1 INFO-VAX	Mon, 04 Apr 2005	Volume 2005 : Issue 188       Contents: Re: %MOUNT-F-SHDWCOPYREQ! ANN: VTfm V2.3-9 is now available * Announcing April 19, 2005 WRUG LUG Meeting Re: cloning a system disk ! Re: EDT question: SET ENTITY WORD 4 Re: Finding The Right Path For Your VMS Applications! Re: indicator lights on SBB disks ! Re: indicator lights on SBB disks 
 Re: InfoTower 
 Re: InfoTower  microvax# Newbie non blocking socket question ' Re: Newbie non blocking socket question $ Re: poor disk I/O performace on ds25$ RE: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25$ Re: poor disk I/O performace on ds25$ RE: poor disk I/O performace on ds25* Re: problem with router---NAT and caching?* Re: problem with router---NAT and caching? Re: Script Kiddie tarpit wanted + Re: There is certainly opportunity in chaos  Time change questions  Re: Time change questions  Re: Time change questions   F ----------------------------------------------------------------------  * Date: Mon, 4 Apr 2005 07:50:07 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)! Subject: Re: %MOUNT-F-SHDWCOPYREQ $ Message-ID: <d2qrjf$lgm$2@online.de>  C In article <1112573090.375991.173850@l41g2000cwc.googlegroups.com>, ' "AEF" <spamsink2001@yahoo.com> writes:    F > Informational would be better than warning, but Fatal tells you thatF > the command has not succeeded yet. And -I- would mean to me that the@ > MOUNT command was taking effect and that the error message wasF > simplying letting me know that a copy operation was likely to begin.H > -F- makes it clear that there is a problem and that the command hasn't > acted yet on the disks.   B I suppose this is the motivation.  But since this is the expected G behaviour, then fatal is a bit off-base.  After all, the whole purpose  - of confirm is to ask a question if necessary.   F There are three unused security levels.  Surely we could have one new  one in VMS 9.0?    ------------------------------  % Date: Mon, 04 Apr 2005 11:48:26 -0500 - From: Hunter Goatley <goathunter@goatley.com> * Subject: ANN: VTfm V2.3-9 is now available* Message-ID: <42516FDA.2070403@goatley.com>  : Thanks to Vladimir Vershinin, VTfm V2.3-9 is now available8 from my FILESERV.  VTfm is a Norton Commander-style file' and directory manager for VT terminals.   % VTfm release 2.3-9 for your FILESERV:   = * Some more enhancements concerning Find, Select and UnSelect 
    functions: =    - possibility of using zero or nigative version numbers in       wildcard masks;9     - VTfm shows now the number of selected files in file        statistics line;<     - VTfm switch OFF the Select (Sel) key when ALL files in"       diectory panel are selected;>     - VTfm switch OFF the Unselect (Rem) key when there are NO(       selected files in directory panel.   * Some minor bug fixes. 1 * User's Guide in PDF format included in the kit.    http://www.process.com/openvms/   4 ftp://ftp.process.com/vms-freeware/fileserv/vtfm.zip9 http://vms.process.com/ftp/vms-freeware/fileserv/vtfm.zip   0 ftp://ftp.tmk.com/vms-freeware/fileserv/vtfm.zip5 http://www.tmk.com/ftp/vms-freeware/fileserv/vtfm.zip    Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/ ; <goathunter@GOATLEY.COM>     http://www.goatley.com/hunter/    ------------------------------  # Date: Mon, 04 Apr 2005 12:26:50 GMT / From: "Joe H. Gallagher" <dtrwiz@ix.netcom.com> 3 Subject: Announcing April 19, 2005 WRUG LUG Meeting , Message-ID: <42513259.6030100@ix.netcom.com>   . /                   Announcing April 19th Meeting .                    Western Reserve Users Group(                         (Northeast Ohio)/                   Local User Group of Encompass   .                   (usual time and usual place)  9 If you are going to attend the LUG meeting on April 19th, ? please RSVP to this message or to Sederholm at TrendCS dot com. 7 We would like to get an estimate of the number who will 9 attend so we can have the appropriate number of slices of = pizza (or other food).  Please RSVP before noon on the day of  the meeting.  8 Topic: Implementation of a 2-Queue model for data backup  ?     In 2002, a client of DRA Consulting decided to migrate from @     a VAXcluster to a networked cluster of Alpha ES45s.  As part@     of this migration, backup procedures were needed that would:?       * maintain compatibility with 10 years of historical data B       * service the needs of the business units that have a common"         need for daily data backup5       * while they insist on different shutdown times <       * and where all demand the shortest possible window of(         "unavailability" due to backups.+     Further, these procedures would need to "       * use off the shelf software9       * be adaptable to changes in the business structure        * be easily maintainable.  	 >     This talk will cover how these constraints were addressed,?     the success of the project over the last 12 months, and the 8     problems that were encountered during the design and     implementation phases.  A     After the presentation there will be some time for discussion A     about other attendees  experiences with this type of problem.   6 Speaker: Denison Rich, President, DRA Consulting, Inc.  :     Denny has been active with DEC equipment and operating=     systems since the mid 70s, starting in 1968 with a PDP8e, >     VMS in 1978, TOPS10 and RSX11 in the 80s, and then back to=     VMS through the present. Since 1999, Denny has been doing 8     OpenVMS consulting for clients in the Cleveland area$     primarily on the Alpha platform.  - Topic: Today's Records Management Environment   7     This presentation is an overview of today's records 8     management environment, including best practices for;     backup/restoration of information systems via remote or ?     local mechanisms, longevity of stored magetic media, pluses 8     and minuses of remote backup for desktops and server=     environments, imaging, Image-on-Demand immediate document ;     retrieval services and data security services including (     HIPAA and Sarbanes-Oxley compliance.  1 Speaker: Dennis Tallerico, President, Assurant RM   >     An industry veteran of 25 years, Dennis Tallerico has been>     engaged in numerous records management practices providing@     software products and services for the legal, pharmacetuical     and other industries.   6     Assurant Records Management is a full service data?     management firm that specializes in offsite data management ?     in paper and electronic format.  A totally accessible, 7x24 8     WWW interface permits clients to do remote backup of7     workstations, servers (Windows, UNIX, OpenVMS) with 8     interactive, user driven requests for automatic (and:     immediate) restoration services for documents (scanned@     images), operating systems, server software even specialized=     restorations of databases and individual mailboxes within      Exchange or Notes.   Date:      Tuesday, April 19, 2005  Time:      4:00 to 7:00 PM 	 Location:      Park Center Plaza III      Lower Level conference room      6050 Oak Tree Blvd.      Independence, OH 44131<     (Call Rich's cell phone at 412-999-6427 if you get lost) Map:     See URL on the Web site    Directions: :     From the area of the interchange of I-480 and I-77, go+     south and take the Rockside Road exit.    >     At the bottom of the ramp, turn right (west) on Rockside.   ;     Go west two blocks (second light); turn left (south) on      Oak Tree Blvd.     See the LUG's web page at   &     http://eisner.decus.org/lugs/wrug/  & Future LUG meetings are scheduled for:       July 12, 2005 (tentatively) "     October 11, 2005 (tentatively)   See you April 19th.      Bill Sederholm     WRUG LUG Chair      Sederholm at TrendCS dot com   ------------------------------  % Date: Mon, 04 Apr 2005 15:45:58 +0200 & From: Paul Sture <paul.sture@decus.ch>" Subject: Re: cloning a system disk, Message-ID: <3bcuolF6ei1j6U1@individual.net>   John Laird wrote:   N > On Sun, 3 Apr 2005 18:59:40 -0400, "Wayne" <bruzeksat@comcastdot.net> wrote: >  >  >>(batch jobs)D >>I believe that it actually uses the FID.  If I submit a batch job N >>HI_WAYNE.COM that prints out HI WAYNE, and then, after submitting, edit the I >>file HI_WAYNE.COM to print out BYE WAYNE, the system will seek out the  J >>originally submitted file and execute that.  Even if I rename the newly I >>edited file to be exactly the same as the original file (same physical  N >>device, directory, name, ext, version, etc.) the system will still look for E >>the original FID to execute, and if it's not there, will error off.  >>) >>That's if memory serves me correctly...  >  > C > No, I think you're right.  However, I think you can get away with  > copy/overlay...  >  I'm pretty sure he's right too.   G In the context of this discussion it is very useful to have batch jobs  G which resumbit themselves with the correct parameters for the next day  I (or whatever time interval is appropriate), and write them so that, say,  ' if P1 .eqs. "SUBMIT" execute that code.   H This enables you to restart failed jobs quickly, and reduces the chance C of human error getting the parameters, log file specifications etc  H wrong. And of course, it makes resubmitting all your jobs after a queue  database rebuild very easy.    ------------------------------  $ Date: Mon, 4 Apr 2005 11:24:19 -04004 From: "Peter Weaver" <newsgroup@weaverconsulting.ca>* Subject: Re: EDT question: SET ENTITY WORD, Message-ID: <3bd4h6F6ibmgoU1@individual.net>  / Phillip Helbig---remove CLOTHES to reply wrote: @ > For many years now, I've used .,?!;:[]()<>*-+=/\{}$^_& as WORDF > delimiters in EDT.  This is pretty much what I want, especially whenG > deleting stuff with the "-" on the keypad.  However, it would be nice > > if, with respect to WRAP, I could specify a different set ofB > characters. For example, if the previous line were one character= > longer, the full stop would be at the beginning of the next C > line---not what I want. Trivial to correct, sure, but it would be @ > nice to have two separate contexts for specifying ENTITY WORD. > A > If TECO can do this, I might switch.  I could even use it in my  > lawnmower as well.  :-)   @ Can you do what you want with macros? (Watch out for line wraps)   !+ !      Word Delimeters Macros  !-, DEFINE MACRO DELIMITERS_TRUNCATE_PROGRAMMING$ FIND=DELIMITERS_TRUNCATE_PROGRAMMINGA INSERT ; DEFINE KEY GOLD CONTROL D AS "EXT DELIMITERS_PROGRAMMING 4 ?*'Press Enter to enter Normal Programming Mode. '." INSERT ; SET NOWRAP  INSERT ; SET TRUNCATE A CHANGE; ISET ENTITY WORD '^Z 9ASC 10ASC 11ASC 12ASC 13ASC 27ASC I  ():,.+/='^Z EX5 ! 9 <TAB> ; 10 LF ; 11 VT ; 12  FF ; 13 CR ; 27 <ESC>  ! # DEFINE MACRO DELIMITERS_PROGRAMMING  FIND=DELIMITERS_PROGRAMMING E INSERT ; DEFINE KEY GOLD CONTROL D AS "EXT DELIMITERS_WORD_PROCESSING 1 ?*'Press ENTER to enter Word_Processing Mode. '."  INSERT ; SET NOWRAP  INSERT ; SET NOTRUNCATE A CHANGE; ISET ENTITY WORD '^Z 9ASC 10ASC 11ASC 12ASC 13ASC 27ASC I  ():,.+/='^Z EX5 ! 9 <TAB> ; 10 LF ; 11 VT ; 12  FF ; 13 CR ; 27 <ESC>  ! ' DEFINE MACRO DELIMITERS_WORD_PROCESSING  FIND=DELIMITERS_WORD_PROCESSING * INSERT ; DEFINE KEY GOLD CONTROL D AS "EXT7 DELIMITERS_TRUNCATE_PROGRAMMING ?*'Press ENTER to enter  Truncate_Programming Mode. '." INSERT ; SET WRAP 78 INSERT ; SET TRUNCATE H CHANGE; ISET ENTITY WORD '^Z 9ASC 10ASC 11ASC 12ASC 13ASC 27ASC I -.,'^Z EX !+ SET MODE CHANGE  DELIMITERS_PROGRAMMING
 FIND =MAIN       --   Peter Weaver Weaver Consulting Services Inc.  Canadian VAR for CHARON-VAX  www.weaverconsulting.ca    ------------------------------  % Date: Mon, 04 Apr 2005 16:31:26 +0100  From: Roy Omond <Roy@Omond.net> = Subject: Re: Finding The Right Path For Your VMS Applications 4 Message-ID: <d2rmke$ds1$1$8300dec7@news.demon.co.uk>   Roy Omond wrote:   > Transoft spouted garbage:  >  >> [...drivel snipped...]  >> [...other stuff snipped...]   *Sigh* again ...   I got a reply from Transoft:  I "sorry if I offended you, we would be more than happy to migrate windows  G to VMS, just we have not been asked to do so up until now. Please note  E it is HP who are recommending migration to Itainium. So please don't   shoot this messenger."    < Note the implication that it is HP recommending migration to> Itainium (sic), and therefore recommending migration away from VMS.  *Sigh* ... FUDsters.  	 Roy Omond  Blue Bubble Ltd.   ------------------------------  % Date: Mon, 04 Apr 2005 09:20:28 +0100 - From: John Laird <nospam@laird-towers.org.uk> * Subject: Re: indicator lights on SBB disks8 Message-ID: <94u151taam4ipjvuqcr2koaebeaaetec9b@4ax.com>  L On Mon, 04 Apr 2005 01:29:08 -0400, Dave Froble <davef@tsoft-inc.com> wrote:   >Tom Linden wrote:D >> On Sun, 3 Apr 2005 20:23:10 +0000 (UTC), Phillip Helbig---remove = >> CLOTHES  to reply <helbig@astro.multiCLOTHESvax.de> wrote:  >>  H >>> On the Storage Works building-block (SBB) disks (e.g. for use in theK >>> BA35x shelves), there are usually (always?) two indicator lights at the L >>> front.  However, what they mean seems to vary with the disk model.  Some> >>> have one light on all the time, some only during activity. >>   >>  K >> I had 4GB drives in the BA356 cannisters, took them out and replced with K >> 74GB Seagate drives, works like a charm.  of course, I only get 40MB not % >> the 160 the drives are capable of.  >  >Huh?  > , >Want to try that again, doesn't make sense.  8 Transfer rates, I think.  40Mb/sec instead of 160Mb/sec.   --  " I'd kill for a Nobel Peace Prize.    Mail john rather than nospam...    ------------------------------  % Date: Mon, 04 Apr 2005 07:45:22 -0700 # From: "Tom Linden" <tom@kednos.com> * Subject: Re: indicator lights on SBB disks( Message-ID: <opsoppxwnazgicya@hyrrokkin>  0 On Mon, 04 Apr 2005 09:20:28 +0100, John Laird  # <nospam@laird-towers.org.uk> wrote:   I > On Mon, 04 Apr 2005 01:29:08 -0400, Dave Froble <davef@tsoft-inc.com>    > wrote: >  >> Tom Linden wrote:D >>> On Sun, 3 Apr 2005 20:23:10 +0000 (UTC), Phillip Helbig---remove> >>> CLOTHES  to reply <helbig@astro.multiCLOTHESvax.de> wrote: >>> I >>>> On the Storage Works building-block (SBB) disks (e.g. for use in the J >>>> BA35x shelves), there are usually (always?) two indicator lights at   >>>> theJ >>>> front.  However, what they mean seems to vary with the disk model.   	 >>>> Some ? >>>> have one light on all the time, some only during activity.  >>>  >>> I >>> I had 4GB drives in the BA356 cannisters, took them out and replced    >>> withJ >>> 74GB Seagate drives, works like a charm.  of course, I only get 40MB   >>> not & >>> the 160 the drives are capable of. >> >> Huh?  >>. >> Want to try that again, doesn't make sense. > : > Transfer rates, I think.  40Mb/sec instead of 160Mb/sec. > C Yes, sorry.  Also I don't know how I manged to post it twice. Sorry  about that too.    ------------------------------  $ Date: Mon, 4 Apr 2005 17:39:53 +1000# From: "Gremlin" <not-here@all.mate>  Subject: Re: InfoTower- Message-ID: <4250ef44@duster.adelaide.on.net>   " -----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1  9 It doesn't - I tried - so, anyone in Australia want one?? 8 "John E. Malmberg" <wb8tyw@qsl.network> wrote in message+ news:j9-dnRBV6qgrWM3fRVn-pg@adelphia.com...  > Gremlin wrote: >>? >> I suppose that it does, I was using it on NT4 and VMS, but I E >> haven't seen a W2K or W2K3 compatible client.  Can you point me in  >> the >> direction of one? > 7 > http://h71000.www7.hp.com/pathworks32/latestinfo.html  > ? > The infoserver client is listed as an unsupported or obsolete B > feature.  I have no idea if it would function on a W2K or a W2K3 > compatible client. > -John  > wb8tyw@qsl.network > Personal Opinion Only    -----BEGIN PGP SIGNATURE-----  Version: PGP 8.1  @ iQA/AwUBQlDvRbBJFNOSUYtbEQI39wCg1ABSQ3jAaIzGiksM1/W2rJe7u04Anifo FPJvwOKJUeUnPg2BmuPvALYv =eCax  -----END PGP SIGNATURE-----    ------------------------------   Date: 4 Apr 2005 09:26:44 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: InfoTower3 Message-ID: <QofOX65HI5Ko@eisner.encompasserve.org>   U In article <4250a81b$1@duster.adelaide.on.net>, "Gremlin" <not-here@all.mate> writes:  > $ > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > 
 > Hi Larry > F > I suppose that it does, I was using it on NT4 and VMS, but I haven't@ > seen a W2K or W2K3 compatible client.  Can you point me in the > direction of one?   ? If Windows is not backward compatible, that is beyond solution. @ Since DEC/Compaq/HP no longer sells those clients, eBay would be? your only source, but it sounds like you already have a Windows ( clients, and there will not be any more.   ------------------------------  # Date: Mon, 04 Apr 2005 17:10:11 GMT / From: "Bruno van Dooren" <microvax@hotmail.com>  Subject: microvax ; Message-ID: <Txe4e.55132$wL4.4122720@phobos.telenet-ops.be>    Hi all,   L i have a microvax 3400 in good working condition -together with a terminal, ! keyboards and some add-in boards.   M i am thinking about selling it on ebay, but the problem is that i don't have  1 any idea about what would be a fair price for it.   & can anyone please give me some advice?  
 kind regards,      Bruno.     ------------------------------   Date: 4 Apr 2005 07:25:35 -0700 ( From: "Lawrie" <stroker_ace@hotmail.com>, Subject: Newbie non blocking socket questionC Message-ID: <1112624735.388390.102580@g14g2000cwa.googlegroups.com>    Hi,   G I'm fairly new to VMS/C socket programming and wondered if anyone could  help me with the following.   B I am working on a legacy application that reads data from an input@ queue, formats the data and writes it to two serial devices. TheC application is synchronous and can be summarised in pseudo code as:    do {      read_input_queue;    format_data; 8    write_to_serial_devices; //Blocking write using qio's   } (while !error)  G If a write to a single serial device fails for any reason the device is G considered 'dead' no attempt is made to re-establish the connection. If E both serial devices the application writes to are 'dead' then this is G considered a fatal error. The application will gracefully exit invoking  a switch to a backup system.  A I am required to add the functionality to write to a configurable F number of remote hosts using TCP/IP. The functionality to write to theF serial devices must also remain (untouched if at all possible) but may+ be switched of with a configuration change.   B I plan to use ioctl() to set the tcp sockets to non blocking. As IF write a message to each socket I check the return status and sent byteB count. If the byte count returned is not the total expected then I/ attempt to send the unsent part of the message.   ' My new application can be summarised as    do {      read_input_queue;    format_data;   #    if (serial writes switched on) { ;       write_to_serial_devices; //Blocking write using qio's     }  @    write_to_sockets;        //Non blocking writes using socket.h,                             //functionality. } (while !error)  D If the write fails due to a network failure/remote disconnect then IF should attempt to recover the connection and this is where my concerns lie!  E Can anyone answer or provide hints regarding the following questions?   F 1. Are connection attempts to a remote port non blocking? I am worriedC that any attempts to (re) connect may cause the application to lock F until a TCP session is established. Obviously this may never happen if; the remote host is disabled causing the application to lock 
 indefinitely.   E 2. The application is required to send data only. Is there any way of E preventing a remote host from sending data to the application? If not F would any unread packets cause the socket input buffer to overflow and cause an error?    Many Thanks for any help   Lawrie   ------------------------------  % Date: Mon, 04 Apr 2005 18:18:13 +0200 , From: Albrecht Schlosser <ajs856@tiscali.de>0 Subject: Re: Newbie non blocking socket question, Message-ID: <bcpr2d.rs5.ln@news.hus-soft.de>  
 Lawrie wrote:    > [...] D > I plan to use ioctl() to set the tcp sockets to non blocking. As IH > write a message to each socket I check the return status and sent byteD > count. If the byte count returned is not the total expected then I1 > attempt to send the unsent part of the message.  > ) > My new application can be summarised as  >  > do { >  >    read_input_queue; >    format_data;  > % >    if (serial writes switched on) { = >       write_to_serial_devices; //Blocking write using qio's  >    } > B >    write_to_sockets;        //Non blocking writes using socket.h. >                             //functionality. > } (while !error) > F > If the write fails due to a network failure/remote disconnect then IH > should attempt to recover the connection and this is where my concerns > lie! > G > Can anyone answer or provide hints regarding the following questions?  > H > 1. Are connection attempts to a remote port non blocking? I am worriedE > that any attempts to (re) connect may cause the application to lock H > until a TCP session is established. Obviously this may never happen if= > the remote host is disabled causing the application to lock  > indefinitely.   Q Usually there's a timeout of about 75 seconds for a connection attempt, but this  N   may be configurable (I don't know). From the linux man page (man 2 connect):           EINPROGRESS O                The socket is non-blocking and the  connection  cannot  be  com- O                pleted  immediately.  It is possible to select(2) or poll(2) for O                completion by selecting the socket  for  writing.  After  select O                indicates  writability,  use  getsockopt(2) to read the SO_ERROR O                option at level SOL_SOCKET to  determine  whether  connect  com- O                pleted   successfully   (SO_ERROR  is  zero)  or  unsuccessfully O                (SO_ERROR is one of the usual error codes listed here,  explain- /                ing the reason for the failure).   L http://node1.yo-linux.com/cgi-bin/man2html?cgi_command=connect&cgi_section=2    O It seems that using a non-blocking socket and polling for the connection would  	 help you.     G > 2. The application is required to send data only. Is there any way of G > preventing a remote host from sending data to the application? If not H > would any unread packets cause the socket input buffer to overflow and > cause an error?   N Once the TCP window would be full, the sending TCP/IP stack would be blocked, Q maybe with 0 bytes output count when trying to write (interesting question, does   anybody know that for sure?).   M Maybe you could "shutdown" the socket for read (man 2 shutdown), but I don't  $ know if that would be what you need.  N http://node1.yo-linux.com/cgi-bin/man2html?cgi_command=shutdownt&cgi_section=2   HTH  Albrecht   ------------------------------  # Date: Mon, 04 Apr 2005 08:09:25 GMT 4 From: Mike Rechtman <michael.rechtman.nospam@hp.com>- Subject: Re: poor disk I/O performace on ds25 & Message-ID: <42512063.130991D4@hp.com>   Nitendra wrote:  > F > can any share the result of copying some 200MB file from one disk to* > one another disk on system running OVMS. > H > pls share your system type and OS version along with time to copy that > fileB > i will try that on my system and compare with what you have got. > 	 > Regards   > AlphaServer 300 4/266, hardware model type 1593 OpenVMS V7.3-1" Memory size: 256MB (8192 byt/pag)    $ty x.x $ $ dir/siz DKA300:[BACKUP]DKA100.BCK; $ show time = $ copy DKA300:[BACKUP]DKA100.BCK; EYAL$DKA600:[000000]tmp.tmp  $ show time  @x.x   Directory DKA300:[BACKUP]    DKA100.BCK;3         3117618    Total of 1 file, 3117618 blocks.    4-APR-2005 10:39:36    4-APR-2005 10:57:03 $ MB = 3117618/2048  $sho sym MB 1   MB = 1522   Hex = 000005F2  Octal = 00000002762  $sec = (57-40)*60 + (63-36)  $sho sym sec2   SEC = 1047   Hex = 00000417  Octal = 00000002027% $time_for_200MB = (1047 * 200) / 1522  $sho sym time_for_200MB 5   TIME_FOR_200MB = 137   ! <<- seconds for 200MB copy    Both disks are RZ28-VW   --  E --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that. ? Mike Rechtman                            *rechtman@tzora.co.il* F Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------  -----BEGIN GEEK CODE BLOCK-----  Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------    ------------------------------  $ Date: Mon, 4 Apr 2005 07:02:54 -0400' From: "Main, Kerry" <kerry.main@hp.com> - Subject: RE: poor disk I/O performace on ds25 R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB59568C@tayexc19.americas.cpqcorp.net>   > -----Original Message-----@ > From: Mike Rechtman [mailto:michael.rechtman.nospam@hp.com]=20 > Sent: April 4, 2005 4:09 AM  > To: Info-VAX@Mvb.Saic.Com / > Subject: Re: poor disk I/O performace on ds25  >=20 > Nitendra wrote:  > >=20H > > can any share the result of copying some 200MB file from one disk to, > > one another disk on system running OVMS. > >=20@ > > pls share your system type and OS version along with time=20 > to copy that > > fileD > > i will try that on my system and compare with what you have got. > >=20 > > Regards  >=20@ > AlphaServer 300 4/266, hardware model type 1593 OpenVMS V7.3-1& > Memory size: 256MB (8192 byt/pag)=20 >=20 [snip..]  C Mike - I noticed your test uses V7.3-1. Keep in mind that VMS 7.3-2 F increased default buffer sizes for COPY command and hence, you may seeE some differences (amount tbd) wrt the original threads question about  V7.3-2.     Reference: (V7.3-2 new features)D http://h71000.www7.hp.com/doc/732FINAL/aa-rv8xa-te/00/00/19-con.htmlG "Performance of the DCL commands COPY and SEARCH has been significantly H improved. It now takes less time and fewer computing resources to copy a file or perform a search."   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Mon, 04 Apr 2005 05:18:49 -0700 ( From: Jeff Cameron <roktsci@comcast.net>- Subject: Re: poor disk I/O performace on ds25 / Message-ID: <BE767EB9.AF59%roktsci@comcast.net>    On 4/3/05 9:52 PM, in article < 1112590336.904305.163640@g14g2000cwa.googlegroups.com, "AEF" <spamsink2001@yahoo.com> wrote:    >  > Jeff Cameron wrote: G >> On 4/3/05 4:51 PM, in article MtDFgd8imW8V@eisner.encompasserve.org,  > "Larry, >> Kilgallen" <Kilgallen@SpamCop.net> wrote: >>  @ >>> In article <BE756316.ADB0%roktsci@comcast.net>, Jeff Cameron! >>> <roktsci@comcast.net> writes: " >>>> On 4/3/05 8:25 AM, in article/ > Cs2pcQljdA$B@eisner.encompasserve.org, "Larry . >>>> Kilgallen" <Kilgallen@SpamCop.net> wrote: >>> B >>>>> As I recall there was a time when highwater marking was less > efficient,D >>>>> but the design and current implementation, as I understand it, > are  >>>>> pretty efficient.  > 
 > Correct. >  >>>>  G >>>> I was always under the impression that the blocks were zeroed when < >>>> allocated, to prevent allocating then reading old data. >  > G > In older versions of VMS (prob. pre-V5, but I'm not sure, I'd have to D > check the manuals if I can even find them for such old releases!). >  > F >>> Zeroing on allocation would not correspond to the name "High Water > Marking", G >>> since the High Water Mark is the point beyond which the data is not  > from >>> this user. >>  H >> I can appreciate the meaning behind the nomenclature, but that is not > whatF >> is happening. I just got finished writing a test program in FORTRAN	 > where I E >> used SYS$ASSIGN to open a file of 2048 blocks and used SYS$QIOW to 
 > read theC >> blocks, and when done on a disk with HFWM active all blocks were  > zero, but 4 >> were not on the same disk when HFWM was disabled. >  > G > They were zero probably because the system is showing you zeros after D > the mark (which I assume is at the beginning of the file), when inB > reality, there is probably random data there. Thus the highwaterI > marking is working. I mean, if you were able to read the random data in F > your newly created file, the high water marking wouldn't be working! > I I am afraid I must still disagree, As I am using QIO to bypass RMS, and I K have not written anything, so there is no filemark. I find it unlikely that H after several runs in each case (at least 2 dozen each  with and withoutB HFWM) with no inconsistencies to the outcome to suggest otherwise. > + >> Also the System manager's manual states:  >>  M > (http://h71000.www7.hp.com/doc/731FINAL/6017/6017pro_077.html#index_x_4370) C >> "High-water marking creates some overhead; the system erases the 
 > previousF >> contents of the disk  blocks allocated every time a file is created > or
 >> extended."  >  > D > Uh, you quoted out of context. Here is the same quote, in context: >  > [begin quote] G > Disable file system high-water marking---This security feature is set H > by default when a volume is initialized to guarantee that users cannot" > read data they have not written. > F > For nonshared sequential files, the performance impact of high-water > marking is  > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > I > minimal. However, for files of nonsequential format, high-water marking G > creates some overhead; the system erases the previous contents of the A > disk blocks allocated every time a file is created or extended. I > Disabling the feature improves system performance by a variable amount, % > depending on the following factors: & > How frequently new files are createdC > For indexed and relative files, how frequently existing files are 
 > extended > How fragmented the volume is > B > Be sure to consider the security implications before you disable > high-water marking. A > To disable high-water marking, you can specify the /NOHIGHWATER G > qualifier when initializing the volume, or you can disable high-water B > marking with the DCL command SET VOLUME in the following format:/ > SET VOLUME/NOHIGHWATER_MARKING device-name[:] 
 > [end quote]  >  > D > Note that it says there is little impact for sequential files. TheG > overhead is significant for files of NON-sequential format and shared  > files. >  >> From the security manual: > ' > Prevention Through High-Water Marking  > C > High-water marking refers to a technique that tracks the furthest H > extent to which each file has been written and prohibits user attempts$ > at reading data beyond that point. > A > The operating system implements true high-water marking for all I > sequential, exclusively accessed files, such as the set of files output H > from various text editors, compilers, and linkers, that is, most filesE > a process writes. The high-water mark is updated in the file header I > whenever the logical end-of-file mark is updated (usually when the file 
 > is closed).  > F > For shared files (both indexed and sequential), the operating systemH > uses the principle of erase-on-allocate to achieve a result similar to@ > true high-water marking. When a file is about to be created orH > extended, the system determines how much disk space (the extent of theH > file) is required and applies the security erasure pattern of zeros toH > the areas (extents) it allocates for writing. The file is then writtenI > into the area just erased for it. Thus, if any user gains access to the G > file (including its full extent) and attempts to read the area beyond G > where the file has been written, only the data security erase pattern  > is readable. > K I don't believe I took the quote out of context. I do not disagree that for K sequential files, the remainder of the block where the file mark is written K must also be zeroed. It is just that section that is done when the filemark H is written thus the line that the performance impact is minimal. That isH because the remaining blocks that have been allocated were zeroed at the# file's initial creation, or extent.e  G Another very simple test you can do with a sequential file to show that3J blocks are zeroed on allocation is to time a file open operation where youK preallocate a large number of blocks. You will find that when you have HFWMaK enabled the open takes longer, where the time is proportional to the numberoK of disk blocks allocated. Then do the same test with HFWM disabled, and allb/ the open file ops take the same amount of time.   L Where HFWM was optimized in sequential files was when your first filemark isK written, then zeroing would not have to occur again except on file extents.a  K In summary I do agree that HFWM works as you describe for sequential files,oI to zero out data between the logical EOF and the physical EOF. It is also L true that for all files (sequential and otherwise) that blocks are zeroed on allocation/extent.   ------------------------------  # Date: Mon, 04 Apr 2005 13:07:42 GMTc4 From: Mike Rechtman <michael.rechtman.nospam@hp.com>- Subject: Re: poor disk I/O performace on ds25-& Message-ID: <4251664D.2379DF7D@hp.com>   Main, Kerry wrote: >  > > -----Original Message-----? > > From: Mike Rechtman [mailto:michael.rechtman.nospam@hp.com]N > > Sent: April 4, 2005 4:09 AMq > > To: Info-VAX@Mvb.Saic.Comc1 > > Subject: Re: poor disk I/O performace on ds25p > >> > > Nitendra wrote:0 > > >.J > > > can any share the result of copying some 200MB file from one disk to. > > > one another disk on system running OVMS. > > >-? > > > pls share your system type and OS version along with timen > > to copy that
 > > > fileF > > > i will try that on my system and compare with what you have got. > > >a
 > > > Regardsa > >rB > > AlphaServer 300 4/266, hardware model type 1593 OpenVMS V7.3-1% > > Memory size: 256MB (8192 byt/pag)  > >s
 > [snip..] > E > Mike - I noticed your test uses V7.3-1. Keep in mind that VMS 7.3-2oH > increased default buffer sizes for COPY command and hence, you may seeG > some differences (amount tbd) wrt the original threads question about9	 > V7.3-2.- > " > Reference: (V7.3-2 new features)F > http://h71000.www7.hp.com/doc/732FINAL/aa-rv8xa-te/00/00/19-con.htmlI > "Performance of the DCL commands COPY and SEARCH has been significantly0J > improved. It now takes less time and fewer computing resources to copy a > file or perform a search." > 	 > Regardsn >  > Kerry Main > Senior Consultanto > HP Services Canada > Voice: 613-592-4660u > Fax: 613-591-4477e > kerryDOTmainAThpDOTcom > (remove the DOT's and AT)  > & > "OpenVMS has always had integrity ..  > Now, Integrity has OpenVMS .."   You are right - mea culpa. repeated test:( OpenVMS V7.3-2  on AlphaServer 800 5/500 $1$DKB600: , device type RZ29B% $1$DKC100: , device type DEC RZ1DB-CAL $ ty x.x  $ $ dir/siz DKC100:[000000]ADB074.RAW; $ s = f$time()A $ copy DKC100:[000000]ADB074.RAW; $1$DKB600:[USERS.RECHTMAN]z.z;0i $ e = f$time() $ diff = f$delta(s,e)f $ sho sym diff $! $ dele z.z; /loi $ @x.x$ $ dir/siz DKC100:[000000]ADB074.RAW;   Directory DKC100:[000000]s   ADB074.RAW;1        600.58MB   Total of 1 file, 600.58MBf $ s = f$time()A $ copy DKC100:[000000]ADB074.RAW; $1$DKB600:[USERS.RECHTMAN]z.z;0z $ e = f$time() $ diff = f$delta(s,e)/ $ sho sym diff   DIFF = "   0 00:03:18.89"G $! $ dele z.z; /loaD %DELETE-I-FILDEL, $1$DKB600:[USERS.RECHTMAN]Z.Z;1 deleted (600.58MB)% $seconds_per_200MB = ((3*60) + 18) /3- $sho sym seconds_per_200MB>   SECONDS_PER_200MB = 66   Hex = 00000042  Octal = 00000000102   Time is halved in V7.3-2!c   Mike   - E --------------------------------------------------------------------- E Usual disclaimer: All opinions are mine alone, perhaps not even that. ? Mike Rechtman                            *rechtman@tzora.co.il*OF Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------  -----BEGIN GEEK CODE BLOCK-----k Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------y   ------------------------------  % Date: Mon, 04 Apr 2005 08:29:48 -0600k4 From: Norman Lastovica <norman.lastovica@oracle.com>- Subject: Re: poor disk I/O performace on ds25e* Message-ID: <42514F5C.2D1C8015@oracle.com>  4 COPY is just about the worst possible way to measure6 performance in that it is basically a synchrous copy. 4 (I'm told that there are plans afoot for VMS V8.3 to make copy really rip however).  8 If you are interested in actually measuring performance,5 try BACKUP or CONVERT to do the copy as step 1.  Thent2 pick up the FAST_IO_COPY program (one reference is7 http://wwwvms.mppmu.mpg.de/vmssig/src/C/FAST_IO_COPY.C)  and try that as well.f   Nitendra wrote:e > F > can any share the result of copying some 200MB file from one disk to* > one another disk on system running OVMS. > H > pls share your system type and OS version along with time to copy that > fileB > i will try that on my system and compare with what you have got. > 	 > Regards    -- m	 - - - - -a0  opinions expressed here are mine and mine alone.  and certainly are not intended in any way to 0  express or represent any opinions or commitment  of oracle corporation.i  *  norman lastovica / oracle rdb engineering   ------------------------------   Date: 4 Apr 2005 09:38:48 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)n- Subject: Re: poor disk I/O performace on ds25c3 Message-ID: <i4Q0tsJXozr8@eisner.encompasserve.org>h  Z In article <BE76088B.AE0F%roktsci@comcast.net>, Jeff Cameron <roktsci@comcast.net> writes:  L > I can appreciate the meaning behind the nomenclature, but that is not whatM > is happening. I just got finished writing a test program in FORTRAN where IlM > used SYS$ASSIGN to open a file of 2048 blocks and used SYS$QIOW to read theoL > blocks, and when done on a disk with HFWM active all blocks were zero, but3 > were not on the same disk when HFWM was disabled.   5 That is exactly the description of Highwater Marking.m< But it does not tell you whether those blocks were zeroed on# disk before you tried to read them.m+ And that does not mimic any useful program.e  C For a test of of overhead from Highwater Marking, try the followingc1 both with Highwater Marking enabled and disabled.e   	Create a new file& 	Write 100000 blocks of non-zero bytes 	Close the filea# 	Report the resources and time used   K A different test would be to allocate the file to a large size on creation.h   ------------------------------  * Date: Mon, 4 Apr 2005 17:09:23 +0000 (UTC)- From: klewis@OMEGA.MITRE.ORG (Keith A. Lewis) - Subject: RE: poor disk I/O performace on ds25k- Message-ID: <d2rsc3$td$1@newslocal.mitre.org>    Kilgallen@SpamCop.net (Larry Kilgallen) writes in article <Cs2pcQljdA$B@eisner.encompasserve.org> dated 3 Apr 2005 10:25:23 -0500:C >If you have an application that explicit asks for a big allocationoA >increase and then writes data near the end of that space without.D >having written data in between then the effect Kerry describes will5 >take effect.  But those applications are quite rare.e  K I curse every time I type the following without disabling highwater markingC first:  . $ MC SYSGEN CREATE PAGEFILE2.SYS /SIZE=2000000  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  * Date: Mon, 4 Apr 2005 07:45:56 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)3 Subject: Re: problem with router---NAT and caching?e$ Message-ID: <d2qrbk$lgm$1@online.de>  E In article <1112561687.f337242c33721a5bb400ea39bd05c79f@teranews>, JFt- Mezei <jfmezei.spamnot@teksavvy.com> writes: )  F > 1- Your terminal (decwindows server) may have a security config thatI > accepts only calls from the old IP address. So change the IP address of G > the host running the application, and the terminal may not accept the:E > connection request from the app because it comes from an unknown IP> > address.     OK.   : > (The DECWIDNOWS software has a security memnnu where you7 > configure who is allowed to connect to that terminal)   E Do you mean the desktop security menu, where one can specify DECNET, tE LOCAL and TCPIP as transport and username and password (at least for H2 DECNET) of the remote account?  If so, that is OK.  E > 2- If you are changing the IP address of the terminal, there may berH > issues with the aplication already listening to the old IP address andH > not listening to the new IP address. (not sure how this works when you: > change the IP address from under a running application).  F So in this case a reboot, or at least a restart of DECwindows, should % clear things up.  I'll give it a try.e   ------------------------------   Date: 4 Apr 2005 09:58:25 -0500  From: briggs@encompasserve.org3 Subject: Re: problem with router---NAT and caching?u3 Message-ID: <Qb7OK$YfrBv3@eisner.encompasserve.org>   r In article <1112402124.167edc32092d62735e61684a0472d1a9@teranews>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes:1 > Phillip Helbig---remove CLOTHES to reply wrote:h > F >> I suppose somewhere within the router there is a mapping between IP >> address and MAC address.  > D > And in your vax too. It is called the ARP table. (TCPIP SHOW ARP). > H > When a node first needs to talk to another node in the same IP subnet,G > it sends a broadcast "would IP 10.1.2.3 please stand up ?".  Then, IPrJ > 10.1.2.3 responds to the sender with "I am 10.1.2.3" and the sender thenJ > knows the ethernert address associated with node 10.1.2.3. This entry inI > the ARP table lasts for a certain amount of time. During this time, the>G > sender need not do any ARP requests since the association between MACn > and IP is still valid.  C Sound diagnosis and good explanation.  I have only a little to add.f  @ A common case is that you have an IP address that you are moving@ from one device to another.  Maybe you're testing.  Maybe you'reG swapping out hardware and don't want to have to mess with configurationa files on a bazillion desktops.  D Anyway, you've moved the IP address and nothing is working.  And, as> JF said, the reason is that the other equipment on the segmentD still has the old MAC address cached in their individual ARP tables.  H You can wait for those ARP cache entries to time out.  But that can take" a long time.  Hours in some cases.  A Or you can go out and visit each piece of equipment and flush theT= tables by hand.  But that can be a lot of equipment to visit,*G a lot of passwords to know and a lot of ARP flush commands to remember.   E What I generally do in these cases (just did one a couple days ago asSC it turns out) is to log onto the piece of equipment with the new IP;@ address, flush its ARP cache and then do a PING sourced from theF new IP address and destined for a random (and preferably non-existent)) workstation on the local network segment.   J On a Cisco router with new IP address 10.37.33.246, netmask 255.255.224.0:   	router# clear arp 	router# ping 10.37.33.3 	Type escape sequence to abort.^D 	Sending 5, 100-byte ICMP Echos to 10.37.33.3, timeout is 2 seconds: 	.....  	Success rate is 0 percent (0/5) 	router#  @ My piece of equipment will check its ARP cache and find no entry> for the target IP address.  It will then issue an ARP request,? "would the node with IP address x.y.z.w please stand up".  This @ request will go out as an ethernet layer broadcast.  In additionB to the requested IP address, this ARP request will also contain myD IP address and my MAC address.  This request will be received by allA stations on the local segment.  All of these stations will update-A their ARP caches with the my machine's IP address and its current  MAC address.  F Voila -- connectivity restored.  And I only had to visit the node that I was working on anyway.  H Note:  You flush the ARP cache first just in case you are trying to pingD        a machine that actually exists or that has existed.  If thereF        were an entry in your ARP cache, you'd never send the broadcast        that fixes things up.  E        You ping a non-existent machine to avoid the corner case where D        the target machine sends its own ARP request between the time@        you flush your cache and the time you try your ping, thusB        populating your ARP cache and eliminating the need for your        ARP request.n   	John Briggs   ------------------------------   Date: 04 Apr 2005 13:45:37 GMT/ From: Thierry Dussuet <thierry@dussuet.lugs.ch>0( Subject: Re: Script Kiddie tarpit wanted0 Message-ID: <slrnd52h82.1gm.thierry@MARS.Family>  2 On 2005-04-03, Doc. <doc@openvms-rocks.com> wrote:I > %NEWS-I-NEWMSG, Paul Sture wrote in news:3b98r7F6eaficU1@individual.netF >l >> Thierry Dussuet wrote:t >>  A >> from a lugs.ch address! (That's Linux User Group Switzerland).l >> e
 >> Ooooh. :-)  >C  > Thierry is a Deathrow convert. >LB > He's responsible for me installing APRIL_FOOL.COM on Friday. :-)   Hey but that was fun! :-)t   Thierry5   ------------------------------  $ Date: Mon, 4 Apr 2005 08:43:51 -0400# From: "John Smith" <a@nonymous.com> 4 Subject: Re: There is certainly opportunity in chaos, Message-ID: <CsOdnTjupqYMq8zfRVn-2Q@igs.net>   Tom Linden wrote:4G > On Sun, 03 Apr 2005 00:53:55 -0500, Dave Froble <davef@tsoft-inc.com>G > wrote: >vG >> Going afield some consider another example.  Some years ago ChryslerdB >> was in trouble.  Lee Iacocoa (probably misspelled) came in fromA >> Ford, partly because he didn't get the head job there, to turn6@ >> things around.  For his salary, he asked for $1 a year, until >> Chrysler became profitable. >-F > AIR, that wasn't the whole story, he got $5B in loan guarantees from
 > CongressE > and his $1/year was a ruse, he ultimately got considerable back payD > later on.z	 > Iacoccak     I've got no problem with that.  E During a period of low market activity and slim profits, the BoD of a4K private stock brokerage firm I once worked at made the decision that during  the period: & a) no bonuses would be paid to anyone,1 b) all senior managment would take a 50% pay cut,tD c) anyone below senior management but VP or above got a 40% pay cut," d) all managers got a 30% pay cut,; e) all non-managerial professional staff got a 20% pay cut,l% f) all support staff had no pay cuts.   K This lasted about 6 months. When the market turned around, everyone who hadlH their pay cut received all their back pay + interest and an extra 20% on their bonuses.   --F OpenVMS - The never advertised operating system with the dwindling ISV base.    ------------------------------   Date: 4 Apr 2005 05:57:56 -0700  From: tadamsmar@yahoo.comh Subject: Time change questionsB Message-ID: <1112619476.518981.28390@g14g2000cwa.googlegroups.com>  2 (My springforward.com failed, apparently due to my/ earlier upgrade to 7.3.2. The command procedureI* sys$manager:utc$configure_tdf is obsolete.  , I disabled ntp and just used SET TIME to get the times close for now.)+   What is the replacement for:  & @sts$manager:utc$configure_tdf -240 60  + and, my X-windows clocks now display a time-, that is one hour later than the system time.$ Why is that?  How can I change that?   Thanks!n   ------------------------------  # Date: Mon, 04 Apr 2005 13:49:48 GMTj3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)t" Subject: Re: Time change questions2 Message-ID: <0Cb4e.2855$L27.2532@news.cpqcorp.net>  ^ In article <1112619476.518981.28390@g14g2000cwa.googlegroups.com>, tadamsmar@yahoo.com writes:3 >(My springforward.com failed, apparently due to my 0 >earlier upgrade to 7.3.2. The command procedure+ >sys$manager:utc$configure_tdf is obsolete.i >t- >I disabled ntp and just used SET TIME to gete >the times close for now.) >l >What is the replacement for:l > ' >@sts$manager:utc$configure_tdf -240 60: >P, >and, my X-windows clocks now display a time- >that is one hour later than the system time.v% >Why is that?  How can I change that?-     Execute-  &     $ @SYS$MANAGER:UTC$TIME_SETUP SHOW  A If AUTO_DLIGHT_SAV is not set to "1", please consider setting it,jG so that OpenVMS will make the seasonal DAYLING<->STANDARD time changes. E (Don't for get to add AUTO_DLIGHT_SAV to MODPARAMS.  For compatiblitylE with older systems, the default is "0", so it will be reset during an1+ OpenVMS upgrade if it is not in MODPARAMS.)s  I If the rest of the information is incorrect, or maybe just for good luck,l" Execute                             #     $ @SYS$MANAGER:UTC$TIME_SETUP  o  M to reset the time variables.  (This is the replacement for utc$configure_tdf eI and does quite a bit more -- some of which may not strictly be necessary,a) but is resonably quick and does no harm.)w  0 (You may have to re-start your x-window clocks.)  F You might also want to look at SYS$EXAMPLES:DAYLIGHT_SAVING.COM, whichD is much changed from the OLD_DAYLIGHT_SAVING.COM that is still there for reference.  B The OpenVMS FAQ hs several topics related to Dayling Savings time.3 See http://h71000.www7.hp.com/faq/openvms_faq.html.o   -- iJ       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------   Date: 4 Apr 2005 09:18:10 -0700m From: tadamsmar@yahoo.comm" Subject: Re: Time change questionsC Message-ID: <1112631490.299361.264790@z14g2000cwz.googlegroups.com>.   hammond@not wrote:D > In article <1112619476.518981.28390@g14g2000cwa.googlegroups.com>, tadamsmar@yahoo.com writes:f5 > >(My springforward.com failed, apparently due to my 2 > >earlier upgrade to 7.3.2. The command procedure- > >sys$manager:utc$configure_tdf is obsolete.t > >s/ > >I disabled ntp and just used SET TIME to get  > >the times close for now.) > >; > >What is the replacement for:e > >d) > >@sts$manager:utc$configure_tdf -240 60w > >!. > >and, my X-windows clocks now display a time/ > >that is one hour later than the system time.e' > >Why is that?  How can I change that?e >n >t	 > Executeo >s( >     $ @SYS$MANAGER:UTC$TIME_SETUP SHOW >dC > If AUTO_DLIGHT_SAV is not set to "1", please consider setting it,a@ > so that OpenVMS will make the seasonal DAYLING<->STANDARD time changes.G > (Don't for get to add AUTO_DLIGHT_SAV to MODPARAMS.  For compatiblityiG > with older systems, the default is "0", so it will be reset during anm- > OpenVMS upgrade if it is not in MODPARAMS.)s >cE > If the rest of the information is incorrect, or maybe just for goode luck,c	 > Execute  >c# >     $ @SYS$MANAGER:UTC$TIME_SETUPo >t< > to reset the time variables.  (This is the replacement for utc$configure_tdft@ > and does quite a bit more -- some of which may not strictly be
 necessary,+ > but is resonably quick and does no harm.)  >{2 > (You may have to re-start your x-window clocks.) >rB > You might also want to look at SYS$EXAMPLES:DAYLIGHT_SAVING.COM, which F > is much changed from the OLD_DAYLIGHT_SAVING.COM that is still there > for reference. > D > The OpenVMS FAQ hs several topics related to Dayling Savings time.5 > See http://h71000.www7.hp.com/faq/openvms_faq.html.n >  > --G >       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL   USA> >           (hammond@not@peek.ssr.hp.com -- remove "@not" when	 replying)c@ >       All opinions expressed are my own and not necessarily my employer's.t  D I and dlighted to get a reply from the guy who's command procedure I was just reviewing!i   A couple of questions:  G 1.  The FAQ seems to imply that I would have to shutdown NTP during thei time change. WhyE is that?  I never had to do that before since the system time and thei tdf were changed at the same time.   D 2.  The system doc implies that I should use sys$set_system_event to look forG sysevt$c_tdf_change event to synchronize with the time change.  Is that  correct?G Is there an easy way to have a command procedure to run just before ther time change?G (I know I can do it the old way, just schedule the application-specifice stuff)   ------------------------------   End of INFO-VAX 2005.188 ************************