0 INFO-VAX	Mon, 23 Jan 2006	Volume 2006 : Issue 45      Contents: Excessive paging problem Re: Excessive paging problem Re: Excessive paging problem FC Tape device ( DX30 ) ) Re: forgetting my DCL: what am I missing? ) Re: forgetting my DCL: what am I missing? ) Re: forgetting my DCL: what am I missing? ) Re: forgetting my DCL: what am I missing? 3 Re: Good OpenVMS presentation (slides # powerpoint) 3 Re: Good OpenVMS presentation (slides # powerpoint) 3 Re: Good OpenVMS presentation (slides # powerpoint) G Re: How about VMS: "Windows XP Gets Independent Security Certification" G Re: How about VMS: "Windows XP Gets Independent Security Certification" G Re: How about VMS: "Windows XP Gets Independent Security Certification" G Re: How about VMS: "Windows XP Gets Independent Security Certification" G Re: How about VMS: "Windows XP Gets Independent Security Certification" G Re: How about VMS: "Windows XP Gets Independent Security Certification" G RE: How about VMS: "Windows XP Gets Independent Security Certification" $ Re: Itanium lap tops are needed now!$ Re: Itanium lap tops are needed now!. Re: Q-Bus serpentine on 4000 series machines ?. Re: Q-Bus serpentine on 4000 series machines ?( Re: running OpenVMS/Alpha on a simulator  F ----------------------------------------------------------------------  % Date: Sun, 22 Jan 2006 22:58:19 +0100 % From: Bernhard Dorninger <bd_@gmx.at> ! Subject: Excessive paging problem . Message-ID: <1137963321.541881@news.liwest.at>   HiM It's me again, the Java programmer slowly turning into a VMS system admin :-I V   Having settled the problems with starting the app detached, we now face performance  issues due to excessive paging. / System is OpenVMS/Alpha 7.3-1. Java 1.4.2-04p2. W Our main app is started with a Java heap of max 96MB, the WSEXTENT is set to 160MB for  + that user (the same as WSMAX for the node). X The app spwans a subproc, which in turn is a Java app (a RMI server) with a max heap of # 64MB. (WSEXTENT has the same value)   X   I have read HP's Optimizing Java Guide for the OpenVMS platform carefully and ordered + the proc quotas to be adjusted accordingly. Y   On starting things up, everything works quite fine. However, after a day of uptime the  / working set grows far beyond the Java heap size T   until WSEXTENT is reached. This is the time, when the excessive paging starts and  performance drops extremely.W   We have noticed ridiculously high PF rates, not only soft faults but also hard faults   Z   I cannot understand, why the process keeps demanding pages, although Java heap has been ! restricted to a much lower value. V   Our main app has a very high object turnover, but shouldn't the Java GC handle this & without ordering additional mem pages?7   If anyone can help me (again), I'll be only too glad. Z See below some info output for our two processes. Nearly all of the output of mon page is : caused by them.  If you need more info, please contact me.   regards, Bernhard  ( ----------------------------------------   $ sho sys /proc=T* /fullL OpenVMS V7.3-1  on node LIS3ER  22-JAN-2006 20:59:01.27  Uptime  12 07:37:30#          AlphaServer 4100 5/400 4MB N    Pid    Process Name    State  Pri      I/O       CPU       Page flts  PagesP 0000223F TG_DEV_193      HIB     10  3807056   0 01:25:47.76  12950933  20480 MSP           [TG_DEV]                                                      163840KbO 00002111 TM_LZU341       HIB     10 23463540   0 16:00:51.77 630243239  20480 M P           [TG_DEV]                                                      163840Kb  ' ---------------------------------------   2                            Working Set Information  E                                   WS    WS    WS     WS   Pages  Page L Username    Processname   State  Extnt Quota Deflt  Size in WS  Faults Image  S TG_DEV      TM_LZU341      HIB  327680  8192  4096327680327680  630243239 JAVA$JAVA M TG_DEV      TG_DEV_193     HIB  327680  8192  4096327680327680  12950933 JAVA   ' ---------------------------------------    monitor page output L                                         CUR        AVE        MIN        MAX  L      Page Fault Rate                 375.33    9471.57     101.33   75682.66L      Page Read Rate                  488.00     382.51       7.33     815.66L      Page Read I/O Rate              116.33     121.19       4.00     232.00L      Page Write Rate                   0.00       2.84       0.00     226.00L      Page Write I/O Rate               0.00       1.00       0.00      74.33  L      Free List Fault Rate            251.33    5543.55       0.00   45827.33L      Modified List Fault Rate          5.00    3778.41       1.00   29646.33L      Demand Zero Fault Rate            0.00       0.15       0.00       6.66L      Global Valid Fault Rate           2.66      26.80       0.00     190.66L      Wrt In Progress Fault Rate        0.00       0.00       0.00       0.00L      System Fault Rate                 0.00       0.00       0.00       0.00  L      Free List Size                 7469.00    9814.45    1218.00   18470.00L      Modified List Size            14307.00   11579.47    3475.00   20590.00   ------------------------------  % Date: Sun, 22 Jan 2006 16:56:17 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>% Subject: Re: Excessive paging problem + Message-ID: <43D40D91.3FD4457F@comcast.net>    Bernhard Dorninger wrote:  >  > HiO > It's me again, the Java programmer slowly turning into a VMS system admin :-I W >   Having settled the problems with starting the app detached, we now face performance ! > issues due to excessive paging. 1 > System is OpenVMS/Alpha 7.3-1. Java 1.4.2-04p2. X > Our main app is started with a Java heap of max 96MB, the WSEXTENT is set to 160MB for- > that user (the same as WSMAX for the node). Y > The app spwans a subproc, which in turn is a Java app (a RMI server) with a max heap of % > 64MB. (WSEXTENT has the same value)    So far, so good.  Y >   I have read HP's Optimizing Java Guide for the OpenVMS platform carefully and ordered - > the proc quotas to be adjusted accordingly. Z >   On starting things up, everything works quite fine. However, after a day of uptime the1 > working set grows far beyond the Java heap size  >   until WSEXTENT is reached.  F Suggests possible resource "leakage". Check for ECOs to the respective RTLs.   8 > This is the time, when the excessive paging starts and > performance drops extremely.Y >   We have noticed ridiculously high PF rates, not only soft faults but also hard faults  > [ >   I cannot understand, why the process keeps demanding pages, although Java heap has been # > restricted to a much lower value.   G Depends how much data the program is trying to keep in memory, not on a  fixed heap size.  W >   Our main app has a very high object turnover, but shouldn't the Java GC handle this ( > without ordering additional mem pages?  > Depends on whether resources are being released as objects are
 dismissed.  9 >   If anyone can help me (again), I'll be only too glad. [ > See below some info output for our two processes. Nearly all of the output of mon page is < > caused by them.  If you need more info, please contact me. > 
 > regards,
 > Bernhard > * > ---------------------------------------- >  > $ sho sys /proc=T* /fullN > OpenVMS V7.3-1  on node LIS3ER  22-JAN-2006 20:59:01.27  Uptime  12 07:37:30% >          AlphaServer 4100 5/400 4MB P >    Pid    Process Name    State  Pri      I/O       CPU       Page flts  PagesR > 0000223F TG_DEV_193      HIB     10  3807056   0 01:25:47.76  12950933  20480 MSR >           [TG_DEV]                                                      163840KbQ > 00002111 TM_LZU341       HIB     10 23463540   0 16:00:51.77 630243239  20480 M R >           [TG_DEV]                                                      163840Kb > ) > ---------------------------------------  > 4 >                            Working Set Information > G >                                   WS    WS    WS     WS   Pages  Page N > Username    Processname   State  Extnt Quota Deflt  Size in WS  Faults Image > U > TG_DEV      TM_LZU341      HIB  327680  8192  4096327680327680  630243239 JAVA$JAVA O > TG_DEV      TG_DEV_193     HIB  327680  8192  4096327680327680  12950933 JAVA  > ) > ---------------------------------------  >  > monitor page output N >                                         CUR        AVE        MIN        MAX > N >      Page Fault Rate                 375.33    9471.57     101.33   75682.66N >      Page Read Rate                  488.00     382.51       7.33     815.66N >      Page Read I/O Rate              116.33     121.19       4.00     232.00N >      Page Write Rate                   0.00       2.84       0.00     226.00N >      Page Write I/O Rate               0.00       1.00       0.00      74.33 > N >      Free List Fault Rate            251.33    5543.55       0.00   45827.33N >      Modified List Fault Rate          5.00    3778.41       1.00   29646.33N >      Demand Zero Fault Rate            0.00       0.15       0.00       6.66N >      Global Valid Fault Rate           2.66      26.80       0.00     190.66N >      Wrt In Progress Fault Rate        0.00       0.00       0.00       0.00N >      System Fault Rate                 0.00       0.00       0.00       0.00 > N >      Free List Size                 7469.00    9814.45    1218.00   18470.00N >      Modified List Size            14307.00   11579.47    3475.00   20590.00  E WOW!!!!!!!!!! What was the /INTERVAL time on MONITOR PAGE? (Hint: the 2 default is 5 seconds, I usually use /INT=2 or =1).   --   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, 23 Jan 2006 01:38:03 +0100 % From: Bernhard Dorninger <bd_@gmx.at> % Subject: Re: Excessive paging problem . Message-ID: <1137972903.681098@news.liwest.at>   David J Dachtera wrote:  > Bernhard Dorninger wrote:  > Y >>  I have read HP's Optimizing Java Guide for the OpenVMS platform carefully and ordered - >>the proc quotas to be adjusted accordingly. Z >>  On starting things up, everything works quite fine. However, after a day of uptime the1 >>working set grows far beyond the Java heap size  >>  until WSEXTENT is reached. >  > H > Suggests possible resource "leakage". Check for ECOs to the respective > RTLs.  Will check that....  >  > 8 >>This is the time, when the excessive paging starts and >>performance drops extremely.Y >>  We have noticed ridiculously high PF rates, not only soft faults but also hard faults  >>[ >>  I cannot understand, why the process keeps demanding pages, although Java heap has been # >>restricted to a much lower value.  >  > I > Depends how much data the program is trying to keep in memory, not on a  > fixed heap size.Z Well, the Java Heap is indeed of a fixed size. I.e. if you have a Java program that tries Z to keep more than max heap (set with the -Xmx cmd line arg of Java) sized data in memory, Y you'll get a java.lang.OutOfMemoryError in your code - at least you should get it. Proof   of this fact is easy. Z This is not the case for the JavaVM itself or any native library, which is loaded (either ( from the VM or from your own Java Code).X  From this point of view, I'd say the VM (or any of the C-RTL libs) is responsible, but X who knows - maybe I have made an error with the VMS proc quota settings, which I am not 	 aware of.    > W >>  Our main app has a very high object turnover, but shouldn't the Java GC handle this ( >>without ordering additional mem pages? >  > @ > Depends on whether resources are being released as objects are > dismissed.Z For our objects (mostly strings and C-struct like data objects), there can't be done more [ than to set the references to null. The rest is usually done by the Java GC. I am not sure  Z but the GC activity should not cause paging, but for sure it should not cause an increase  of process memory. >  >>monitor page output N >>                                        CUR        AVE        MIN        MAX >>N >>     Page Fault Rate                 375.33    9471.57     101.33   75682.66N >>     Page Read Rate                  488.00     382.51       7.33     815.66N >>     Page Read I/O Rate              116.33     121.19       4.00     232.00N >>     Page Write Rate                   0.00       2.84       0.00     226.00N >>     Page Write I/O Rate               0.00       1.00       0.00      74.33 >>N >>     Free List Fault Rate            251.33    5543.55       0.00   45827.33N >>     Modified List Fault Rate          5.00    3778.41       1.00   29646.33N >>     Demand Zero Fault Rate            0.00       0.15       0.00       6.66N >>     Global Valid Fault Rate           2.66      26.80       0.00     190.66N >>     Wrt In Progress Fault Rate        0.00       0.00       0.00       0.00N >>     System Fault Rate                 0.00       0.00       0.00       0.00 >>N >>     Free List Size                 7469.00    9814.45    1218.00   18470.00N >>     Modified List Size            14307.00   11579.47    3475.00   20590.00 >  > G > WOW!!!!!!!!!! What was the /INTERVAL time on MONITOR PAGE? (Hint: the 4 > default is 5 seconds, I usually use /INT=2 or =1).X used the default interval. changing it to a lower value does not deliver more beautiful  results :-)    Thanks,  Bernhard   ------------------------------    Date: 22 Jan 2006 19:04:03 -0800) From: "kiwi-red" <antonywardle@gmail.com>   Subject: FC Tape device ( DX30 )C Message-ID: <1137985443.168564.174670@g49g2000cwa.googlegroups.com>    Hi  E Has anyone tried using a DX30 on an Alpha. I saw a post from a couple  of years ago, but nothing else.   B I have an ES40 running 7.3-2 with a KGPSA-C connected to a brocade switch. C I have a DX30 ( http://www.ultriumlto.com/quantum/quantumdx30.htm )  which + is set up to be 12 DLT devices and a robot.   G The switch says that both devices are In Sync. According to the Brocade G Fabric OS manual this is normal and shows that you have a laser in sync  with the device attached  E The SAN guy here thinks that the devices should be broadcasting their 
 WWID numbers.   E The KGPSA was set to loop with the DX 30 direct connected but we were D unable to see any device. We've connected it to a switch and changed= the KGPSA to Fabric topology and we still can't see anything.   % mc sysman io a doesn't find anything. * a show dev from >>> doesn't show anything.C a mc sysman io find says No previously undiscovered WWIDs have been  found.  G I haven't done any wwidmgr command from the >>> prompt. As I understand   it, this is for disks not tapes.    9 Any idea what I need to do to make these tapes available?    antony   ------------------------------  + Date: Sun, 22 Jan 2006 19:44:22 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)2 Subject: Re: forgetting my DCL: what am I missing?$ Message-ID: <dr0nam$6bo$1@online.de>  5 In article <43D356EF.2F11ECD0@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:    ! > say = "write sys$output ""===>"  >  > $say/hello world" . > ==> /hello world            DCL adds a space >  >  >  > $'say'/hello world" 4 > ===>/hello world           DCL doesn't add a space >  > $'say /hello world" / > ===>/hello world          DCL removes a space  >  > $'say/hello world ; > ===>/hello world          Doesn't add nor remove a space.  >  > H > Looks like the code that checks for possible symbol substitutions in aF > DCL command has code to add or remove spaces. Mayby Guy Peleg should > look into this.   G Interesting.  In my case, though, this hack won't work, since I have a  / foreign command in the first symbol definition.    ------------------------------  + Date: Sun, 22 Jan 2006 19:45:40 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)2 Subject: Re: forgetting my DCL: what am I missing?$ Message-ID: <dr0nd4$6bo$2@online.de>  5 In article <43D3576A.88CB4EEB@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:     > Manage host ip.ad.re.ss  > Manage port 931   E OK, but it has to be done explicitly, so knowing about 931 is no big  
 deal.  :-)   ------------------------------  % Date: Sun, 22 Jan 2006 14:52:52 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>2 Subject: Re: forgetting my DCL: what am I missing?+ Message-ID: <43D3F0A4.3209AEAA@comcast.net>   / Phillip Helbig---remove CLOTHES to reply wrote:  > E > In article <1137893795.722320.292690@g44g2000cwa.googlegroups.com>, ( > "AEF" <spamsink2001@yahoo.com> writes: >  > > > Consider the following:  > > > W > > > $  PRIVREQUEST :== $DISK$SOFT:[HTTP_SERVER_3-10A.'ARCH'.BASE_CODE]PRIVREQUEST.EXE 	 > > [...] 4 > > > $  WWW_NEWTRACE :== 'PRIVREQUEST' 931 NEWTRACE	 > > [...]  > > " > > >    $  privrequest newtrace/3+ > > > then it works fine.  This also works: E > > The line above is missing the 931 even after symbol substitution.  > G > Sorry, I should have written that "privrequest 931 newtrace/3" works.  > J > > 931: The line with /3 is missing 931. The line with /11 is not missing > > the 931. > 4 > Sorry for my typos; here's a direct cut-and-paste: >  > $ sh sym privrequest@ >   PRIVREQUEST == "$DISK$SOFT:[HTTP_SERVER_3-10A.VAX.BASE_CODE] > PRIVREQUEST.EXE" > $ show symbol www_newtrace^ >   WWW_NEWTRACE == "$DISK$SOFT:[HTTP_SERVER_3-10A.VAX.BASE_CODE]PRIVREQUEST.EXE 931 NEWTRACE" > $ www_newtrace > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ www_newtrace/11 F > error connecting to 'localhost': -1(can't assign requested address ) > $ www_newtrace /11F > error connecting to 'localhost': -1(can't assign requested address ) > $ privrequest 931 newtrace > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ privrequest 931 newtrace/0 > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ privrequest 931 newtrace /0 F > error connecting to 'localhost': -1(can't assign requested address ) > I > So, it looks like JF's remark about "DCL inserting a space" is correct. , > However, WHY does DCL insert a space here?   Generally, it's valid syntax.   B From DCL's perspective "command /qualifier = keyword" is generallyG identical to "command/qualifier=keyword". When DCL (CLI$PARSE) is *NOT* G used to parse out the command line, sensitivity to variations in syntax 0 becomes more unpredictable and program-specific.  G Recall that in UN*X-land, "whitespace" is used to separate command line 
 arguments.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------    Date: 22 Jan 2006 21:56:44 -0800$ From: "AEF" <spamsink2001@yahoo.com>2 Subject: Re: forgetting my DCL: what am I missing?C Message-ID: <1137995804.769111.112900@g44g2000cwa.googlegroups.com>   / Phillip Helbig---remove CLOTHES to reply wrote: E > In article <1137893795.722320.292690@g44g2000cwa.googlegroups.com>, ( > "AEF" <spamsink2001@yahoo.com> writes: >  > > > Consider the following:  > > > W > > > $  PRIVREQUEST :== $DISK$SOFT:[HTTP_SERVER_3-10A.'ARCH'.BASE_CODE]PRIVREQUEST.EXE 	 > > [...] 4 > > > $  WWW_NEWTRACE :== 'PRIVREQUEST' 931 NEWTRACE	 > > [...]  > > " > > >    $  privrequest newtrace/3+ > > > then it works fine.  This also works: E > > The line above is missing the 931 even after symbol substitution.  > G > Sorry, I should have written that "privrequest 931 newtrace/3" works.  > J > > 931: The line with /3 is missing 931. The line with /11 is not missing > > the 931. > 4 > Sorry for my typos; here's a direct cut-and-paste: >  > $ sh sym privrequest@ >   PRIVREQUEST == "$DISK$SOFT:[HTTP_SERVER_3-10A.VAX.BASE_CODE] > PRIVREQUEST.EXE" > $ show symbol www_newtrace^ >   WWW_NEWTRACE == "$DISK$SOFT:[HTTP_SERVER_3-10A.VAX.BASE_CODE]PRIVREQUEST.EXE 931 NEWTRACE" > $ www_newtrace > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ www_newtrace/11 F > error connecting to 'localhost': -1(can't assign requested address ) > $ www_newtrace /11F > error connecting to 'localhost': -1(can't assign requested address ) > $ privrequest 931 newtrace > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ privrequest 931 newtrace/0 > Port 80 response: F > 200 New version of trace log opened at logger level: 0, old level: 0 >  > $ privrequest 931 newtrace /0 F > error connecting to 'localhost': -1(can't assign requested address ) > I > So, it looks like JF's remark about "DCL inserting a space" is correct. , > However, WHY does DCL insert a space here?    % I think it's more that in the command        www_newtrace/11   G ww_newtrace is "substituted" in phase 2 of symbol substitution. In this F phase, the command is already DCL-parsed, i.e., as Gotfryd Smolik onceF put it: "broken into parts". So at this stage the command is no longer- a simple text string. For example, you can do        $ DT="1-DEC-2005 12:21"      $ DIR/SINC=&DT  A and it will work. So the symbol substituion for a foreign command E probably happens AFTER the command is parsed, also. The same goes for & ampersand substitution as shown above.  D So you can't simply append the /11 -- it is too late. You can appendC only in phase 1. This means that 'www_newtrace'/11 should work. The ? foreign command appears to be designed to accept paramters, not  appended qualfiers.    Also, check out this:        $ DIR:==DIRECTORY/DATE/SIZE   
     $ DIR.OBJ   * This DIR.OBJ command will work the same as       $ DIR .OBJ  E For DIR.OBJ to work as DIR .OBJ!, DCL must "insert a space" (emphasis  on quotes).   E I think this is even alluded to in the User's Manual but I could only = find an example like this regarding automatic foreign command > substitution. I don't have the time to look further right now.  F I think this also has to do with how $AAA:==BLAH is parsed, or perhaps vice versa! Hey, it's late!    This is how DCL works.     AEF  &-)    ------------------------------  % Date: Sun, 22 Jan 2006 21:56:08 +0100 ( From: Paul Sture <paul.sture@bluewin.ch>< Subject: Re: Good OpenVMS presentation (slides # powerpoint), Message-ID: <43i9r8F1nmg91U1@individual.net>   BRAD wrote: " > VAXman- @SendSpamHere.ORG wrote: > 2 >> This whole thread is moot if you can't view it. >> > H > Be grateful you can't view it - it's not very readable.  Way too much I > text information in the slides themselves.  The notes accompanying the  A > slides are more suitable for the amount of textual information  K > presented.  This was produced by a company (avitage) that is supposed to  K > be presentation-savvy; if this is a typical example of the presentations  7 > they produce, avoid this company at all costs.    :-)  >    <snip>   > I > I don't suppose you have M$ Office on your PowerBook?    :-)  I wonder  I > if OpenOffice would render it (I think I used OpenOffice several years  B > ago to view a PP presentation, just to see if it could be done)?    D It's the first PPT file I've come across which Keynote (Apple's own H presentation tool) can't make reasonable sense of. At points it's as if G the previous transparency or three were left on the overhead projector.   E OO suffered in a different way, and I'm sure it missed some text out   completely.    ------------------------------  % Date: Sun, 22 Jan 2006 16:55:14 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> < Subject: Re: Good OpenVMS presentation (slides # powerpoint), Message-ID: <43D3FF40.6AA4B4C4@teksavvy.com>  ( > >> PUT THIS CRAP INTO A USEFUL FORMAT.  B In fairness, this was produced by a wintel company (HP) to be usedC during presentations with the powerpoint stuff projected onto a big 6 screen while the speaker elaborates on all the points.  F Like it or not (and I don't like it either), Powerpoint has become theD de facto standard, to such an extent that the main competitor (Aldus. Persuasion) dropped out of the race years ago.  G If HP has such as sweet relationship with Microsoft, why don't they get H MS to do a Powerpoint viewer for VMS on VAX, Alpha and whatever else VMS	 runs on ?    ------------------------------  % Date: Sun, 22 Jan 2006 16:49:03 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>< Subject: Re: Good OpenVMS presentation (slides # powerpoint)+ Message-ID: <43D40BDE.4DBA8AAB@comcast.net>    JF Mezei wrote:  > * > > >> PUT THIS CRAP INTO A USEFUL FORMAT. > D > In fairness, this was produced by a wintel company (HP) to be usedE > during presentations with the powerpoint stuff projected onto a big 8 > screen while the speaker elaborates on all the points. > H > Like it or not (and I don't like it either), Powerpoint has become theF > de facto standard, to such an extent that the main competitor (Aldus0 > Persuasion) dropped out of the race years ago. > I > If HP has such as sweet relationship with Microsoft, why don't they get J > MS to do a Powerpoint viewer for VMS on VAX, Alpha and whatever else VMS > runs on ?   H Which raises the question: is there an open source PPT viewer that could be ported to VMS+DECwindows?   --   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: Sun, 22 Jan 2006 14:03:53 -0500 ' From: Dave Froble <davef@tsoft-inc.com> P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"0 Message-ID: <11t7lnv71a4sc53@corp.supernews.com>   Main, Kerry wrote: > J > [Note - because of the one app, one server model, the average x86 serverI > peak utilization in prime time today is between 5-20% .. Ask any CIO if I > they would be satisfied if they got that much productivity out of their  > employees..]  ' You've beat upon this issue many times.   $ Hardware prices continue to go down.    Employees continue to cost more.  # The two are not a valid comparison.   G That said, I do agree that the employee costs for running many systems  I is an issue, and as stated, those employee costs only go up, or to India.    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------    Date: 22 Jan 2006 12:32:32 -0800 From: davidc@montagar.com P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"B Message-ID: <1137961952.215981.10780@o13g2000cwo.googlegroups.com>  E >Most of the public don't deal in confidence.  If they did they never D >would have adopted MS in the first place.  I would imagine the rushA >to certification is needed to continue their push in gov (and in 0 >particular DOD) sales.  "Perception is Reality"  D Not just the government. Their "Secure Computing" initiative has notB been enough, because sloppy programming continues to be exploited,B business impacting, and news-worthy. CIEIO's do hear this, and areG listening to places that tell people to abandon IE in favor of Firefox, G hear of the losses due to bugs on Microsoft software - and do ask "What  alternatives are there?".    ------------------------------    Date: 22 Jan 2006 12:53:43 -0800 From: davidc@montagar.com P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"C Message-ID: <1137963223.846360.109830@g49g2000cwa.googlegroups.com>   E True - the certifications are basically "User make file non-readable, F other users cannot read file" type of checks.  While this is good from@ a basica security standpoint, it fails to recognize the numerousF exploits where "User makes file non-readable, others users pwn system, read file anyway".  G The security of the O/S API needs to be fully considered.  Is every API @ safe?  Are all parameters fully checked to ensure safe access ofC memory?  Are all appropriate object rights validated before access?    ------------------------------    Date: 22 Jan 2006 12:40:09 -0800 From: davidc@montagar.com P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"C Message-ID: <1137962409.265580.307510@g43g2000cwa.googlegroups.com>   A I agree on the password issue.  One of my clients is undergoing a G review fora HIPAA certification, and one of the suggestions is (now get G this!) users should not set their own passwords.  The Supervisor  is to : dispense passwords to the user.  Since doing this means anC organizations IT department needs to change everyones password, and B distribute everyones new password to the supervisors.  Nice - talkF about widening the exposure of passwords, incidence ot passwords being? put on post-its on the terminals.  Might as well post it on the # bulletin board in the break room...    ------------------------------  % Date: Sun, 22 Jan 2006 15:03:58 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"* Message-ID: <43D3F33E.2BD11BB@comcast.net>   davidc@montagar.com wrote: > C > I agree on the password issue.  One of my clients is undergoing a I > review fora HIPAA certification, and one of the suggestions is (now get I > this!) users should not set their own passwords.  The Supervisor  is to < > dispense passwords to the user.  Since doing this means anE > organizations IT department needs to change everyones password, and D > distribute everyones new password to the supervisors.  Nice - talkH > about widening the exposure of passwords, incidence ot passwords beingA > put on post-its on the terminals.  Might as well post it on the % > bulletin board in the break room...   ) SSSSHHHHHHH!!!! Don't give 'em any ideas!    --   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, 23 Jan 2006 00:58:02 +0100 + From: Karsten Nyblad <nospam@nospam.nospam> P Subject: Re: How about VMS: "Windows XP Gets Independent Security Certification"= Message-ID: <43d41bff$0$78283$157c6196@dreader1.cybercity.dk>    JF Mezei wrote: I > VMS has separate DATA and CODE memory areas, and you cannot branch to a D > DATA memory location, nor can you write to a CODE memory location. > (correct ?). > D > Would it be too late to give Windows the same memory protections ?  C Yes, because it already in Windows XP SP2.  The chips implementing  H x86-64/EMT64/AMD64 has a feature where you can mark pages as executable I and not executable, and the latest versions of Windows use that feature.  I   The user may disable this new feature either for the entire machine or  E for specific applications.  There are a few video decoding programs,  5 that generate code in data areas and then execute it.    ------------------------------  % Date: Mon, 23 Jan 2006 00:12:14 -0500 ' From: "Main, Kerry" <Kerry.Main@hp.com> P Subject: RE: How about VMS: "Windows XP Gets Independent Security Certification"R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB7D33F8@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Dave Froble [mailto:davef@tsoft-inc.com]=20   > Sent: January 22, 2006 2:04 PM > To: Info-VAX@Mvb.Saic.Com = > Subject: Re: How about VMS: "Windows XP Gets Independent=20  > Security Certification"  >=20 > Main, Kerry wrote: > >=20< > > [Note - because of the one app, one server model, the=20 > average x86 server? > > peak utilization in prime time today is between 5-20% ..=20  > Ask any CIO ifA > > they would be satisfied if they got that much productivity=20  > out of their > > employees..] >=20) > You've beat upon this issue many times.  >=20& > Hardware prices continue to go down. >=20  H And server utilization continues to fall while at the same time softwareC licensing, management costs continue to be a major and rising cost. C Remember that most third party management and maintenance costs are E based on per server and/or per cpu basis. So, you end up paying these G costs for monitoring, maintaining and managing servers that are grossly  under utilized.     " > Employees continue to cost more. >=20% > The two are not a valid comparison.  >=20> > That said, I do agree that the employee costs for running=20 > many systems=20 A > is an issue, and as stated, those employee costs only go up,=20  > or to India.  D While off shoring is still a major consideration, many companies areG being much more selective these days as they have begun to realize that A rising costs in India, additional time required to QA work, staff G retention, and the lack of technical expertise that understands how the F business operates is causing these companies to re-think some of their off-shoring plans.  H Some companies are beginning to re-focus their attention to still havingD fewer IT staff, but these staff would be more highly trained and use= higher value tools to make them more proactive. This is where F consolidation (servers and data centers) is becoming a major player to drive this.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------  % Date: Sun, 22 Jan 2006 15:00:02 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>- Subject: Re: Itanium lap tops are needed now! + Message-ID: <43D3F252.B4D2BAF8@comcast.net>    Bill Gunshannon wrote: > [snip]E > I wold love to know what machine you are trying that the CD doesn't 1 > boot and go right into the install program on!!   D I first tried FreeBSD on a '386 machine (Packard Bell - my first PC)C which does not have a CD drive and only had 4MB of RAM at the time. @ Hence, the genesis of my very stale FreeBSD pages on my website.  G Most recently, I tried using a DECpc 450/ST which does not boot from CD D (the BIOS does not support it) and only has 12MB of RAM. (SYSINSTALL needs 16MB last I tried it).  F With my limited budget, I don't have a newer machine I can dedicate toE FreeBSD or Linux. I've tried using my Celeron-366 machine with an Orb A drive so I could change out the boot disk at will, but that never  worked, either.    --   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: 22 Jan 2006 22:00:43 GMT( From: bill@cs.uofs.edu (Bill Gunshannon)- Subject: Re: Itanium lap tops are needed now! , Message-ID: <43idkbF1o24oiU2@individual.net>  + In article <43D3F252.B4D2BAF8@comcast.net>, 5 	David J Dachtera <djesys.nospam@comcast.net> writes:  > Bill Gunshannon wrote:	 >> [snip] F >> I wold love to know what machine you are trying that the CD doesn't2 >> boot and go right into the install program on!! > F > I first tried FreeBSD on a '386 machine (Packard Bell - my first PC)E > which does not have a CD drive and only had 4MB of RAM at the time. B > Hence, the genesis of my very stale FreeBSD pages on my website. > I > Most recently, I tried using a DECpc 450/ST which does not boot from CD F > (the BIOS does not support it) and only has 12MB of RAM. (SYSINSTALL > needs 16MB last I tried it).  B So what, isn't that like complaining about VMS because the currentD version won't run on your 11/780?  (I understand it didn't boot from CDROM either :-)   > H > With my limited budget, I don't have a newer machine I can dedicate toG > FreeBSD or Linux. I've tried using my Celeron-366 machine with an Orb C > drive so I could change out the boot disk at will, but that never  > worked, either.   8 I hardly think your limited budget is FreeBSD's fault.     bill      --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  # Date: Mon, 23 Jan 2006 00:49:49 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>7 Subject: Re: Q-Bus serpentine on 4000 series machines ? 0 Message-ID: <NKVAf.1952$uK3.76@news.cpqcorp.net>   JF Mezei wrote:   I > On my new VAX 4000s, (-500A that says it is a 600a , and a -200), there J > are no neat drawings. And I have a KLESI card which occupies only the ABD > slots, and there is a plastic insert for the CD slots that have no6 > connectors at all (which means no grant continuity).  D All slots within the BA200-series and in the BA400-series boxes are C Q22/CD.  These boxes are also known as S-boxes, and these boxes do  / require Q-bus grant continuity for empty slots.   J > Looking at dj/google, I can see plenty of posts asking people to referC > to "hoffman" and "serpentine". But can't seem to find a post that B > actually provides the information about the 4000s'q QBUS :-( :-(  H Please explain where you looked, and what you searched for.  I ask this F as I seek to learn where and what folks look for, as that can help me A make that information more accessible; this can allow me to seed  F discussions with more appropriate keywords, or pointers to tools, etc.  5 I found 31 hits with the google groups search string:   '    hoffman serpentine group:comp.os.vms   G Here are the three (oldest, too) that were returned by google that are   relevant here:  a http://groups.google.com/group/comp.os.vms/browse_thread/thread/4221981c0f27b970/11bae473dbf85720 a http://groups.google.com/group/comp.os.vms/browse_thread/thread/b1b38b24974f7304/8d8bbfc0445e3132 a http://groups.google.com/group/comp.os.vms/browse_thread/thread/e9a310f767150494/1c42735fe042b138   G To reduce the numbers of off-topic hits on a google web search for the  , information, I had to use the search string:     hoffman serpentine q-bus  / The first hit from the search is the following:   . http://h71000.www7.hp.com/wizard/wiz_1149.html  G I do see the cited Ask The Wizard topic does not show up if you search   for the variant spelling qbus.  G To make it even easier, I'll get a reference added to the next edition  G of the Frequently Asked Questions.  (I'd thought there was a reference  < within, but I don't see it.  That can and will be remedied.)  @ Here are the VAX 4000 manuals that are available at the HP site:  > http://h18002.www1.hp.com/alphaserver/vax/archive/vax4000.html  I I don't see a service guide posted there, but I did not look through the   list particularly carefully.   ------------------------------  % Date: Sun, 22 Jan 2006 20:25:29 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 7 Subject: Re: Q-Bus serpentine on 4000 series machines ? , Message-ID: <43D43079.F0795E9E@teksavvy.com>   Hoff Hoffman wrote: E > All slots within the BA200-series and in the BA400-series boxes are D > Q22/CD.  These boxes are also known as S-boxes, and these boxes do1 > require Q-bus grant continuity for empty slots.   G From how it was later explained to me, these QBUSes are NOT serpentine. F All board go in the AB slots. The is different than the BA23 box whereH slots 1-2-3 are not serpentine, but then goes 4 AB to 4CD to 5CD to 5 AB to 6AB to 6CD etc.  H On the 4000s, from what I was told (and what seems to work), they all goE into AB and no need to fill CD with grant cards when the card doesn't 4 fill AB and CD (what I think is called quad height).        = > Please explain where you looked, and what you searched for.   B Looked for q-bus or Q22 and serpentine and VAX 4000 foudn nothing.C Removed the VAX 4000 and found a whole lot of posts all alluding to 5 someone doing a serach for hoffman and serpentine :-)   E I may have found posts that do describe the true ser]entine nature in A the BA23 , but not for the more modern machines. On my All mighty F Microvax II, the serpentine layout was perfectly documented inside the( cabinet so it was never an issue for me.  0 > http://h71000.www7.hp.com/wizard/wiz_1149.html  D I had stumbled on that text. However, I was perhaps looking for 4000C instead of BA400 and this page begins with a question about a 3500. D There is a mention in there that it isn't serpentine, but it had notG been obvious to me because I was looking at VAX model number instead of 
 cabinet type.   F However, the explanations still focus mostly on the MVII style of Qbus5 and not so "in your face" for the more modern models.   @ > http://h18002.www1.hp.com/alphaserver/vax/archive/vax4000.html  H The manuals  don't mention anything obvious  about QBUS organisation. MyE guess is that all the boards supplied for that class of machine had a G cab kit that was quad height with whatever plastic insert necessary and C thus no worry about whether arrangement is serpentine or not and no - worry if a boartd was AB or AB-CD or just CD.   E Another thing: the VAX 4000-200 has a different backpane with the CPU H and memory getting onto the QBUS, while on the 4000-500/600, the CPU hasG its onw interconnect, the momeory has its own plugs and then there is a E completely independant QBUS, so it isn't obvious that they all behave 9 the same in terms of phsycal configuration of the boards.    ------------------------------  % Date: Sun, 22 Jan 2006 21:16:53 -0500 2 From: "Timothy Stark" <fsword7_nospam@comcast.net>1 Subject: Re: running OpenVMS/Alpha on a simulator 0 Message-ID: <NdSdncLzF6QGoUneRVn-gA@comcast.com>  ; "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message  - news:87VDD4r4HlQk@eisner.encompasserve.org... F > In article <1137820459.725149.229040@z14g2000cwz.googlegroups.com>,  > davidc@montagar.com writes: E >> Cool!  Somewhere, there should be a program called "Manequin" that J >> provides an emulation of Alpha hardware - which I think may run on VAX.H >>  I think it was used during the initial OpenVMS porting efforts until5 >> sufficent hardware was available to do real stuff.  > 2 > I believe Mannequin ran at .1 VUPS or .01 VUPS.\  G Well, that remind me about my TS10 emulator development.  I was trying  L developing the Alpha emulator and it only took 10 seconds to uncompress SRM M firmware from ROM image into physical memory. Yes, that was very slow on 350  M Mhz PIII processor.  I suspended that Alpha emulator developement until much  ) faster P4 processor or 64-bit AMD, etc...   L Where is "Manequin" Alpha emulator?  I was searching for that on Google but  can't find.    Thanks,  Tim    ------------------------------   End of INFO-VAX 2006.045 ************************