000 7B074800   0    BDB E 7B085370   0  00004800 00000000 00000000 00000000 7B079000   0    BDB E 7B085400   0  00004800 00000000 00000000 00000000 00138000   0    BDB E 7B085470   0  00004800 00000000 00000000 00000000 0013C800   0    BDB  : @ and on and on.... for 100 buffers in total. Allocated, but never touched.  I >>> For indexed files, I assume that the buffer count determines how many B buckets are maintained in memory by default, helping keep parts of index  and data buckets,    Yes, pretty much. C Come to the VMS bootcamp this May and I'll explain you all about it  :-).3 Better still.. sign up for some private consulting!   Y >> especially if there are multiple streams accessing the same file in the same program ?    Ah, watch out there.F If the program, like so many programs, simply opens the files multiple? times then those opens could just as well have been coming from < multiple processes. They are entirely unaware of each other.G The only way to 'pool' buffers is to perform multiple RAB connects to a  single FAB with shr=mse active. E The only language to support this nicley is BASIC: OPEN... CONNECT... 3 The other languages need USEROPEN support for this. < The better way to share buffers is of course GLOBAL buffers.   Cheers,  And happy newyear!   Hein.    ------------------------------  % Date: Tue, 03 Jan 2006 03:34:14 +0000 7 From: David B Sneddon - bigpond <dbsneddon@bigpond.com>   Subject: Re: OT: Mozilla/Firefox* Message-ID: <43B9F0B6.2070401@bigpond.com>  ) bob@instantwhip.com mentioned in passing:  > TCPware is not free! > and ucx is still suffering > Palmer and his brilliant > move to cut spending on  > development ...   A Have you considered quoting the text to which you are responding? ) (Not that it would probably help much...)    Regards, Dave --  B David B Sneddon (dbs) VMS Systems Programmer dbsneddon@bigpond.comB Sneddo's quick guide ...   http://www.users.bigpond.com/dbsneddon/B DBS freeware   http://www.users.bigpond.com/dbsneddon/software.htm   ------------------------------  % Date: Mon, 02 Jan 2006 13:47:36 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> M Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert , Message-ID: <43B9753C.AA69A5B9@teksavvy.com>   Larry Kilgallen wrote:L > > Many ATMs around here don't "swallow" and regurgitate the card anymore -/ > > you just swipe it (insert / remove) on cue.  > H > I think that choice is made based on the goal of using cheap hardware.F > Many PCs do not eject CDroms and floppies when you dismount, whereas* > Macintoshes do (the ones I use, anyway).  E There are many issues around this. Cost of reader is probably least.  G When you have an external card swipe, not only does the customer always E retain his card (no cards forgotten at ATM), but the bank also has no H way of capturing cards. In the early days of ATMs, banks would swallow aF card after a number of unsuccesful PIN entry attempts for instance (or" if the card was declared stolen).   G Capturing stolen cards costs a lot in management etc as well as causing D many headaches when the card doesn't belong to that bank. Once banksF realised that it was possible to evade card capture by using a foreign= ATM that didn't capture cards, the whole practice was quietly E abandonned. And this now permits the use of simpler card swipers that  don't capture the card inside.  G Personally, I prefer a system when the card is captured to be read, but E then half ejected. This way, if you grab your card and leave, the ATM F knows to cancel your unfinished session. With a simple card swipe, youD can leave the ATM with the session still oppened and the next person having access to your accounts.    ------------------------------  % Date: Mon, 02 Jan 2006 12:57:42 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>M Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert + Message-ID: <43B977A6.B4E588D7@comcast.net>    Larry Kilgallen wrote: > b > In article <43B95609.984EAA06@comcast.net>, David J Dachtera <djesys.nospam@comcast.net> writes: > L > > Many ATMs around here don't "swallow" and regurgitate the card anymore -/ > > you just swipe it (insert / remove) on cue.  > H > I think that choice is made based on the goal of using cheap hardware.  G Perhaps a mixture of the two. No card intake means less moving parts as D well as fewer customer service headaches (both are cost savers, both initially and operationally).   > > Many PCs do not eject CDroms and floppies when you dismount,  C I've never actually seen a "dismount" option for a removable drive, G though in WhineBloze Exploder you can right-click a CD, DVD, Zip or Orb F drive (maybe Jaz too, dunno) and select "Eject" to have it regurgitateG the media. USB devices use a similar function to "stop" the peripheral,  as WhineBloze terms it.   	 > whereas * > Macintoshes do (the ones I use, anyway).  H About the only floppy with an actual "eject" function I ever saw was the LS-120 "SuperDisk" variety.   E Something I'd try (in my "copious spare time") if I had a newer Alpha D that handled IDE better would be to try an LS-120 drive with VMS. At9 120MB raw, it's not real useful, just a curiosity really.    --   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/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  % Date: Mon, 02 Jan 2006 13:40:29 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>M Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert + Message-ID: <43B981AD.EC938D4F@comcast.net>    JF Mezei wrote:  > [snip]I > Personally, I prefer a system when the card is captured to be read, but G > then half ejected. This way, if you grab your card and leave, the ATM H > knows to cancel your unfinished session. With a simple card swipe, youF > can leave the ATM with the session still oppened and the next person! > having access to your accounts.   @ Typically, you need to swipe your card for each transaction. The@ terminal does not stay in an "authorized" mode at the end of theE transaction. They did initially; however, the banks quickly corrected E this. Arguably safer (since both the card number and PIN traverse the = network multiple times - once for each transaction), but less  convenient.    --   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/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  # Date: Tue, 03 Jan 2006 03:30:05 GMT   From: John Santos <john@egh.com>Y Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert   ...)   .. ) Message-ID: <1dmuf.8684$7x.6735@trnddc03>    JF Mezei wrote:  > David J Dachtera wrote:  > J >>Many ATMs around here don't "swallow" and regurgitate the card anymore -- >>you just swipe it (insert / remove) on cue.  >  > I > It gets worse here. The bank thay now asks to confirm a deposit 3 times F > has "upgraded" to Diebold machines with fancy graphics and sound etcF > etc.  Branch staff had to put duck tape over the speaker because theJ > "beeps" were so loud. (untile they got tech folks to come and "fix" it). > E > But it gets worse, with all the fancy stuff, the display and sounds J > don't match the keyboard. I tend to type in my PIN number quickly and byH > the time I am finished, I can still watch the "*" appear on the screen. > and the beeps take even longer to complete.  > J > Reminds me of time sharing in the 1970s where if you typed too fast, the( > screen wouldn't be able to keep up :-) > G > Still cannot understand how a bank's quality control people would let I > this machine out in production. And I imagine that kids of the nintendo D > generation are even faster at the keyboard, so if they have such aG > weenie in their Q/A assurance comittee, Shirley he would have spotted  > the delay problem. >  > J > (The sound delay is probably caused by the "beep" being a recorded soundJ > with a certain amount of silence at the end, so when you press a key, itE > "queues" the sound to be produced, and if the previous sound hasn't I > completed, it waits for it to complete before starting the next sound).   4 Vain attempt to get this thread remotely on-topic...  F The VT52 used to process multiple <BEL> characters by ringing its bellH (actually a buzzer)  for somewhat longer.  Then along came the VT100 (orE was it the VT220? - I don't have a VT100 around to test this on :-(), C which gave a nice little "beep" for each <BEL>.  Suddenly 10 <BEL>s G changed from "bzzzt" to "beep...beep...beep...beep...beep...beep...beep C ...beep...beep...beep"  Since non-printing characters also rang the G bell, accidentally typing a binary file, and it would take hours before $ the stupid terminal finally shut up!         --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 2 Jan 2006 12:10:21 -0800  From: bob@instantwhip.com R Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...)C Message-ID: <1136232621.578924.108260@o13g2000cwo.googlegroups.com>   ? I check mine ... our site works with both Internet Exploder and  Mozilla ... :)   ------------------------------   Date: 2 Jan 2006 12:11:54 -0800  From: bob@instantwhip.com R Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...)C Message-ID: <1136232714.765543.114150@o13g2000cwo.googlegroups.com>   @ that is what you get with freeware ... you get what you pay for!  ? DEC software still is superior to all other trash out there ...    ------------------------------   Date: 2 Jan 2006 12:16:20 -0800  From: bob@instantwhip.com R Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...)C Message-ID: <1136232980.782698.234700@z14g2000cwz.googlegroups.com>   < and what fact of linux is dead on the desktop don't you get?  8 It is too CONVULUTED!  OpenVMS is superior if you want a= OS that is user friendly and actually has commands in english > that you can actually understand ... VMS is more dos like thanC linux and a heck of a lot more user friendly ... it was written for : the user who may or may not be a programmer, and it can be  taylored to your preferences ...   ------------------------------   Date: 2 Jan 2006 18:56:24 -0800  From: bob@instantwhip.com R Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...)C Message-ID: <1136256984.747080.230310@g43g2000cwa.googlegroups.com>    TCPware is not free! and ucx is still suffering Palmer and his brilliant move to cut spending on  development ...    ------------------------------   Date: 2 Jan 2006 20:51:02 -0800 $ From: "AEF" <spamsink2001@yahoo.com>R Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...)C Message-ID: <1136263862.302885.104690@g44g2000cwa.googlegroups.com>   
 AEF wrote: > AEF wrote: > > BRAD wrote:  > > > Rich Jordan wrote:	 > > [...]  > > > K > > > It continues to amaze me that so many people have not heard of (or do ? > > > not use) Mozilla/Firefox - people who consider themselves M > > > "internet-savvy".  It sure makes surfing easier and safer.  I guess too K > > > many folks are used to beating themselves over the head with Internet A > > > Exploder, and have gotten used to the dull ache of Windows.  > > K > > [Normally, I really hate Microsoft, but there are some exceptions. This  > > is one of them.] > > J > > Well, some of us prefer Internet Exploder. It does some things MozillaH > > can't. If you hate the mouse, and I surely do, IE is better. In most > [...]  > ? > And here's another thing I don't like about Firefox: Bookmark ; > management. I can't use a keyboard shortcut to get to the F > search-results pane: I am forced to use the stupid mouse. In IE, theB > bookmarks (favorites) are just files I can manipulate in WindowsC > Explorer. I don't even bother with the IE Favorites Organize crap D > because it is total crap. (It gives you one of those really stupidI > windows that are way too small and can't be maximized. It is one of the I > most worthless pieces of software I have ever seen in my life.) Instead F > I just have a shortcut on my desktop that opens the Favorites folderC > and I can manipulate the favorites and search them with the usual A > Windows stuff -- all without touching the mouse. And why is the @ > Location column in the Firefox bookmark-search window severelyF > truncated leaving a huge blank column for non-existent descriptions. > That's a Microsoft stupidity!  > H > I'm glad that many of you are happy with Firefox. Just want to let youF > know it's not better for everyone as many seem to think -- and maybe5 > find out secret ways to make it work better for me!   @ OK, I just downloaded and installed Firefox 1.5. Some things areB actually better! When I go back, the keyboard focus is on the same link! Cool.   B Bookmarks Manager is now navigable via the TAB key (COOL!), but itA still hides most of the each URL leaving a huge blank Description 9 column. And why is the Add Bookmarks dialog box so small?   C Another thing I prefer about IE is that I can be on a page, click a G link which takes me to another page or site, then Ctrl/N gives me a new G window copying the same page AND THE COMPLETE HISTORY OF THE SESSION to F the new window. This lets me alt-left-arrow to go back to the previousG page and try another link while still having the first link open in the F first window. Firefox still doesn't do this. Someone at work showed meD a workaround but it's awkward and doesn't work for button links. TheE other option is to always successfully anticipate when you might want B to open a new window and go back one page and you have to copy and9 paste the address (unless there's a better way I forgot).   G There basically doesn't seem to be an equivalent at all to IE's Ctrl/N, " which I really like and use often!  D Also, the Firefox fonts are now better for Google groups, but then I@ can't have the Times Roman on other pages with the same setting.( Haven't tried it on other pages, though.  G Also, even with IE the fonts are screwy: On other PC's I can get Google C groups to display articles in Courier (Courier New?) with the fixed B font setting but never on my own computer. Ugh! Fonts are always aE pain. For Google Groups I want Roman for all of it except the article F content for which I want Courier or Courier New. For most other thingsE I want Roman that isn't microscopic. I haven't checked recently but I C always wondered why default fonts on most Web sites used to be (and D perhaps still are) microscopic. View/Text Size/Large used to fix it,G but that broke after which I use the ignore-font-sizes trick to fix the 4 size. Why is it so damn hard to get the fonts right?   AEF    ------------------------------  % Date: Mon, 02 Jan 2006 15:35:18 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> Y Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) ...).. , Message-ID: <43B98E73.1E32C23E@teksavvy.com>   bob@instantwhip.com wrote:  A > DEC software still is superior to all other trash out there ...     B You mean TCPIP Services is superior to TCPware and Multinet ?????     D you left the door opened for this one, you had to expect a splash of cold air to come in :-)    ------------------------------  % Date: Mon, 02 Jan 2006 17:08:34 -0500 ( From: Bill Todd <billtodd@metrocast.net>Y Subject: Re: OT: Mozilla/Firefox (was:Re: Windoze ends year with a major cert ...) ...).. G Message-ID: <urqdnWzW2aT6OSTenZ2dnUVZ_sidnZ2d@metrocastcablevision.com>    bob@instantwhip.com wrote:> > and what fact of linux is dead on the desktop don't you get?  H What you think you're responding to is not at all clear.  Have you been / forgetting to take your meds over the holidays?    >  > It is too CONVULUTED!   I It won't help with problems like the above, but you might consider using   a spell-checker as well.   - bill   ------------------------------   Date: 2 Jan 2006 10:42:23 -0700 1 From: nothome@spammers.are.scum (Malcolm Dunnett) / Subject: Re: Samba v3 on VMS and HP VMS Roadmap , Message-ID: <wloiQvKD4zSE@malvm9.mala.bc.ca>  8 In article <g_auf.501$H37.96934@news20.bellglobal.com>, /     "Neil Rieck" <n.rieck@sympatico.ca> writes:  > M > HP did a good job of porting Apache-2 to OpenVMS so I can only assume that  . > SAMBA-3 will also be of fairly high quality. >   J     Has anyone heard what will happen to the Microsoft Authentication ACMEP module (which right now is part of Advanced Server) if Pathworks/Advanced Server is replaced by Samba?   I Will an equivalent module be distributed with Samba or through some other  channel?   ------------------------------  $ Date: Mon, 2 Jan 2006 13:03:35 -07006 From: "Michael D. Ober" <obermd.@.alum.mit.edu.nospam>/ Subject: Re: Samba v3 on VMS and HP VMS Roadmap . Message-ID: <rGfuf.16$sW2.733@news.uswest.net>  ; I think that's the question I started this thread with.  ;)    Mike.   > "Malcolm Dunnett" <nothome@spammers.are.scum> wrote in message& news:wloiQvKD4zSE@malvm9.mala.bc.ca...9 > In article <g_auf.501$H37.96934@news20.bellglobal.com>, 1 >     "Neil Rieck" <n.rieck@sympatico.ca> writes:  > > I > > HP did a good job of porting Apache-2 to OpenVMS so I can only assume  that0 > > SAMBA-3 will also be of fairly high quality. > >  > L >     Has anyone heard what will happen to the Microsoft Authentication ACMEK > module (which right now is part of Advanced Server) if Pathworks/Advanced  Server > is replaced by Samba?  > K > Will an equivalent module be distributed with Samba or through some other 
 > channel? >  >    ------------------------------  $ Date: Mon, 2 Jan 2006 14:35:33 -07006 From: "Michael D. Ober" <obermd.@.alum.mit.edu.nospam>/ Subject: Re: Samba v3 on VMS and HP VMS Roadmap . Message-ID: <F0huf.31$sW2.745@news.uswest.net>  8 Not a problem.  I just thought it was slightly humorous.   Mike.   > "Malcolm Dunnett" <nothome@spammers.are.scum> wrote in message& news:9T1edx9BAPDF@malvm9.mala.bc.ca...0 > In article <rGfuf.16$sW2.733@news.uswest.net>,< >   "Michael D. Ober" <obermd.@.alum.mit.edu.nospam> writes: > ? > > I think that's the question I started this thread with.  ;)  > >  > : >    Oops, sorry for coming into the thread in the middle. >  >    ------------------------------  % Date: Mon, 02 Jan 2006 13:40:26 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 1 Subject: TCPIP Services config files in a cluster + Message-ID: <43B9738E.ECB0C1F@teksavvy.com>   C I have a dual root disk that started off with TCPIP Services 5.0-9. ! Things worked fine on both nodes.   E Then, I upgrade the system disk, from node1 to TCPIP services 5.3.  I D did not reboot node1. But rebooted node2, only to find that it would crash when starting TCPIP.    G After much investigation, it turns out that the configuration files are H in SYS$COMMON:[SYSEXE] and this seems to be hardcoded in tcpip$config to define it in sys$common.     Is this normal ?  D What I find interesting is that @TCPIP$CONFIG on node1 finds node1'sH info, while TCPIP$CONFIG on node2 finds node2's information.  What "key"E is used by the software to find the configuration records of one node E versus the other in the same config file ? ethernet address ?  system  root ? (SYS0 vs SYS1) ???   D What reason exists to have that file in SYS$COMMON: ? Seems far more% sensible to have it in sys$specific ?     C Also, dumping the file shows there are records left from a previous E temporary incandation of a node when it had a temporary node name. Is D there a way to remove this ? And if I wish to restart node2's configC file from scrath, is that possible without zapping node1's config ?    ------------------------------   Date: 2 Jan 2006 21:29:01 +0100 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)5 Subject: Re: TCPIP Services config files in a cluster * Message-ID: <43b99b1d@news.langstoeger.at>  [ In article <43B9738E.ECB0C1F@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: D >I have a dual root disk that started off with TCPIP Services 5.0-9." >Things worked fine on both nodes. > F >Then, I upgrade the system disk, from node1 to TCPIP services 5.3.  IE >did not reboot node1. But rebooted node2, only to find that it would  >crash when starting TCPIP.   F What crash ? Why did you upgrade TCPIP ? Did you upgrade VMS as well ?L Is TCPIP V5.3 supported on this VMS version you run ? Did you install ECOs ?  H >After much investigation, it turns out that the configuration files areI >in SYS$COMMON:[SYSEXE] and this seems to be hardcoded in tcpip$config to  >define it in sys$common.   A Config file is there where TCPIP$CONFIGURATION logical points to.  Normally in SYS$COMMON:[SYSEXE]    >Is this normal ?   + For a single archtitecture system disk yes.   L You can't have 2 versions of TCPIP on the same system disk at the same time.L But you can have TCPware and TCPIP on the same system disk at the same time.L Of course you can't run TCPIP and TCPware on the same node at the same time,M but you can run it in the same cluster (on different nodes) at the same time.   E >What I find interesting is that @TCPIP$CONFIG on node1 finds node1's I >info, while TCPIP$CONFIG on node2 finds node2's information.  What "key" F >is used by the software to find the configuration records of one nodeF >versus the other in the same config file ? ethernet address ?  system >root ? (SYS0 vs SYS1) ???   SCSNODE   E >What reason exists to have that file in SYS$COMMON: ? Seems far more & >sensible to have it in sys$specific ?  E Not really. Two possibilites, both with advantages and disadvantages.   D I myself find it much more disturbing to have service directories inD SYS$SYSDEVICE and others in SYS$SPECIFIC. Here would a SYS$COMMON be much more logical...  D >Also, dumping the file shows there are records left from a previousF >temporary incandation of a node when it had a temporary node name. IsE >there a way to remove this ? And if I wish to restart node2's config D >file from scrath, is that possible without zapping node1's config ?  I Change your SYSGEN nodename, change/delete the config and change it back, E or build your own tool to play with indexed files (eg. in DCL) or zap E the (config) file(s) alltogether and reconfig all nodes from scratch.    --   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: Mon, 02 Jan 2006 15:45:27 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 5 Subject: Re: TCPIP Services config files in a cluster , Message-ID: <43B990D3.DC93D817@teksavvy.com>    Peter 'EPLAN' LANGSTOEGER wrote:H > What crash ? Why did you upgrade TCPIP ? Did you upgrade VMS as well ?  F I upgraded to get IMAP, DHCP etc. Although I need to keep a 5.0-9 copyF of DHCPGUI to edit fields which tyhe 5.3 versiosn crashes when editingH those fields. And before I upgrade 5.3 with the ECOs, I need to preserveB all the IMAP stuff because it crashes in subsequent ECOs.  I can'tD complain about IMAP because the made it quite clear that it was madeC available on VAX for our pleasure but isn't supported.  The DHCPGUI H crashing when entering DNS server IPs (or any field that allows multipleB items) is not very acceptable though. Oh, and in 5.3, the help nowG points to a web site that doesn't exist anymoe, whereas in 5.0 the help  is self contained.     N > Is TCPIP V5.3 supported on this VMS version you run ? Did you install ECOs ?  F This is a catch up. I have been through this before with another node.G But that other node hadn't crashed. And when I had upgraded my MV II to = have 5.3 it didn't crash unless I enabled the slip interface.     C > Config file is there where TCPIP$CONFIGURATION logical points to. ! > Normally in SYS$COMMON:[SYSEXE]   G I ended up deleting all the config files and using TCPIP$CONFIG on both B nodes to recreate all the parameters. Now, the second node doesn'tC crash, but after having started, "Internet" shutds down. I do see a H message about insufficient dynamic memory now. So I have a clue of where to go.  K > Change your SYSGEN nodename, change/delete the config and change it back, G > or build your own tool to play with indexed files (eg. in DCL) or zap G > the (config) file(s) alltogether and reconfig all nodes from scratch.     H The config files are not indexed. I tried editing the configuration fileE to delete records belonging to a defunct node, but after that, TCPIP> A complained about not finding the config data, so I deleted it and  started from scratch.     F I can see the point of services and hosts being in SYS$COMMON. But notA so obvious why node specific data such as interfaces should be in 6 sys$common especialyl since there are no tools to edit3 tcpip$configuration.dat to remove unwanted records.    ------------------------------   Date: 2 Jan 2006 21:59:56 +0100 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)5 Subject: Re: TCPIP Services config files in a cluster , Message-ID: <43b9a25c$1@news.langstoeger.at>  \ In article <43B990D3.DC93D817@teksavvy.com>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:I >The config files are not indexed. I tried editing the configuration file F >to delete records belonging to a defunct node, but after that, TCPIP>B >complained about not finding the config data, so I deleted it and >started from scratch.   Check again. ... 3 File organization:  Indexed, Prolog: 3, Using 1 key '                              In 2 areas  ...  --   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: Mon, 02 Jan 2006 16:08:07 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 5 Subject: Re: TCPIP Services config files in a cluster , Message-ID: <43B99622.9B317891@teksavvy.com>    Peter 'EPLAN' LANGSTOEGER wrote: > Check again. > ... 5 > File organization:  Indexed, Prolog: 3, Using 1 key ) >                              In 2 areas     G Mea culpa. I had looked in a small window and the top info had scrolled H past me and I had only remember the "variable length record" portion and assumed it was sequential.  C Ok, so if it is an indexed file I guess it becomes easier to make a - small DCL routine to delete unwanted records.    ------------------------------  % Date: Mon, 02 Jan 2006 23:43:07 +0100 & From: Paul Sture <paul.sture@decus.ch>5 Subject: Re: TCPIP Services config files in a cluster , Message-ID: <41tojsF1ffspqU1@individual.net>   JF Mezei wrote: " > Peter 'EPLAN' LANGSTOEGER wrote: >  >>Check again. >>... 5 >>File organization:  Indexed, Prolog: 3, Using 1 key ) >>                             In 2 areas  >  >  > I > Mea culpa. I had looked in a small window and the top info had scrolled J > past me and I had only remember the "variable length record" portion and > assumed it was sequential. > E > Ok, so if it is an indexed file I guess it becomes easier to make a / > small DCL routine to delete unwanted records.   H Also worth checking is the version number if you have been doing edits. I I know I ran into one area (though I don't remember which version) where  % TCPIP has version ;1 hard programmed.    ------------------------------  % Date: Mon, 02 Jan 2006 10:33:24 -0800 # From: "Tom Linden" <tom@kednos.com>  Subject: Re: tcpip5.4 ? ( Message-ID: <ops2rkhyd6zgicya@hyrrokkin>  9 On 2 Jan 2006 19:27:06 +0100, Peter 'EPLAN' LANGSTOEGER    <peter@langstoeger.at> wrote:   J > In article <ops2rfm4iizgicya@hyrrokkin>, "Tom Linden" <tom@kednos.com>  	 > writes: < >> On 2 Jan 2006 17:53:52 +0100, Peter 'EPLAN' LANGSTOEGER    >> <peter@langstoeger.at> wrote:J >>> PS: Don't forget to also check the existance/protection/owner of the   >>> dir J >>> SYS$SYSDEVICE:[TCPIP$GATED] and/or SYS$SYSDEVICE:[TCPIP$ROUTED] and ifI >>> GATED the content of the SYS$SYSDEVICE:[TCPIP$GATED]TCPIP$GATED.CONF.  >>K >> Well, on the old system I have both, but on the new one, neither. So how  >> do I created the routed?  > @ > With CREATE/DIR/OWNER=/PROT= or with TCPIP$CONFIG of course... > G > You need the SYS$SYSDEVICE:[TCPIP$GATED] for the config file and both E > directories for the log and dump files of the respective processes.  > % > What did TCPIP START ROUTING tell ?  > 5 > -TCPIP-E-CONFLICT or %COB-W-CLOSEDEL or what else ?  > G it started it:-|   I thought it would have started with shut down and    startup  form TCPIP$CONFIG menus.   ------------------------------   Date: 2 Jan 2006 15:20:59 -0800 . From: "Schnootling" <chuckmoore55@hotmail.com>0 Subject: Re: Using VT420 as Alpha System ConsoleC Message-ID: <1136244059.390911.267730@g44g2000cwa.googlegroups.com>    Hello David,   I did as you suggested. I set :  BOOT_RESET = ON  AUTO_ACTION = HALT CONSOLE = SERIAL and INIT-ted, then BOOT-ed.   F  I didn't get a (as in anything) response on the VT420, but the system, did not appear to use the PC monitor (much).  = The, by now familiar, phrase "Alphaserver 4100 Console V6.0-4 G 10-May-2001 10:11:42" appeared on the PC monitor (at the top, on a blue C background), but other than that nothing happend on it. The HP-logo G screen did NOT appear. I'd say the system was waiting for something. (I F did not have the keyboard plugged in as I had unplugged it immediately after doing the BOOT).  D So, after this, I went POWERED OFF+ON back to the SRM prompt (on the. monitor) and found through SHOW * | more that: COM1_BAUD = 9600 COM1_FLOW = SOFTWARE GRAPHICS_TYPE = VIDEO  TTY0_TYPE (?) = VIDEO   B I then set GRAPHICS_TYPE to HARDCOPY and INIT-ted + BOOT-ed. I wasC thinking maybe I could force the monitor to be treated like (say) a B Teletype...and got the lovely HP multiple-logo/logon boxes. (sigh)  9 My thanks for the assistance, though. It was worth a try.  Chuck   , P.S. What would you recommend as a monitor ?A P.P.S. I can't read what it's asking for at initial boot. Is it a C userid (if so, which one -- I don't remember setting up a userid at % system generation, just a password) ?    ------------------------------   Date: 2 Jan 2006 13:01:03 -0800 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>, Subject: Re: VMS MAIL: DIRECTORY strangenessB Message-ID: <1136235663.787629.69840@z14g2000cwz.googlegroups.com>  D This appears to be a MAIL GUI problem, and not just with the message number perhaps. D It may require the large subject lines and such, requirig lots of VM perhaps.B I would try the same command with the old BLISS GUI instead of the current C re-write. D Check out DIR SYS$SYSTEM:MAIL*.* and SYS$LIBRARY:MAIL*.* and take it from there.   E I just tried with 13000 messages and 65000 very short messages and it  works fine. B However... when I tried 100,000 messages mail went CPU bound after reading the file. A Looks like a 16 bit counter problem as  66K messages also failed.   G I hope that when you consider 10,000+ messages in a folder that you RMS D FDL tune the mail file a litte. Give it an allocation of 10,000, and/ extent of 5,000 and  bucket size of 24 or more.   D With the default of 5 block buckets and no extent you will have manyD many fragments and the folder requires dozens of buckets. At 50,000+B each new message will take 200+ read IOs before RMS knows where to write the last new message.   ; To test this i did not actually send thousands of messages. @ I just used a DCL loop to fake it. I think that is close enough. Mail understood :-).? I first mailed one real message with no body, just from/to/subj < Then filed that in a test mail file file, and replicated it,E incrementing a the main key, and counter in the subject. Feel free to C retry with a large subject if you have enough time and diskquota on  hand.    Here is how I did the test:    $MAIL/SUBJ=TEST100000 NL: ME $MAIL  :  FILE TEST TEST Create new? yes    $@MANYMAIL 13000 $MAIL  : 
 SET FILE TEST  DIR/START=999999 :  :    The script follows.  Cheers,  Hein.      $IF p1.EQS."" THEN p1 = 100  $test_file = "TEST.MAI"  $folder = "TEST" $record = "" $records_read = 0  $records_written = 0 $   $OPEN/READ/WRITE file 'test_file% $READ/KEY=&folder/INDEX=1 file record  $read_loop:   $records_read = records_read + 1 $READ/END=write file record  $GOTO read_loop  $  $write: 0 $WRITE sys$output "''records_read' records read" $IF record.EQS."" THEN EXIT 8 $number_loc = F$LOC("TEST1",record) + F$LEN("TEST1") - 1$ $number = F$EXT(number_loc,6,record) $key = F$CVUI(0,24,record) $  $write_loop:& $records_written = records_written + 1 $number = number + 1 $key = key + 1 $record[0,24] = key $ $record[number_loc,6] := "''number'" $write file record $p1 = p1 - 1  $IF p1.gt.0 THEN GOTO write_loop6 $WRITE sys$output "''records_written' records written" $CLOSE file  $EXIT    ------------------------------   Date: 2 Jan 2006 13:14:36 -0800 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>, Subject: Re: VMS MAIL: DIRECTORY strangenessC Message-ID: <1136236476.431996.309480@g43g2000cwa.googlegroups.com>   F Oh, forgot to make clear that there is no actual message number stored with the message. D The message number is a concept maintained by the mail routines, not the file contents.  ? And those stars are of course likely to be an F$FAO feature for 7 printing a number too large for the fields to print it.    Hein.    ------------------------------  * Date: Mon, 2 Jan 2006 22:54:07 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply), Subject: Re: VMS MAIL: DIRECTORY strangeness$ Message-ID: <dpcauf$ulb$1@online.de>  H In article <1136235663.787629.69840@z14g2000cwz.googlegroups.com>, "Hein9 RMS van den Heuvel" <heinvandenheuvel@gmail.com> writes:    F > This appears to be a MAIL GUI problem, and not just with the message > number perhaps. F > It may require the large subject lines and such, requirig lots of VM
 > perhaps.D > I would try the same command with the old BLISS GUI instead of the > current C re-write. F > Check out DIR SYS$SYSTEM:MAIL*.* and SYS$LIBRARY:MAIL*.* and take it
 > from there.   E I should have mentioned that the output was from VAX.  I was thinking G that the Bliss-to-C rewrite (I put that up there with the background of F 1-pixel white gifs on the HP web pages in my list of stupid decisions)G was just on ALPHA, but apparently that's not the case.  However, I just G confirmed that MAIL and MAIL/OLD (IIRC this undocumented qualifier runs G MAIL_OLD.EXE which is derived from BLISS, not C) on 7.3 VAX produce the E same output.  With 7.3-2 ALPHA, the output is also the same.  (Why oh B why was MAIL_OLD.EXE and the MAIL/OLD qualifier deleted with a VMSH upgrade on ALPHA even though the problems introduced by the rewrite wereD not solved with the same upgrade?  It's embarrassing enough that theG rewrite was so bad that MAIL/OLD and MAIL_OLD.EXE had to be introduced, G and utterly inexcusable that they were dropped even though the problems C remain.  Can we please have MAIL/OLD and MAIL_OLD.EXE back in 7.3-2  ALPHA via a MAIL ECO?)    G This is all with the normal text-based default MAIL interface.  I never @ use it otherwise, but I wanted to try the DECwindows interface. G However, I don't know how to tell it to open a file (drawer?) which is   not in the main MAIL directory.   I > I hope that when you consider 10,000+ messages in a folder that you RMS F > FDL tune the mail file a litte. Give it an allocation of 10,000, and1 > extent of 5,000 and  bucket size of 24 or more.   F This folder normally just archives stuff and is rarely read.  This is  what it looks like now:   8    Allocation: 17527, Extend: 15, Maximum bucket size: 5  E Is just the number of messages something to consider when tuning, or  E would the message size need to be considered as well?  What would be  H really nice would be a TUNE_MAIL_FILE.COM which would do the tuning you  mentioned.  :-)   F > With the default of 5 block buckets and no extent you will have manyF > many fragments and the folder requires dozens of buckets. At 50,000+D > each new message will take 200+ read IOs before RMS knows where to > write the last new message.   H Despite the small gripes mentioned above, I will probably stick with VMSD MAIL until I die.  I do notice some performance problems, though.  IH haven't done any RMS tuning (although I do a MAIL> COMPRESS from time toH time if I have substantially reduced the size of a file and don't expect; it to grow much again soon).  What I have done is create 22 D subdirectories in my main mail directory (one for each letter of theH alphabet for which a mail file exists which starts with that letter) andC distributed the 128 files among these directories.  There are about C 42000 message files altogether.  This keeps the number of files per , directory down to a more reasonable number.    ------------------------------   End of INFO-VAX 2006.005 ************************                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  h
l6hשv4-S!W~R	r1r&^!l2gk)]2^"Oe>
+
Ο&at)al+]Ͱ2|q:x}Ya?&3Y_R{xNȓ6Bi#$l626*q:xvI/aζo<?l\6N~q~!N43L^_Z	MW0!naHBX_})A 㮡tcW93D lZt7iR6:
MlUβŌ^V96*gq:!ln_lZLϸIBaڭjKLXh-{M?%4]oo6ہ7߮R6*=l&琄˨0$3|E2..$a(
(G8|b݁-t6atp[,MCG٭WPvSF%,4]\F	Ce//#KŅC$]e!ttu0,2"LakO.]Ї+at6	ۡoy<-?u,4];5UJ8lbm¯;6H|X!.M3!tG`#̳Л `;0skHw0¼bFw^oĵqԵ䩏ڸI8\#:ۑsЈ7Xl֝`;BoAx`;d0f
FNr=p1[!r
'~9H^!Փ6vca؅iza1XG8fJ`Tgkp9oԏzfҀO\֍$/Q;f"l4]\!02Z]߄ў|.$)a-t0QeJhl0<70Kh^.o\|bWeu
uc{wf[\	k)[m\ tɟt0(gatھ 9gaHB42,ƛw	[՘6BoX
ޅIFaN?wl3u"[~;O+[xF^B?w%K6=wϰ;/FN!aߟ+/)3AUXw[?OMYi%~Fؒ6=Ǐ0l1~oMAث1L^^tY>g[~ >Qp!s!-F7 lVSl^l-`gp6akVWT\5sHzFdu186Lq:;2~zۄ.MCOh0,FNe~N
3~&a-S|h
3Ԯƥۮ9vll+}($,(K6kHgYn)tglOrm)޷'lM?aSS7qEl`EM3-
6NZ&l;2?s_[<	;ӷyN0~3}1>f l:4Ű͓q:a{ c8t<atp@.1lwq:~5LxӡIm{
nie`ߕ¦d߹fN]ex`;6	[u7,'kS6M׽٨iU6:ct`!϶IH0lf
a{p׸L3l2'lOM:fO	#Mݔჲp0l{ƚf6kHI8F!̈́-	aSb،|w쿡CtGL
]z@7y^v?Qjk6!q!~l`͗`kؐҩS6pBޓq$[az6v^݄y f6a[3lo[ l#¶omq:xGw:[Xf#Ҫla4/wD	s¦(j"&l#20Y t|0?iV̰f#}Sg{a:۔uC`F	[?aLgKyFe5@شzncLvh
GcX=r~3q:x,xW}^=6N6v1lv!lnOۖ6Nvov'at
5:,M+f3Pg
et|o@4߻ǰɄloa;7,Ag!vlڬ!:edFYB>*`"l.aq0l2&l^=9VϷ6aty2a{a<)f'o6Nϓ3:[#̓f̓-ec	]<On"Ii#˵>=$yii8=1NO!gw
F?3eZJVfiAM:{0,@ȫ=ҬG6W\Q:.g#Nl?Wު?'e~q·ҭRq^.ΫzGonzFh;?-aSaL=sq:x_a[i_sp6"?u`OuvI鷃`'ɟm֐͆ҁWf4
c׽O5w~ұ
F]&٢6~	f.-]{c*pL
mC2N5#z{USũwJӎcEz[ƵH*?NU{q\ZG4.
s}I??s˗[v2c|&"~ii9?$\FXD	1&VwCVIJoRQ;hj8;Ĭ	'2)Rēi"L.ouZfbrt<+7~yڃ4!OÏ'h?W//ٖ5&glu8Qe}:s)|ĩNnj3ðV|lٴ:{)fg_͏O6j
G2'	1|gOgƥ\~a%ONӚ
kɷ+ˣ5C*BO 	^jiN
a
Iض5{nfBhCk]\jrDZc}	3看LD-PG\<
զ#G`(dC|E^ZgmpŞP2rTk뻶N2pxRgMñ.oÈsYB5;
Kvy(dFAU'-O6#vI+3ȤG6#Msm=|Yk49--NFZ[yIe[y&v9	Ĝ[|)a