1 INFO-VAX	Sun, 24 Sep 2000	Volume 2000 : Issue 534       Contents: Re: AlphaPC164 & decwindows , Re: Deleting 250,000+ files from a directory, Re: Deleting 250,000+ files from a directory% Re: F$GETSYI for multiple processors? # HELP ! VMS License Units Question - ' Re: HELP ! VMS License Units Question -  Re: Incremental backup blues Re: Is VMS worth learning....? Re: Is VMS worth learning....?3 Re: Memo:  Deleting 250,000+ files from a directory 3 Re: Memo:  Deleting 250,000+ files from a directory % Re: No more SYSDUMP.DMP? No more NTP? $ Re: Problem with COMPAQ 9.1GB drives$ Re: Problem with COMPAQ 9.1GB drives  reading VMS files with Delphi...4 Re: Replacing system disk, will licenses still work? Re: SRAM Corruption Thread Re: Suggested VMS new feature! Re: Suggested VMS new feature! RE: Suggested VMS new feature! Re: Suggested VMS new feature!! RE: Sun Hardware problems persist & Re: Unusual hard disk activity at boot VMS and UNIX Re: VMS and UNIX Re: VMS and UNIX Re: VMS and UNIX Re: VMS and UNIX Re: VMS Memory Tuning  Re: VMS Memory Tuning   F ----------------------------------------------------------------------  % Date: Sat, 23 Sep 2000 17:07:03 -0400 ) From: yyyc186.illegaltospam_@flashcom.net $ Subject: Re: AlphaPC164 & decwindows9 Message-ID: <39cd1c0a$1$lllp186$mr2ice@news.flashcom.com>   1 In <8qf65b$3b0$1@mordred.cc.jyu.fi>, on 09/23/00  :    at 05:07 PM, Osmo Kujala <kujala@kanto.cc.jyu.fi> said:  + >yyyc186.illegaltospam_@flashcom.net wrote:   I >> My AlphaPC164 has a Diamond PCI based video adapter in it.  Decwindows 1 >> refuses to recognize it as a graphics adapter.  >.... % >> What Graphics adapter do I need...  >.... K >> command file for an ATI Mach64.  I can pic one of these cards up if need  >....   F >S3 trio64 cards work, Diamond's 64 too. (You didn't say which DiamondG >model you have.) ATI Mach64 is VMS-supported, so it should work too. I I >haven't tested. When I tried Diamond S3 trio64, VMS didn't see it first, E >but after io autoconfiguration it showed up and DECwindows worked. I E >guess I haven't done all the necessary configuration for DECwindows?   J Mind had a Diamond with S3 chipset and Stealth64 v3.02 BIOS.  No amount ofH auto configing I did made it work...then again I'm a software developer,G not a hardware developer or systems manager.  I have purchased the Elsa I card from Island and am having same difficulty.  Trying fresh VMS install  now with Elsa card.   H What autoconfig command did you use to get Decwindows to see the device?   Roland   --  ; ----------------------------------------------------------- D yyyc186@flashcom.net              To Respond delete ".illegaltospam"6                             MR/2 Internet Cruiser 1.528                             For a Microsoft free univers; -----------------------------------------------------------    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:47 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>5 Subject: Re: Deleting 250,000+ files from a directory 7 Message-ID: <200009231631_MC2-B46D-12D8@compuserve.com>   J         If you are going to make a directory listing into a command file,=   you probably want to do:& $ DIR /NOHEADER /NOTRAILER /COLUMN=3D18 particularly, if you are going to sort the list somehow.  % Message text written by Chris Scheers  > =    Jim Agnew wrote: > =   J > you can do a dir/size:all of the dir, sort it in reverse order, trim of= f   the extraneous info in a editor,H > and turn it into a delete com file....  for a one-shot deal like this,< that is probably the fastest solution you can comeup with...J > now, it won't run as fast as dfu, but the simplicity of it and the fact=  7 it can run in batch w/o your attention, gives it a tiny  > peice of merit..    1 Out of curiosity, why do you specify "/size:all"?   E If the intent is to get the listing into a single column, you can use J "/columns=3D1".  This eliminates the need to trim the size from the end o= f @ the line.  It also eliminates the time required to read the fileG attributes to get the size, which would be considerable for 250K files.  <    ------------------------------  # Date: Sat, 23 Sep 2000 21:32:25 GMT + From: "Matt Jacobs" <mjacobs1@twcny.rr.com> 5 Subject: Re: Deleting 250,000+ files from a directory : Message-ID: <Jj9z5.22009$O37.3605720@typhoon.nyroc.rr.com>  J The DFU utility worked very well.  I deleted 136,000 files in less than an# hour.  Thanks to all who responded.   
 - Matt Jacobs    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:46 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>. Subject: Re: F$GETSYI for multiple processors?7 Message-ID: <200009231631_MC2-B46D-12D6@compuserve.com>   / Message text written by INTERNET:stan@stanq.com H >Is there is a parameter that I can specify to F$GETSYI to find out if =  2 the system is single-processor or multi-processor?   <   B         Yes, there is.  Two of them, in fact.  One to get the CPUs9 available and the other to get the number of active CPUs.   F         Go thou and RTFHF!   Or the "System Services Reference Manual"J (title approximate) has descriptions of all the things that SYS$GETSYI ca= n  return.   7         Oh, all right!  AVAILCPU_CNT and ACTIVECPU_CNT.    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:24 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>, Subject: HELP ! VMS License Units Question -7 Message-ID: <200009231630_MC2-B46D-12CB@compuserve.com>   8         I believe that "units" are completely arbitrary.  J         The reason for the difference between the VMS Base License and th= e J DECwindows license is that they are a different class of license.  The VM= S J Base would be Class A or Class D, I believe, while the DECwindows would b= e ! Class H (Alpha Layered Products).   F         Confused?  So is everyone.  I seem to recall that a VAX 11/750H required more license units for layered products  than many systems thatJ you could tuck underneath your arm and that would blow the doors off it!!= ! =   J  All the VAXstations after the VAXstation 1, from the least to the greate= stJ used a class D license of 100 units, a class E license of 50 units  and a=  J Class F (Layered Products) license of 10 units.  That covered a range of = .90 to 40 or 50 VUPS.   All supported a single user.  J         A SHOW LICENSE /CHARGE should show you the license requirements f= o / each type of license supported on your machine.   " Message text written by "Islandco"H >Can someone kindly advise me of the "actual" meaning behind the OpenVMS base license "units"   H The Alpha PWS500 I have in here has OpenVMS base (QL-MT1AE-6X) 12 units, but H a Customer said he needed 15(Ql-MT1AE-63) to make DECwindows work on his	 APWS500au   J Now I have loaded DECwindows many many times and have never had a problem=     Any info would be appreciated    Am I missing something ???  2 Kindly email me with info at dbturner@islandco.com <    ------------------------------  # Date: Sat, 23 Sep 2000 21:56:16 GMT * From: Paul Walker <walkerp1@earthlink.net>0 Subject: Re: HELP ! VMS License Units Question -- Message-ID: <39CD287B.1F18E939@earthlink.net>   L Someone once told me that license units were influenced by DEQ's anticipatedO support costs; therefore, units for older systems like 11/750's might be higher B since there are less of them now to spread the fixed costs across.   Paul   "Richard B. Gilbert" wrote:   : >         I believe that "units" are completely arbitrary. > H >         Confused?  So is everyone.  I seem to recall that a VAX 11/750J > required more license units for layered products  than many systems thatL > you could tuck underneath your arm and that would blow the doors off it!!!   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:44 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>% Subject: Re: Incremental backup blues 7 Message-ID: <200009231631_MC2-B46D-12D5@compuserve.com>   =         You can still do incremental or differential backups!   D         Just note the time you broke the shadow sets and, instead of /SINCE=3DBACKUP, use /SINCE=3D'TIME_OF_SPLIT'!     0 Message text written by INTERNET:_RAQ_@yahoo.comH >The way we do backups today is to shut down the DB's, splitt the shadowE sets, restart the DB's, and then let the backup run on the "inactive" 2 shadows set members. This gives us alot of uptime.  F Problem is, we cant do incremental backup, since the recording pass isC done to the "inactive" disks. When remounting the shadow member the C former "inactive" member with the recording pass gets overwritten -  fair enough.   How could this be improved?  <    ------------------------------  # Date: Sat, 23 Sep 2000 20:36:20 GMT ! From: "dls2" <dlshearer@home.com> ' Subject: Re: Is VMS worth learning....? < Message-ID: <8v8z5.13265$ib7.1150218@news1.rdc1.nj.home.com>  0 "William H." <t18_pilot@hotmail.spam.com> wrote:A > VMS *is* indeed behind the scenes in many places.  For example, D > I contracted at Hoffman LaRoche last year, and 65% of their Oracle> > ran on OpenVMS, the balance ran on UNIX, and a small portion@ > was on NT.  (When I asked about the low NT usage, the database9 > department said they simply didn't trust NT enough yet)   > It is not surprising that they wouldn't trust NT "enough yet".  D > If you're into finance, you'll be happy to know that lots' of wall@ > street applications are on VMS.. the bank I work for transfersF > a trillion bucks a week over the wire (to the Fed, other banks, etc)D > over VMS.  They wouldn't do it any other way.  (we have 150 vaxes)   150 vaxes, and how many alphas? % Or is alpha not trusted "enough yet"?   @ I can't help but wonder, considering the threads on SRAM latchup> and Sun hardware, if DEQ's newer alphas are as reliable as the) older VAXen, for underlying platforms....        --  Derrick Shearer    ------------------------------  % Date: Sat, 23 Sep 2000 17:15:31 -0400 ) From: yyyc186.illegaltospam_@flashcom.net ' Subject: Re: Is VMS worth learning....? 9 Message-ID: <39cd1daa$2$lllp186$mr2ice@news.flashcom.com>   - In <39CC19A7.E066A629@MPCM.com>, on 09/23/00  1    at 05:15 PM, Matt Morley <Matt@MPCM.com> said:   H You can pick up a VMS Alpha box for under a grand on Ebay.  Won't be theJ fastest thing in the world and probably won't come with a monitor, but youI should be able to scrounge up a VGA monitor for a workstation pretty easy 
 at a college.    Roland   >Is VMS worth learning....  @ >That's a loaded question. Especially since this is a vms group.  I >I've been reading up on vms manuals and it seems like it would be fun to + >learn, but what is the state of VMS usage?   3 >What roles does VMS fulfill that other OS's don't?   9 >And where, besides expensive classes could I learn this?   J >I have a feeling that I would be better setting up a box and hacking awayJ >with it, until I understand it. But not knowing it, I going to figure out@ >hardware... and cheap at that (us college students struggle for >hardware).   ! >I have joined decus... now what?    >--  >Matt Morley (MPCM)  --  ; ----------------------------------------------------------- D yyyc186@flashcom.net              To Respond delete ".illegaltospam"6                             MR/2 Internet Cruiser 1.528                             For a Microsoft free univers; -----------------------------------------------------------    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:40 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>< Subject: Re: Memo:  Deleting 250,000+ files from a directory7 Message-ID: <200009231630_MC2-B46D-12D3@compuserve.com>   F         This approach would involve 250,000 image activations which, I3 believe, would be even slower that DELETE *.*;*!!!!   E         Instead, construct delete commands with as many filenames, in A descending alphanumeric order, as will fit in the 1024 characters J available.  With a bit of luck you should be able to pack at least 40 fil= esC per command which will cut the image activations by a factor of 40.   % Message text written by Paul BEAUDOIN  >Try this|: @ Get the directory listing to a file and sort it in reverse order (alphabeticaly (z-a)) : Convert each file name to a delete command and execute it.C Should be much faster as you are now deleting bottom to top and the ' directory file does not have to shuffle % the remaining entries around so much. @ Good luck  - and don't forget to kill the errant user/programme<   ------------------------------   Date: 23 Sep 2000 23:08 -0400  From: hein@eps.enet.dec.c*m < Subject: Re: Memo:  Deleting 250,000+ files from a directory& Message-ID: <23SEP200023085237@miasys>  n In article <200009231630_MC2-B46D-12D3@compuserve.com>, "Richard B. Gilbert" <DRAGON@compuserve.com> writes...G >        This approach would involve 250,000 image activations which, I 4 >believe, would be even slower that DELETE *.*;*!!!!  C Nah... never underestimate the pain of shuffling down a 1000+ block " directory files a block at a time.    F >        Instead, construct delete commands with as many filenames, inB >descending alphanumeric order, as will fit in the 1024 charactersL >available.  With a bit of luck you should be able to pack at least 40 filesE > per command which will cut the image activations by a factor of 40.   B I wrote that once. It helped some, not much. At that time I tried , only 200 bytes. Easy enought to up hat some.  & Attached a basic reverse-delte script.( Watch out... works on current directory!   hth, 	Hein    $if p1.eqs."" then goto help- $directory/out=sys$Login:delete.tmp/col=1 'p1 H $sort /key=(pos:1,siz:-1,desc) sys$Login:delete.tmp sys$Login:delete.tmp5 $open/read/error=clean_up sorted sys$Login:delete.tmp  $init: $read/end=clean_up sorted f  $l2 = f$len(f)% $if f$loc(";",f).eq.l2 then goto init 
 $files = f $loop: $read/end=done sorted f  $l1 = f$len(f)% $if f$loc(";",f).eq.l1 then goto loop 
 $if l2.lt.200  $  then  $    files = files + "," + f $    l2 = l2 + l1 + 1  $  else  $    delete 'p2' 'files' $    files = f $    l2 = l1 $endif
 $goto loop $done: $delete 'p2' 'files'
 $clean_up: $close/nolog sorted  $delete sys$Login:delete.tmp;,;  $exit  $help: $type sys$input      Usage:  M   P1 (required) = wild card file spec, with optional DIR selection qualifiers >   P2 (optional) = delete command qualifiers (/LOG or /CONFIRM)   $exit    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:31 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>. Subject: Re: No more SYSDUMP.DMP? No more NTP?7 Message-ID: <200009231630_MC2-B46D-12CE@compuserve.com>   F         If TCP/IP starts before DECnet phase 4, DECnet will change theJ Ethernet address and all sorts of strange things may happen on the TCP/IP=   side.   J         If you start DECnet in batch, ensure by whatever means necessary,=  : that TCP/IP does not start until after DECnet has started.     Message text written by Cthulhu , >Peter LANGSTOEGER <eplan@kapsch.net> wrote:  ( >>And SYS$SPECIFIC:[TCPIP$NTP] is there. > What's in it ?  < LOGIN.COM, TCPIP$NTP.CONF, two .LOG files, the sample .CONF.  7 >>Maybe that try to contact 127.0.0.1 is too premature? * > Don't think so. NTP is started this way.  D Even now that a Kind Soul posted me TPCIP 5.0A, it is doing the same thing.F And OSU-HTTPD is almost doing the same: at the first launch it doesn't- start, generating a PTHREAD*.LOG in WWW_ROOT. C I discovered that a little PERL program launched just before HTTPD, F that simply open and then close a socket somewhere, solve this problem: at boot... maybe it is a timeout problem on a slow system?  B Could be de DECnet (Phase IV) subsystem launched as a batch job be implied in some way?<    ------------------------------  % Date: Sat, 23 Sep 2000 16:30:38 -0400 2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>- Subject: Re: Problem with COMPAQ 9.1GB drives 7 Message-ID: <200009231630_MC2-B46D-12D2@compuserve.com>   G         What's missing is the "forced error flag".  Delete the affected G files.  Run ANALYZE /DISK again just to be sure.  If it has not already J done so, it should assign the bad blocks to BADBLK.SYS.  Restore your fil= es from backup or reinstall.   J         DEC raid controllers; e.g. the HSZ40/50/70/80 have the ability to=  J simulate the forced error flag but they are a little expensive for casual=   use.  ' Message text written by Bill McLaughlinhG >Hello all. I purchased a COMPAQ (Quantum, I think) 9.1 GB drive for anr alpha =i  J in April. This drive failed last month and COMPAQ sent a replacement. Onc= e =r  J the data was restored, the drive seemed to work OK. However, during a lar= ge  J software installation, the drive loged a few parity errors. There appear = to  I be some bad blocks. COMPAQ tells me that "drives usually have a few bad =a  + blocks" and that I should work around them."  J Since there doesn't seem to be the usual bad block flagging and replaceme= nt  E mechanism for local SCSI drives, I am not sure what to do. COMPAQ's =o  J suggestion was to leave this old file there and create a new one. The fil= e =a  @ is very large, and I don't feel this is a valid solution anyway.  , 1) Should I keep this drive or send it back?8 2) Has anyone else had problems with these 9.1GB drives?   BTW, here is the error info:   $  anal/disk/read dka200:l( %ANALDISK-I-BBLHEADER, file (8583,3,1) =  & [ORACLE8.816.RDBMS]LIBSERVER8_64.OLB;1%         contains suspected bad blockse% %ANALDISK-W-READFILE, file (8583,3,1) & [ORACLE8.816.RDBMS]LIBSERVER8_64.OLB;1         error reading VBN 59135  -SYSTEM-F-PARITY, parity error% %ANALDISK-W-READFILE, file (8583,3,1) & [ORACLE8.816.RDBMS]LIBSERVER8_64.OLB;1         error reading VBN 59215  -SYSTEM-F-PARITY, parity error< %ANALDISK-W-DELHEADER, file (10418,1,1) ORACLIENT_V816.EXE;1         marked for deleteo> %ANALDISK-W-DELHEADER, file (12577,1,1) ORACLIENT64_V816.EXE;1         marked for delete, <r   ------------------------------  % Date: Sat, 23 Sep 2000 18:43:59 -0400-) From: Richard <dr_vmsREMV@techieREMV.com>5- Subject: Re: Problem with COMPAQ 9.1GB driveso2 Message-ID: <ADLNOSSnuMAak4aio7E3TtAM8Am1@4ax.com>  9 On Fri, 22 Sep 2000 21:32:09 +0000 (GMT), Bill McLaughlinn <mcbill20@hotmail.com> wrote:o  G >It's an RZ2DC-CA, ALphstation 500/266, internal. (I'd love a desk-side 6 >Storageworks enclosure, but they are VERY expensive.)  N There was a BA-356 on eBay that could work for you for $250.  You just need to look around.  :)   ------------------------------  + Date: Sat, 23 Sep 2000 18:55:04 -0500 (EST)o* From: Pedro Viena <contracer2@hotmail.com>) Subject: reading VMS files with Delphi...i5 Message-ID: <F283t859I09Ek8fIpWI00001ac6@hotmail.com>a   Hello guys:e  A Is there any way to read VMS files using Borland Delphi in a PC ?    Thank you in advance...lI _________________________________________________________________________eI Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.   D Share information about yourself, create your own public profile at  http://profiles.msn.com.   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:26 -0400I2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>= Subject: Re: Replacing system disk, will licenses still work?-7 Message-ID: <200009231630_MC2-B46D-12CC@compuserve.com>o  J         The DECpaq licenses are not tied to a specific Ethernet address o= r2J machine serial number.  The only reason the licenses would not work would=  J be if you upgraded your processor to one requiring more license units tha= n 	 you have.2  J         You could look in OPERATOR.LOG for the error messages.  If you go=  J through your license database with LICENSE /LIST /FULL and make a note of=  J the expired license which have not been disabled and then disable them, y= ou9 should be rid of most, if not all, of the error messages.s  & Message text written by Charles GilleyJ > I want to cold-spare a system drive on a VAX4300.  I make a stand-alone=   backup =  J tape of the existing system image and blast it onto a replacement disk.  = I =   5 would think that the licenses should still load, yes?i  J Under what hardware circumstances would the licenses not load, replacemen= tB ofJ motherboard?  Network card?  From unix-land, many licenses like to tie to=   ther, physical network address.  Does VMS do this?  @ Lastly, when my machine boots, I see a number of license manager complaints.t Are these logged anywhere? <e   ------------------------------  # Date: Sat, 23 Sep 2000 18:18:26 GMTh% From: hg/jb <shsrms@bellatlantic.net>l# Subject: Re: SRAM Corruption Thread-0 Message-ID: <39CCF446.E14021C6@bellatlantic.net>   Jerry Leslie wrote:s > ( > hg/jb (shsrms@bellatlantic.net) wrote: > : Interesting!!a > : Very good Jerry, thanks!!0 > : bob7 >  > You're welcome.< > E > It could explain the problems SUN (allegedly) has with their E10Ks.mD Exactly my thought!  there had been some work on this in years past,F and some speculation that the chip densities could show up the effects of normal particle hits.< The effects would certainly look like totally unexplainable ; phenomena until sufficient data points could be correlated. ; Static ram chips would certainly show this phenomena beforen9 dynamic chips - especially in systems with autocorrectinge- capabilities!  Really is good timing on this!k bobe   > 6 > --Jerry Leslie     (my opinions are strictly my own)   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:36 -0400h2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>' Subject: Re: Suggested VMS new feature!l7 Message-ID: <200009231630_MC2-B46D-12D1@compuserve.com>t  D         It's not at all difficult.  You can do it yourself!  Write a> substitute for DELETE.EXE that does what you want, store it inF SYS$COMMON:[SYSEXE] and you're done.  I beleive that there are libraryH routines to delete and/or rename files if you don't want to get into theJ RMS calls and data structures.  Can't manage it yourself?  Want to pay me= ?   J         Frankly, I think making regular backups is a lot easier and almos= t 4 as effective.  Defining a symbol in my LOGIN.COM:  $: DEL*ETE:=3D=3DDELETE/CONFIRM has saved my a** a few times.  J         FWIW, there are several undelete utilities available.  If you use=   them quickly enough$ they work just fine.  $ Message text written by "John Nixon"G >I bet none of you will ever guess why I am asking for this new featuren now! ;-(   J How difficult would it be to inocorporate an undelete utility within VMS?=   WithE the huge disk drives that are coming available, it seems to me that a  wastebasketiF concept, where all "deleted" files are temporarily deposited, would be	 extremelysJ usefull.  Especially in light of the VMS propensity to use the last block=  	 freed forE the next disk space allocation.t  J The utility could allocate a percentage of the disk for deleted files.  A=   showH device command wouldn't display this space as available.  There could beF various ways of emptying the waste basket, but the point would be that rightgF after you hit the delete key, and then utter "Oh  Sh*t" , you would at leastn have! a prayer of retrieving the file.<h   ------------------------------  # Date: Sun, 24 Sep 2000 01:02:01 GMTN/ From: "John Nixon" <jorlnixon@worldnet.att.net>o' Subject: Re: Suggested VMS new feature!aF Message-ID: <docz5.4534$tl2.343539@bgtnsc07-news.ops.worldnet.att.net>  G Well,  I appreciate all the comments, but if I wanted a "roll your own"t operating system, ( I would probably become a unix advocate.  H And I disn't mind suddenly stopping production, dismounting the disk andD using a third party undelete tool, I would probably become a M$shaft windoze advocate.i  L But, I am stuck being a VMS bigot, and I would like such a simple feature to beI incorporated in my vanilla OS.  I am quite sure I am not the first one to- accidentallyI delete a file. Even using a delete command procedure that does a confirm,m and then asks if you are really sure.  L I am sure there also some very fine shelving products that will continuously moveF files to near-line storage, but I dont' think continuous backup is the solution either.  L I think that a simple "undelete" feature that actually gives you a prayer of
 retrievingK a file, even after a period of time elapses, and without shutting down your  system4 would be a very good feature for VMS to incorporate.    : "John Nixon" <jorlnixon@worldnet.att.net> wrote in message? news:5gLy5.1782$6T.157328@bgtnsc05-news.ops.worldnet.att.net...0H > I bet none of you will ever guess why I am asking for this new feature now! > ;-(U >sK > How difficult would it be to inocorporate an undelete utility within VMS?p > WithG > the huge disk drives that are coming available, it seems to me that a 
 > wastebasket0H > concept, where all "deleted" files are temporarily deposited, would be > extremelyoK > usefull.  Especially in light of the VMS propensity to use the last blockr > freed forM! > the next disk space allocation.  >hK > The utility could allocate a percentage of the disk for deleted files.  Ar > showJ > device command wouldn't display this space as available.  There could beH > various ways of emptying the waste basket, but the point would be that rightlH > after you hit the delete key, and then utter "Oh  Sh*t" , you would at leastN > have" > a prayer of retrieving the file. >  >    ------------------------------  % Date: Sat, 23 Sep 2000 21:36:03 -0400-# From: John Vottero <John@MVPSI.com>-' Subject: RE: Suggested VMS new feature!:D Message-ID: <C15945A9D9EFCF11BA8B08002BBF1CCC0CD81D@berry.mvpsi.com>  J This whole "UNDELETE Utility" idea is silly.  It's silly because, about 10K minutes after a slick soft delete/wastebasket utility is added to VMS, some  bozo is going to ask:t  D  "How do I recover a file after I've done a EMPTY/ERASE/DO_IT_NOW ?"   :)   > -----Original Message-----6 > From: John Nixon [mailto:jorlnixon@worldnet.att.net], > Sent: Saturday, September 23, 2000 9:02 PM > To: Info-VAX@Mvb.Saic.Com.) > Subject: Re: Suggested VMS new feature!6 >  > @ > Well,  I appreciate all the comments, but if I wanted a "roll  > your own"s > operating system,s* > I would probably become a unix advocate. > > > And I disn't mind suddenly stopping production, dismounting  > the disk andF > using a third party undelete tool, I would probably become a M$shaft > windoze advocate.i > = > But, I am stuck being a VMS bigot, and I would like such a t > simple feature to  > be? > incorporated in my vanilla OS.  I am quite sure I am not the n > first one to > accidentally< > delete a file. Even using a delete command procedure that  > does a confirm,m
 > and then > asks if you are really sure. > = > I am sure there also some very fine shelving products that l > will continuouslyd > moveH > files to near-line storage, but I dont' think continuous backup is the > solution either. > ? > I think that a simple "undelete" feature that actually gives s > you a prayer of  > retrieving; > a file, even after a period of time elapses, and without r > shutting down your > system6 > would be a very good feature for VMS to incorporate. >  > < > "John Nixon" <jorlnixon@worldnet.att.net> wrote in messageA > news:5gLy5.1782$6T.157328@bgtnsc05-news.ops.worldnet.att.net...a? > > I bet none of you will ever guess why I am asking for this ,
 > new feature  > now! > > ;-(s > >D: > > How difficult would it be to inocorporate an undelete  > utility within VMS?a > > With@ > > the huge disk drives that are coming available, it seems to  > me that aT > > wastebasket17 > > concept, where all "deleted" files are temporarily . > deposited, would ben
 > > extremelyr? > > usefull.  Especially in light of the VMS propensity to use   > the last block
 > > freed ford# > > the next disk space allocation.  > >e< > > The utility could allocate a percentage of the disk for  > deleted files.  Ah > > show> > > device command wouldn't display this space as available.   > There could be= > > various ways of emptying the waste basket, but the point c > would be thati > right > > > after you hit the delete key, and then utter "Oh  Sh*t" ,  > you would at > least  > > have$ > > a prayer of retrieving the file. > >u > >o >  >  >    ------------------------------  % Date: Sat, 23 Sep 2000 21:00:22 -0500 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net>a' Subject: Re: Suggested VMS new feature!h, Message-ID: <39CD6036.3B54890@earthlink.net>   John Nixon wrote:- > I > Well,  I appreciate all the comments, but if I wanted a "roll your own"o > operating system,u* > I would probably become a unix advocate.   IMHO...e  @ If you think about it, VMS is _THE_ premier "roll your own" o.s.  F You get the basics, an excellent foundation. What you build with those pieces is up to you.  C All the various "stuff" you "get with" UN*X is the result of othersa: rolling THEIR own, for the lack of the pieces they needed.  B Perhaps my rather long post really would have been better directed: toward VMS Engineering, with a general "Cc:" to the group.   -- g David J. DachteraN dba DJE Systems. http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/t  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.s   ------------------------------  % Date: Sat, 23 Sep 2000 19:26:01 -0400h+ From: "Main, Kerry" <Kerry.Main@compaq.com>-* Subject: RE: Sun Hardware problems persistJ Message-ID: <910612C07BCAD1119AF40000F86AF0D805284802@kaoexc4.kao.dec.com>   Andrew, Andrew ..i  & Feeling a bit warm these days are we ?   :-)_   re: your quote on eTrade ..   K >>> But eTrade did not have one outage associated with this issue, they hadn( a number each of which lasted hours. >>>  L Yep, a long time ago (Feb 99 timeframe) they went through some configurationI issues over the total space of about 4-5 days. No question. No vendor canD+ say they have NEVER had hardware problems. d  E >>> So you have a number of outages at eTrade which according to your 5 information was caused by Compaq and the 22 hour eBayrD outage which was not caused by Sun hardware, software or engineers.   H I suspect the issue is not with the single 22 hour outage, but the many,> many outages eBay has experienced over the last 12-18 months.   " Are the problems all Sun related?    Nope.   K Are a good deal of them ? Don't know, but press articles do offer some very H interesting insight. As I recall, a number of the posted articles quotedL eBay spokespeople as stating cpu hardware or database server issues were the problem.  H Anyway, this subject of the cache issues has likely started to bore many	 readers. k  L The bottom line is that, imho, when faced with hardware (or software) issuesK that impact many Customers, it is the responsibility of the vendor to admit,7 it with a formal statement, then fix it, then move on.    K Where vendors get in trouble, is when they find a workaround (like loweringmJ the temperature in a room where no other vendors are having problems) thatI reduces or hides the problem, but then try and suggest to their Customersm& that this workaround is the final fix.   Regards,  
 Kerry Main Senior Consultant,
 Compaq Canadal Professional Servicesu Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.com9       -----Original Message-----7 From: andrew harrison [mailto:andrew.nospam@uk.sun.com]c( Sent: Friday, September 22, 2000 6:30 AM To: Info-VAX@Mvb.Saic.Com'* Subject: Re: Sun Hardware problems persist     Jordan Henderson wrote:c > = > Yes, I've heard the explanation... Now, enlighten us.  What < > _exactly_ was the patch that was not available for "over a; > year"?  When was it first made available and what problem0< > did it address?  Most patches have a URL, where is the URL< > for this one?  I think you'll find that these details have > never been revealed. > 5 So you heard the explanation but thought why believe e4 Sun, eBay and the two software vendors all of whome 1 say exactly the same thing. Instead you choose tou5 believe the rumour and FUD which as you must realise s have no actual basis in fact.t  > > Who is protected by not revealing this information?  Indeed,9 > other customers might be similarly exposed and might be3! > interested in this information.  > 5 The 3rd party supplier and eBay are protected. It is t2 no advantage to Sun to keep the information quiet.  ; The fact that you choose not to believe Sun, eBay, Veritas  = and Oracle and instead choose to believe sources that are at n/ best biased against Sun says volumes about you.e  = > More convenient finger pointing.  It could all well just bes > a cover story. > 6 You of course a entitled to your opinion, sad for you 7 that very few people outside the confines of this groupm hold the same views.    B > > Remember eTrade, what happened there, was it a Compaq engineerD > > doing the wrong thing and a bug in the OpenVMS cluster software,B > > who knows, neither Compaq or eTrade have been very forthcoming: > > about what really caused eTrade to have their outages. > < > I heard that a Compaq engineer had improperly put a lot of> > non cluster traffic on a Fiber Cluster Interconnect.  SoundsC > reasonable.  Case closed.  And, we don't hear continued instanceseE > of downtime amounting to many dozens of hours like we do with eBay.  > 7 But eTrade did not have one outage associated with this85 issue, they had a number each of which lasted hours.    : So you have a number of outages at eTrade which according = to your information was caused by Compaq and the 22 hour eBayr6 outage which was not caused by Sun hardware, software  or engineers.   5 Yet you still persist in trying muddy Sun's name overO9 the eBay outage. Not a very credible stance to be taking.a     > >rC > > > I'm waiting for those people to be shown to be liars.  Sorry, > > > > I can't take your word for it, you're just not credible.; > > > Everytime I do catch you in a clear lie, anyone who's-C > > > interested please trawl through deja, you change the subject.r > > >h? > > So when they said that it was Sun hardware or software thatQ? > > caused the outage when in fact it was a bug in an unpatched>; > > 3rd party product, they were not lying or being idiots.n > > > It's called credibility.  Since you've never had any I don't > expect you to recognize it.e >     5 No Jordan it does not matter how credible you appear )3 if you are not telling the truth. A lie is a lie nou7 matter how credible the source is, in the same way that-4 truth is also independant of the credibility to you  of the source.  8 You appear to be wrestling with a problem which is that 5 I work for Sun therefor I cannot be a credible sourcek4 in your eyes. But all the credible sources (in your 2 eyes) like Rob and Kerry have been shown up to be 5 often wrong particularly if their pronouncements have_4 anything to do with the capabilities or otherwise of to be released Compaq products.   6 On the other hand when Rob for example said Oracle and6 Sybase were really interested in Galaxies, I said they5 wern't and that there was no information from either -( Oracle or Sybase to back up Rob's claim.  5 You no doubt thought "lying idiot from Sun" credible >2 reliable Rob, guess who was right you do remember ' what did happen with Sybase and Oracle.C  5 Kerry for example said that I was spreading FUD when e8 I said that the WildFire architecture would force people4 to tune their applications specifically for the NUMA5 architecture to overcome the memory latency issues ono
 WildFire.   6 You thought credible Kerry "lying idiot from Sun" and 4 guess what Compaq go an publish a TPC-C result where> they have done exactly what I predicted, tuned the application: specifically to get around the shortcoming of the WildFire NUMA architecture. e  7 Seems to me that you need to find some credible sourceso% that actually get the odd fact right.w  ; > When Sun has been shown to have the unethical practice of < > forcing customers, who are entitled to Platinum service toA > sign gag-order NDAs before they get that service, we understand = > why a customer might initially express one opinion and then  > change their tune later. > 6 That is yet another allegation which does not hold up.  7 Sun has contractual obligations to its suppliers which  ; may require customers to sign NDA's to get the full detailsi5 but Sun would and does fix systems regardless of the l1 presence of an NDA. Another ludicrous allegation.h     Regards- Andrew Harrison- Enterprise IT Architect2   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:34 -0400R2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>/ Subject: Re: Unusual hard disk activity at boot 7 Message-ID: <200009231630_MC2-B46D-12D0@compuserve.com>t  G         Well, now you know why you should have been making regular disk  backups!!!!!  J         Your choices are: to wait for it to finish whatever its doing, to=  G try to login from a character cell terminal, or try rebooting.  I wouldaD much rather have a very recent backup than have to try any of these.    Message text written by "laurie"J >This morning, when I booted our Alpha, it got to the login screen, and t= heH hard disk wouldn't stop churning away. It allowed me to login, but stillH wouldn't stop once it had done that. It won't allow me to open a consoleA either, although DECterm appears in the 'work in process window'.(C Please help, because I'm scared of losing months work on that disk.      <a   ------------------------------   Date: 23 Sep 2000 19:32:36 GMT) From: begcompuser@aol.com (Beg Comp User)n Subject: VMS and UNIXD: Message-ID: <20000923153236.16269.00000927@ng-cg1.aol.com>  O What is the difference between VMS and UNIX?  What are the advantages of each? X TIA.   ------------------------------    Date: 23 Sep 2000 18:53:08 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)o Subject: Re: VMS and UNIX + Message-ID: <lk$a72Pa4AjQ@eisner.decus.org>c  f In article <20000923153236.16269.00000927@ng-cg1.aol.com>, begcompuser@aol.com (Beg Comp User) writes:Q > What is the difference between VMS and UNIX?  What are the advantages of each? $  . What makes you think there is any similarity ?  1 A closer comparison would be between VMS and MVS.e  4 But perhaps that was not the homework assignment :-)   ------------------------------  % Date: Sat, 23 Sep 2000 17:58:28 -0400n, From: Howard S Shubs <hshubs@mindspring.com> Subject: Re: VMS and UNIX > Message-ID: <hshubs-534DD6.17582723092000@news.mindspring.com>  ; In article <20000923153236.16269.00000927@ng-cg1.aol.com>, a* begcompuser@aol.com (Beg Comp User) wrote:  J >What is the difference between VMS and UNIX?  What are the advantages of  >each? s  5 We don't have space for a book here.  A -large- book.t -- s Howard S Shubs# "Run in circles, scream and shout!"n   ------------------------------  % Date: Sat, 23 Sep 2000 18:56:26 -0400., From: Howard S Shubs <hshubs@mindspring.com> Subject: Re: VMS and UNIXh> Message-ID: <hshubs-7BC1F6.18562623092000@news.mindspring.com>  N In article <lk$a72Pa4AjQ@eisner.decus.org>, Kilgallen@eisner.decus.org.nospam  (Larry Kilgallen) wrote:  2 >A closer comparison would be between VMS and MVS.  O If you take it -that- way, yer right.  There's no intersection between the set <! of {u,n,i,x} and that of {v,m,s}.c -- : Howard S Shubs# "Run in circles, scream and shout!"$   ------------------------------  % Date: Sun, 24 Sep 2000 00:37:24 +0200h" From: "Hans Vlems" <hvlems@iae.nl> Subject: Re: VMS and UNIXm( Message-ID: <8qjass$f1h$1@news.IAEhv.nl>  F In brief: the differences are too vast to explain in acouple of words.K The advantages: it depends what you want to do with the O/S of your choice.i  
 Hans Vlems  ) Beg Comp User heeft geschreven in berichtI1 <20000923153236.16269.00000927@ng-cg1.aol.com>...tI >What is the difference between VMS and UNIX?  What are the advantages ofl each?e >TIA.l   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:29 -0400s2 From: "Richard B. Gilbert" <DRAGON@compuserve.com> Subject: Re: VMS Memory Tuning7 Message-ID: <200009231630_MC2-B46D-12CD@compuserve.com>   F         The "global valid" faults are due to mapping global pages intoF somebody's working set.  This suggest that some, or some parts, of the) images are already installed with /SHARE.     & Message text written by David A Froble >Rob Young wrote:r > =d  < > In article <20000921100524.26592.00000003@ng-ch1.aol.com>,# briannfo@aol.com (BrianNFO) writes:sE > > Wanted to get some thoughts on a poorly performing VAX 7610.  Theo
 system hasJ > > 128 MB memory (maybe that's a big part of the problem right there) an= df it'sJ > > hard faulting way too much.  I think this is driving up Interrupt and=   Kernel
 > > modes. > >oA > > If I monitor the type of page faults, hard faulting is almostwJ > > constant...clearly a bad thing.  Free list is around 50,000 pages...b= ut one.J > > thing I notice is that about 50-75% of the faults are Demand Zero and=   GlobalJ > > Valid faults.  I'm trying to recall...are Demand Zero because of imag= ef > > activations? > >iJ > > Individual user processes do not seem to be faulting excessively, so = Ih thinkl' > > working sets are sized at least ok.e > >e# > > Thoughts?  Thanks in advance...l > >f > =o  H >         Yes... using -- Destroyer of the Environment -- Google revealsE >         image activitions are demand zero related , also hard fault-F >         recommendations are below that.  Running a common image?  IfJ >         so, is it INSTALLed?  Running a bunch of DCL scripts?  What you=   >         running?   Rob's pretty much explained it.h  J If you have many image activations of commonly used functions, installing=   themG as 'KNOWN' images will help in some ways.  I'd wonder about help on the  'DEMANDaJ ZERO' faults since it's my impression that these are process specific and=   not 
 shared pages.a  J Another possibly fix is that if you have a number of programs that are pa= rt of aJ specific application, such as a menu and a number of programs, linking al= leJ together into one image and installing it shared does help, but that isn'= tI an trival solution.  J With a free list of 50000 pages, you either have enough memory, or you ha= veJ working set numbers that are too small.  You indicate that faulting is no= t J excessive, so it may be the former.  Still, you could experiment with som= e J increases in working set numbers.  Pages on the free list are not working=   fori you. <u   ------------------------------  % Date: Sat, 23 Sep 2000 16:30:33 -0400e2 From: "Richard B. Gilbert" <DRAGON@compuserve.com> Subject: Re: VMS Memory Tuning7 Message-ID: <200009231630_MC2-B46D-12CF@compuserve.com>o  C         Demand zero and global valid faults are diagnostic of imagepG activations.  With "Polycenter Peformance Solution", you could find outeJ which users and which images but who can afford that now?  (It's now know= n   as Computer Associates AdviseIT)  9         Turning on image accounting might provide a clue.i  ?         It is also worth noting that some hard page faulting isdH unavoidable!  The first time any image is activated, it must be paged inF from disk.  If it is installed /SHARE some or all of it will remain inJ memory as long as someone is using it.  If no process is executing it, so= me  or all of it might be paged out.  J         Global valid and demand zero page faults are not hard faults; i.e= .lG they do not result in disk I/O.  A demand zero fault is a request for a F page of memory set to all zeros; the system will grab one off the freeJ list, erase it, and hand it to the process requesting it.  A global valid=  G fault is a request to map a page that is already in a global section ini memory.e      Message text written by BrianNFOJ >Wanted to get some thoughts on a poorly performing VAX 7610.  The system=   hasoJ 128 MB memory (maybe that's a big part of the problem right there) and it= 'sE hard faulting way too much.  I think this is driving up Interrupt andl Kernel modes.  = If I monitor the type of page faults, hard faulting is almosttJ constant...clearly a bad thing.  Free list is around 50,000 pages...but o= neE thing I notice is that about 50-75% of the faults are Demand Zero and. GlobalF Valid faults.  I'm trying to recall...are Demand Zero because of image activations?  F Individual user processes do not seem to be faulting excessively, so I thinkg# working sets are sized at least ok.     Thoughts?  Thanks in advance...<   ------------------------------   End of INFO-VAX 2000.534 ************************ur own"o > operating system,u* > I would probably become a unix advocate.   IMHO...e  @ If you think about it, VMS is _THE_ premier "roll your own" o.s.  F You get the basics, an excellent foundation. What you build with those pieces is up to you.  C All the various "stuff" you "get with" UN*X is the result of othersa: rolling THEIR own, for the lack of the pSpn7&$v^?֝]7Nj!'X^gDEngЀgT+-ͫ|o={G{eހT/e7Xui#9<b׍f1"$baeCiY*(C,!'BxAhOb+ |M\xֱwѹގ?piMvLnbA6|2HA=ۂ@^z?^qu{G8d(CzH؇!UsQ;C!FhWX4ҾC
xx~Fdq.Ne~U*(CߧarKIcU^/өk]U̕m@;Ԟt)?@yk6p-iKXX\vbVNAmWsL__sH6)-u;"VAxsiAzjik]f"g4'D
oƇpzMbs)NgLFτbXnq4
V^JmʦBYagJG1iΟTO׳GFwxwHRګw'\bzIl!}w}C'#=rlwCk
S^qbݘiX>mհf\e bs䣺:d~C9qǽpcwq4۔>p+j}$MVp,P^Hje8sp?ՉN\cḻaC}8>.f`;81I2lv42WwH{5%=څHbfAFS3Q{3%X1.X\{D}~Ɋccq<5%<zH%ϫ4AESe=kж(\E1D՗>KfW<bkX!
=?7s4/Ty[hr1lϰ#ᢐ	YJ8_{bF!4p}%/{p+.$b)ڻ֫Wy#)H๞==,sNFJt=LʲT8S<`'x>*
J;}OІỈ[aZ09`pR᠂8	5膋JC~/mS5r?YKo㲞Wg#qВwQw̅oiS>fGPkl2U׆z#
VǔvӠIusl4b_ D8a=|aYYXo'}ӿե3V7cϚp
x0dUKeJڴ* r:WbI'øX6/lcrmжM4_o3^A9rs;wWg1BmwuX
^0̃ed,BJ^>%-z~j?~b
<F1Rga= ,F@v&X)#<Ncj
ꮒ]_kmQ gI9y;Cط\wܠѸB}$9z^G:{0\]?'),̦sc{LBVkw\qcyQ<JSDk!=R픴B1֥8Ho?`E*!{iF
$1?w$55h@{\׭2A
ut}٬} ~!ؠcq.n>\6۽rr[CvUGyRI]}GSAӥE|x.[wZyj1H)'_Izax[.\_ϯ]כc&+AAUU`mM"!hAlaO"Y2n8X:E^9bm>K#UDw0|SiYY$sgA>pw'{0H )_]x֞&@c*,S
4hB4WJ}iK1;,TiyA_j7Tx`-G|%%Ҙ2Pkf%hbYuk7n/WzEMk~IJ!wbi=E-Էu~ܲN2FXgQO{m-0E*Ѧ6ކiϘ'tlu[q{f_uTB=w>Cn(1vpk!{퍷9FfajP7ŰEz|4֕3룝^=? TP^~3,ݿA
j랅50^h ,fRk{mnF==HAXp܎7yg3qy{p&&]bL%Ĝ
^gnXH\0p1K\ǯf牻߯	T5&hk naדзpvb]JC若F}|G ׳^,]&4nn[8طzELJ˓eք~\+*d%ސ0I}dCv*[{Ti6xn`s~^:,;xN2yn?`8mΔ
.b~nCl;]yg*q	 ǁ]Wߚ@v[>MfɭE
<_CvUiR{j$, _3\`e1,L3QC|tݼ
KA>׶iH[zl17}tnB?r\Y0{
^n+2͠9
{eU8+'B
gG^%*&Hvȶ |+u!q_k@(xy2߆<g$&~p-Ϸx'=fL#/{f~bIh(r3Y~ٴ7،v}&C<޶ Y^X\X,0A?؎;
e6&ڻb<w9]A̭Hxxc,xݹ~BH7m(po
~"ܣr{4kfBq;g»D vw$n9ƵgLqLW@N.%hgbrZa9	2/,`hTmoͽ	je<5֣m(9ւ^2x- >R	"GA}SA@}YmBf5mS3R_}S=Ҏ0洨'ݠEA6#kv	A7wte\|-O(cBs7׷\Qx|C}&iDg[S