0 INFO-VAX	Fri, 05 Jan 2001	Volume 2001 : Issue 10      Contents:P Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable in LOGIN.COP Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable in LOGIN.COP Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable inLOGIN.COMP Re: APACHE (CSWS) + MOD_PERL = No private process logicals definablein LOGIN.COMP Re: APACHE (CSWS) + MOD_PERL = No private process logicals definablein LOGIN.COM  RE: AQUARIUM OWNERS READ THIS!!!  Re: AQUARIUM OWNERS READ THIS!!!, Re: Big Operator.log full of strage messages, Re: Big Operator.log full of strage messages! Re: Can DCL talk to TCP/IP ports? A Re: Can't get ghostscript 6.0 to create postscript file from pdf. A Re: Can't get ghostscript 6.0 to create postscript file from pdf. # con$convert problems on VAX-VMS 7.2 ' Re: con$convert problems on VAX-VMS 7.2 ' Re: con$convert problems on VAX-VMS 7.2  Re: CSWS Newbie Questions  Re: CSWS Newbie Questions F Re: DCL Temp Files Was: DCL: Retrieving file version from OPEN commandF Re: DCL Temp Files Was: DCL: Retrieving file version from OPEN command2 RE: DCL: Retrieving file version from OPEN command2 RE: DCL: Retrieving file version from OPEN command2 RE: DCL: Retrieving file version from OPEN command DTR install problem  Re: DTR install problem # RE: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # RE: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again # Re: Ebay dead as a post - yet again  Re: Happy New Years  Re: How can I setup ucx$ntpd?  Re: How can I setup ucx$ntpd?  Re: RMS for unix?  Re: RMS for unix?  Re: Source dor TSM021? Re: Source dor TSM021?+ Re: Suggestion for improvement for DECterm. + Re: Suggestion for improvement for DECterm. ! Used DS20E Special - One Only !!! D Using Telnet create_session (was: Re: Can DCL talk to TCP/IP ports?)H Re: Using Telnet create_session (was: Re: Can DCL talk to TCP/IP ports?)" uVAX 3100 model80/90 hardware info& Re: uVAX 3100 model80/90 hardware info Re: VAX to NT migration? Re: VAX to NT migration? Re: VAX to NT migration?& Re: voor de belgische Alpha gebruikers& Re: voor de belgische Alpha gebruikers ZIP and UNZIP for VAX-VMS 7.2 ! Re: ZIP and UNZIP for VAX-VMS 7.2 ! Re: ZIP and UNZIP for VAX-VMS 7.2 ! Re: ZIP and UNZIP for VAX-VMS 7.2   F ----------------------------------------------------------------------   Date: 5 Jan 2001 12:41:49 GMT - From: "Phil Tregoning" <ptregoni@esoc.esa.de> Y Subject: Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable in LOGIN.CO = Message-ID: <01c07712$d855eb40$4b53b083@ptregoni.esoc.esa.de>   3 Andy Stoffel <acs@fcgnetworks.net> wrote in article 0 <a5856.200416$DG3.4584400@news2.giganews.com>... > I > I thought maybe that there just wasn't enough room for all the process   > logicals but:  > < >     LNMPHASHTBL on this system is set to 512 (the default)< >    and, when mod_perl is disabled, an APACHE$WWW_# process >    has 244 process logicals.   >   1 This only affects the size of the hash table used 2 for process logicals - it doesn't limit the number3 you can have. Process logicals go into the P1 pool, , the size of which is controlled by CTLPAGES.  2 244 sounds like quit a lot of process logicals. It4 is quite likely CTLPAGES needs to be increased. This is an autogen parameter.   Phil T      ------------------------------  # Date: Fri, 05 Jan 2001 15:20:41 GMT * From: "Andy Stoffel" <acs@fcgnetworks.net>Y Subject: Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable in LOGIN.CO 9 Message-ID: <dDl56.167401$IP1.5978949@news1.giganews.com>   8 "Phil Tregoning" <ptregoni@esoc.esa.de> wrote in message7 news:01c07712$d855eb40$4b53b083@ptregoni.esoc.esa.de...   4 > 244 sounds like quit a lot of process logicals. It6 > is quite likely CTLPAGES needs to be increased. This > is an autogen parameter.  / Looking at the VMS 7.2-1 documentation it says:   D    CTLPAGES specifies the size of P1 pool. CTLPAGES is automaticallyD    changed only when the process logical name table, DCL symbols, orB    some layered products require an increase in the size of the P1
    pool area.   K This (to me anyway) implied that this is automatically adjusted at the same 8 time as it is listed as an AUTOGEN adjustable parameter.  J How do I determine WHAT to adjust it to (rule of thumb?) ? I couldn't find? any information on this in the VMS documentation I've searched.    Using H "Google" I've found several documents that say to adjust it for specific? products but not explanation of WHY they recommend those values @ other than one product (HIERARCHY?) that includes the CHANNELCNTD value as part of the calculation. But not WHY & just because SOMEONE8 uses that calculation doesn't mean it applies to Apache.  @ And what other system parameters need to be adjusted if CTLPAGES
 is adjusted ?   * Any thoughts would be appreciated. Thanks.   -Andy-   ------------------------------  % Date: Fri, 05 Jan 2001 11:51:49 -0500 ) From: Rick Barry <barry@star.zko.dec.com> Y Subject: Re: APACHE (CSWS) + MOD_PERL = No private process logicals definable inLOGIN.COM 0 Message-ID: <3A55FBA5.9125D8B7@star.zko.dec.com>  1 It does seem  like CTLPAGES might need adjusting.   N As a quick experiment, you might try doubling CTLPAGES. This requires a rebootN and affects all process system-wide. If that works, then adjusting it downward until 5 the problem reappears would give you the right value.   F Please let me know if this corrects the problem, so we can improve our
 documentation 
 in that area.    Thanks.    Andy Stoffel wrote:   : > "Phil Tregoning" <ptregoni@esoc.esa.de> wrote in message9 > news:01c07712$d855eb40$4b53b083@ptregoni.esoc.esa.de...  > 6 > > 244 sounds like quit a lot of process logicals. It8 > > is quite likely CTLPAGES needs to be increased. This > > is an autogen parameter. > 1 > Looking at the VMS 7.2-1 documentation it says:  > F >    CTLPAGES specifies the size of P1 pool. CTLPAGES is automaticallyF >    changed only when the process logical name table, DCL symbols, orD >    some layered products require an increase in the size of the P1 >    pool area.  > M > This (to me anyway) implied that this is automatically adjusted at the same : > time as it is listed as an AUTOGEN adjustable parameter. > L > How do I determine WHAT to adjust it to (rule of thumb?) ? I couldn't findA > any information on this in the VMS documentation I've searched.  >  > Using J > "Google" I've found several documents that say to adjust it for specificA > products but not explanation of WHY they recommend those values B > other than one product (HIERARCHY?) that includes the CHANNELCNTF > value as part of the calculation. But not WHY & just because SOMEONE: > uses that calculation doesn't mean it applies to Apache. > B > And what other system parameters need to be adjusted if CTLPAGES > is adjusted ?  > , > Any thoughts would be appreciated. Thanks. >  > -Andy-   --
 Rick Barry  3 Compaq Secure Web Server (CSWS) - Development Group  Compaq Computer Corporation 
 Nashua, NH   ------------------------------  % Date: Fri, 05 Jan 2001 10:33:58 -0500 " From: Dan Sugalski <dan@sidhe.org>Y Subject: Re: APACHE (CSWS) + MOD_PERL = No private process logicals definablein LOGIN.COM : Message-ID: <5.0.2.1.0.20010105102919.0216e120@24.8.96.48>  - At 11:56 PM 1/4/01 +0000, Andy Stoffel wrote: > >What appears to have happened is that, with mod_perl enabled,6 >the APACHE$WWW processes no longer seem to have their; >process-private logicals as created in Apache's LOGIN.COM. C >(Ouch!) and this "legacy" application stops working. But any other ; >logicals (job or system) or symbols ARE defined correctly.   J Yowtch. I can't think off-hand of a good reason for this, but a couple of 
 questions:  * 1) Are the process logicals defined /USER?- 2) Are you throwing any sort of quota errors? L 3) Is the process logical table readable by the user the process ultimately  ends up as?   L There shouldn't be any reason that mod_perl in and of itself will throw you K off here, but it does suck down a lot of extra quota (including subprocess  K quotas) at least temporarily, and spawns off some subproceses temporarily,  H so that may be where you're running into issues. Maybe. It also defines = some logicals in the process table, hence the quota question.      					Dan  I --------------------------------------"it's like this"------------------- 2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and even ;                                       teddy bears get drunk    ------------------------------  # Date: Fri, 05 Jan 2001 17:08:11 GMT * From: "Andy Stoffel" <acs@fcgnetworks.net>Y Subject: Re: APACHE (CSWS) + MOD_PERL = No private process logicals definablein LOGIN.COM 9 Message-ID: <%bn56.167533$IP1.5987682@news1.giganews.com>   B [Oops, hadn't realized Dan's response was posted also... so here's what I told him.... - Andy-]  / "Dan Sugalski" <dan@sidhe.org> wrote in message 4 news:5.0.2.1.0.20010105102919.0216e120@24.8.96.48...? # Yowtch. I can't think off-hand of a good reason for this, but  # a couple of  questions:  # , # 1) Are the process logicals defined /USER?   No.   / # 2) Are you throwing any sort of quota errors?    Not that I've seen.   : # 3) Is the process logical table readable by the user the! # process ultimately  ends up as?    Yes.  8 I haven't found much info on adjusting system parameters) to affect the process logical name table. < The only suggestion I've gotten so far is to adjust CTLPAGES7 but I can't just randomly reboot this particular system ; (It's in Iowa & I'm in NH) without any clear idea on how to : adjust it, especially on a Friday %-). And my local system: only vaguely resembles the system I'm having problems with/ (Comparing a DEC 2000 w/192MB memory to a 2 CPU 7 Alphaserver 1200 w/512MB memory, even  running the same = versions of VMS & other applications, is a large stretch %-))   8 Does mod_perl really create that many process logicals ?* This is my first encounter with it on VMS.   Thanks for your suggestions.    -Andy-    ------------------------------  % Date: Fri, 05 Jan 2001 12:04:18 +0000 8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>) Subject: RE: AQUARIUM OWNERS READ THIS!!! L Message-ID: <35666012DF4CD411BE940090279FA24011201C@ppnt41.physics.ox.ac.uk>  J This message is in MIME format. Since your mail reader does not understand< this format, some or all of this message may not be legible.  ' ------_=_NextPart_001_01C0770F.A14CE5C0  Content-Type: text/plain;  	charset="iso-8859-1"   K It's generally a bad idea to share the same water amongst a large number of J tanks as disease, or any other form of contamination would spread rapidly.G It's rather like having all your computing resources in one cluster and * attaching all disks to a single interface.    John     C Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.uk  I Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UK  B Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)     ' ------_=_NextPart_001_01C0770F.A14CE5C0  Content-Type: text/html; 	charset="iso-8859-1"   > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD>H <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">    5 <META content="MSHTML 5.00.2919.6307" name=GENERATOR>  <STYLE></STYLE>  </HEAD>  <BODY bgColor=#ffffff>N <DIV><FONT color=#0000ff face=6x13 size=2><SPAN class=533120112-05012001>It's P generally a bad idea to share the same water amongst a large number of tanks as N disease, or any other form of contamination would spread rapidly. It's rather P like having all your computing resources in one cluster and attaching all disks * to a single interface.</SPAN></FONT></DIV>0 <DIV><FONT color=#0000ff face=6x13 size=2><SPAN 3 class=533120112-05012001></SPAN></FONT>&nbsp;</DIV> 0 <DIV><FONT color=#0000ff face=6x13 size=2><SPAN 1 class=533120112-05012001>John</SPAN></FONT></DIV>  <DIV>&nbsp;</DIV> K <P><FONT color=#800000 face=Verdana size=1>Name: John B. Macallister&nbsp;  G E-mail: j.macallister1@physics.ox.ac.uk</FONT> <BR><FONT color=#800000  K face=Verdana size=1>Post: Nuclear and Astrophysics Laboratory, Keble Road,  L Oxford OX1 3RH,UK</FONT> <BR><FONT color=#800000 face=Verdana size=1>Phone: L +44-1865-273388 (direct)&nbsp; 273333 (reception)&nbsp; 273418 (Fax)</FONT>  </P></BODY></HTML>  ) ------_=_NextPart_001_01C0770F.A14CE5C0--    ------------------------------  # Date: Fri, 05 Jan 2001 15:53:12 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net>) Subject: Re: AQUARIUM OWNERS READ THIS!!! < Message-ID: <I5m56.33061$1M.7688606@typhoon.ne.mediaone.net>  E "John Macallister" <J.Macallister1@physics.ox.ac.uk> wrote in message F news:35666012DF4CD411BE940090279FA24011201C@ppnt41.physics.ox.ac.uk...K It's generally a bad idea to share the same water amongst a large number of J tanks as disease, or any other form of contamination would spread rapidly.G It's rather like having all your computing resources in one cluster and * attaching all disks to a single interface.  K At the risk of perpetuating this piscatorial diversion, John sure knows his  astronotos ocellatus!    ------------------------------  % Date: Fri, 05 Jan 2001 08:30:06 +0000 * From: Ed Dennison <ed.dennison@compaq.com>5 Subject: Re: Big Operator.log full of strage messages * Message-ID: <3A55860E.C3C211CB@compaq.com>  R It may be worth checking the CLUSTER_SERVER process on all the remote nodes in theQ cluster. Is it running with all privileges and under UIC [1,4]? These are the two N conditions under which the "clm privilege violation" message should be logged.K (This assumes that the incomming message is comming from the CLUSTER_SERVER 	 process).                        Ed Dennison    ------------------------------  # Date: Fri, 05 Jan 2001 14:29:26 GMT  From: rocoto@my-deja.com5 Subject: Re: Big Operator.log full of strage messages ) Message-ID: <934lnu$h7o$1@nnrp1.deja.com>   * In article <3A55860E.C3C211CB@compaq.com>,-   Ed Dennison <ed.dennison@compaq.com> wrote: G > It may be worth checking the CLUSTER_SERVER process on all the remote  nodes in theG > cluster. Is it running with all privileges and under UIC [1,4]? These  are the two H > conditions under which the "clm privilege violation" message should be logged. > > (This assumes that the incomming message is comming from the CLUSTER_SERVER > process).    Right.  E I've seen this on my 7.2 (update 1) systems also.... quite rare. It's E happened 3 times in a year or so, out of the blue - not at startup or C anything, it just starts for no discernable reason. Each time, I've > fixed the problem by using DELTA to correct the CLUSTER_SERVERH PSB$Q_WORKPRIV from FFFFFFFF.5FFFFFFF back to FFFFFFFF.FFFFFFFF which is! what OPCOM seems to be expecting.   C Dunno if the Q found & fixed this yet -- Does seem like it'd be the H dickens to find though... I do know that there are some good folks there! who were made aware of the issue.    Regards. -- David 	 rold5(at) 
 tditx(dot)com      Sent via Deja.com  http://www.deja.com/   ------------------------------  % Date: Fri, 05 Jan 2001 10:48:31 -0800 0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>* Subject: Re: Can DCL talk to TCP/IP ports?, Message-ID: <3A55A67F.7754F661@Mvb.Saic.Com>   Dean Woodward wrote (in part): >  > .  > .  > . P > > Another new feature for V5.1 will be that you can define the TCPIP$SMTP_FROMK > > logical name, specifying an alternate address for your "From:" field in L > > outbound mail.  This feature is specifically intended for users who wantL > > their mail to appear to be from their Microsoft Outlook/Exchange mailbox% > > address.  We hope you'll like it.	 > H > So DEFINE/JOB TCPIP$SMTP_FROM "deanw@rdrop.com <Dean Woodward>" shouldG > work?  Can I point out that some mailers expect to see brokets around=E > the friendly name field, but that VMSMAIL /PERSONAL_NAME won't take G > them, and uses parentheses?  It's a nit, but in the MS-centric office(B > environment I work in, Exchange won't recognize (Dean Woodward).  F That would result in an invalid address.  Parts inside parentheses areE comments to an email address, anything inside angle brackets ( the <>I> characters ) must be a vaild, machine-readable address.  Thus:  ( deanw@rdrop.com (Dean Woodward) is valid andk- Dean Woodward <deanw@rdrop.com> is also valid.  & but the example you give above is not.  
 Mark Berrymans Mark.Berryman@Mvb.Saic.Com   ------------------------------  $ Date: Fri, 5 Jan 2001 02:05:39 -0500* From: Marc Tardif <intmktg@Gloria.CAM.ORG>J Subject: Re: Can't get ghostscript 6.0 to create postscript file from pdf.E Message-ID: <Pine.LNX.4.10.10101050201400.3354-100000@Gloria.CAM.ORG>a  J In my reply to your original posting, I pasted the following line from the pdf2ps script:  U exec gs -q -dNOPAUSE -dBATCH -sDEVICE=pswrite -sOutputFile=$2 $OPTIONS -c save pop -$   H Perhaps my message was delayed, but in case you missed it, the answer to2 your latest question is to use the -dNOPAUSE flag.  ! On 5 Jan 2001, Mark London wrote:g  O > Hi -  That does create the file, but now I get the showpage prompt, no matterE5 > what I try.  Is there a way to avoid that?  Thanks.p > F > $ gs "-sDEVICE=pswrite" "-sOutputFile=abc.xyz" example.pdf "-dBATCH"& > Aladdin Ghostscript 6.0 (2000-02-03)O > Copyright (C) 2000 Aladdin Enterprises, Menlo Park, CA.  All rights reserved.eH > This software comes with NO WARRANTY: see the file PUBLIC for details. > Processing pages 1 through 1.s > Page 1* > >>showpage, press <return> to continue<< >  > I > In a previous article, Helge Blischke <H.Blischke@srz-berlin.de> wrote:y > ->Mark London wrote: > ->> S > ->> I'm trying to get ghostscript 6.0 to create a postscript file from pdf.  I doe > ->> the following: > ->> S > ->> $  gs "-dNODISPLAY"  "-sOutputFile=abc.xyz" example.pdf "-dBATCH" "-dNOPAUSE"t* > ->> Aladdin Ghostscript 6.0 (2000-02-03)S > ->> Copyright (C) 2000 Aladdin Enterprises, Menlo Park, CA.  All rights reserved.sL > ->> This software comes with NO WARRANTY: see the file PUBLIC for details.# > ->> Processing pages 1 through 1.X > ->> Page 1 > ->> = > ->> No abc.xyz is created.  What am I doing wrong?  Thanks.  > -> o4 > ->Replace the "-dNODISPLAY" by "-sDEVICE=pswrite". > -> o	 > ->Helges > -> h > ->-- e > ->H.Blischke@srz-berlin.de > ->H.Blischke@srz-berlin.com  > ->H.Blischke@acm.org > 
 > Mark London  > MRL@PSFC.MIT.EDU >  >    ------------------------------   Date: 5 JAN 2001 16:19:07 GMT-$ From: mrl@psfc.mit.edu (Mark London)J Subject: Re: Can't get ghostscript 6.0 to create postscript file from pdf.* Message-ID: <5JAN01.16190722@psfc.mit.edu>  G Thanks. I actually did try dNOPAUSE, but I had put it at the end of thesM command line rather than the beginning.  I won't even ask why it doesn't work@ that way.  e  B In a previous article, Marc Tardif <intmktg@Gloria.CAM.ORG> wrote:L ->In my reply to your original posting, I pasted the following line from the ->pdf2ps script: ->  W ->exec gs -q -dNOPAUSE -dBATCH -sDEVICE=pswrite -sOutputFile=$2 $OPTIONS -c save pop -$s -> bJ ->Perhaps my message was delayed, but in case you missed it, the answer to4 ->your latest question is to use the -dNOPAUSE flag. -> o# ->On 5 Jan 2001, Mark London wrote:  -> aQ ->> Hi -  That does create the file, but now I get the showpage prompt, no matterC7 ->> what I try.  Is there a way to avoid that?  Thanks.s ->> H ->> $ gs "-sDEVICE=pswrite" "-sOutputFile=abc.xyz" example.pdf "-dBATCH"( ->> Aladdin Ghostscript 6.0 (2000-02-03)Q ->> Copyright (C) 2000 Aladdin Enterprises, Menlo Park, CA.  All rights reserved.tJ ->> This software comes with NO WARRANTY: see the file PUBLIC for details.! ->> Processing pages 1 through 1.-
 ->> Page 1, ->> >>showpage, press <return> to continue<< ->>  ->> K ->> In a previous article, Helge Blischke <H.Blischke@srz-berlin.de> wrote:e ->> ->Mark London wrote: ->> ->> U ->> ->> I'm trying to get ghostscript 6.0 to create a postscript file from pdf.  I dos ->> ->> the following: ->> ->> U ->> ->> $  gs "-dNODISPLAY"  "-sOutputFile=abc.xyz" example.pdf "-dBATCH" "-dNOPAUSE"e, ->> ->> Aladdin Ghostscript 6.0 (2000-02-03)U ->> ->> Copyright (C) 2000 Aladdin Enterprises, Menlo Park, CA.  All rights reserved.sN ->> ->> This software comes with NO WARRANTY: see the file PUBLIC for details.% ->> ->> Processing pages 1 through 1.j ->> ->> Page 1 ->> ->> ? ->> ->> No abc.xyz is created.  What am I doing wrong?  Thanks.t ->> -> A6 ->> ->Replace the "-dNODISPLAY" by "-sDEVICE=pswrite". ->> -> t ->> ->Helgen ->> -> d	 ->> ->-- n ->> ->H.Blischke@srz-berlin.de ->> ->H.Blischke@srz-berlin.com  ->> ->H.Blischke@acm.org ->>  ->> Mark Londonn ->> MRL@PSFC.MIT.EDU ->>  ->>  -> o   Mark Londone MRL@PSFC.MIT.EDU   ------------------------------  $ Date: Fri, 5 Jan 2001 15:14:49 +0100+ From: "Peter Kille" <peter.kille@volvo.com>t, Subject: con$convert problems on VAX-VMS 7.2, Message-ID: <934ktk$884$1@vg170.it.volvo.se>  2 I upgraded a VAX system from 7.1 to 7.2, a programL converting files (been running since -90) suddenly crashed. It crashes doing CONV$CONVERT, im getting this message in the logfile:  ( Retur kod CONV$PASS_OPTIONS:           1# Retur kod CONV$CONVERT:    11829388e> %DEBUGBOOT-W-CHN, assign channel system service request failed! %FDL-F-OPENFDL, error opening !ASe  6 I can see in SDA> that im running out of channels when9 looking at the process. It converts 338 files and crashes " every time, (channelcount is 400).  / Anybody else had any problems with CONVERT /7.2d   Peter Killer   ------------------------------  # Date: Fri, 05 Jan 2001 15:13:53 GMT 1 From: "Mark D. Jilson" <jilly@clarityconnect.com>o0 Subject: Re: con$convert problems on VAX-VMS 7.22 Message-ID: <3A55E5BF.8211F37A@clarityconnect.com>  A WAG increase CHANNELCNT to 1000 vua MODPARAMS.DAT and AUTOGEN andeH reboot.  Also make sure the processes FILLM is going to be at least 850.   Peter Kille wrote: > 4 > I upgraded a VAX system from 7.1 to 7.2, a programN > converting files (been running since -90) suddenly crashed. It crashes doing > CONV$CONVERT, im getting > this message in the logfile: > * > Retur kod CONV$PASS_OPTIONS:           1% > Retur kod CONV$CONVERT:    11829388e@ > %DEBUGBOOT-W-CHN, assign channel system service request failed# > %FDL-F-OPENFDL, error opening !ASl > 8 > I can see in SDA> that im running out of channels when; > looking at the process. It converts 338 files and crashes $ > every time, (channelcount is 400). > 1 > Anybody else had any problems with CONVERT /7.2t > 
 > Peter Killem   -- nD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Fri, 05 Jan 2001 11:14:37 -0500H2 From: norm lastovica <norman.lastovica@oracle.com>0 Subject: Re: con$convert problems on VAX-VMS 7.2* Message-ID: <3A55F2ED.32B87FB7@oracle.com>  G you should get and install all available patches for 7.2 (and I believekD that there are a few).  if the problem persists, I think that you'llF find a "OLD" version of the sort and convert images in sys$share.  Try these.   Peter Kille wrote: > 4 > I upgraded a VAX system from 7.1 to 7.2, a programN > converting files (been running since -90) suddenly crashed. It crashes doing > CONV$CONVERT, im getting > this message in the logfile: > * > Retur kod CONV$PASS_OPTIONS:           1% > Retur kod CONV$CONVERT:    11829388e@ > %DEBUGBOOT-W-CHN, assign channel system service request failed# > %FDL-F-OPENFDL, error opening !ASt > 8 > I can see in SDA> that im running out of channels when; > looking at the process. It converts 338 files and crashesu$ > every time, (channelcount is 400). > 1 > Anybody else had any problems with CONVERT /7.2  > 
 > Peter Kille    -- a> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Fri, 05 Jan 2001 11:00:23 -0500)) From: Rick Barry <barry@star.zko.dec.com>h" Subject: Re: CSWS Newbie Questions0 Message-ID: <3A55EF97.4FFA2885@star.zko.dec.com>  
 Problem 1:  ? CSWS uses C run-time library time functions to obtain the locale? time. These functions rely on a  time differential factor (TDF) ? from UTC to correctly interpret the local  time value accordinge: to UTC time. If the system time is correct, but the TDF is5 not set correctly, the  local time will be1 hour off.n   To set the TDF, use either  $ $ @sys$examples:daylight_savings.com   or  $ $ @sys$manager:utc$configure_tdf.com  D These procedures will set the logical name SYS$TIMEZONE_DIFFERENTIAL< and the internal TDF cell EXE$GQ_TDF (which the C RTL uses).  
 Problem 4:  9 The SSL_SCACHE.DIR and SSL_SCACHE.PAG files are temporary K files used to cache SSL session data. These files exist while the server isRG running and are deleted when the server is shut down. The error messageIC you are seeing is due to a bug in CSWS V1.0 which causes some child F processes to re-create the session cache files. This adversely affectsI performance, but the server will continue to run correctly. (This will be - fixed in CSWS V1.0-1 due out "any time now".)6   "Barratt, Chris (FMC)" wrote:f  L > Last week I installed Compaq Secure Web Server on my workstation ("macca")  > to have a bit of a look at it.J > At this stage I have not installed the mod_perl or mod_jserv components. > I am running TCP/IP Services.  >iN > After reading the documentation over and over, and searching a couple of theN > Apache sites on the web, I have some questions which perhaps someone who has2 > been there done that might be able to help with. > M > 1.  The time being displayed in the log files is one hour different to thats > from a $SHOW TIME at DCL.nF >      Where does CSWS get it's time from ?    (I am expecting this is. > something to do with an incorrect NTP setup) >tN > 2.  I have the option "Indexes" on the document root, but whenever I specifyN > it in my browser (ie.   <http://macca/> http://macca/), it does not find theK > index.html file in the directory. If the index.html file is not there, it K > does not do the usual listing of the directory contents. I have specifiedhG > index.html (and all case variations) in the DirectoryIndex directive. M > Additionally, if I create a directory within my document root directory, it1# > also exhibits the same behaviour.R > J >     Error log shows a "File does not exist" message for the directory in > question.o >cL >     However, if I goto URL   http://macca/icons/ <http://macca/icons/>   IH > do get the listing of directory contents, and URL http://macca/manual/F > <http://macca/manual/>  will return the appropriate index.html file. >"N >     When I access http://macca/index <http://macca/index>   it works, but ifJ > I access http://macca/index.html <http://macca/index.html>  it does not.* > (which leads me into the next question). >cM > 3.  Using http, if I specify the HTML or HTM extension for a file, I get anc( > error, but it works if I leave it off.N >      Using https, it works only if I supply the extension and not if I leave	 > it off.K >eB > 4.  On startup of the server, I am getting an error "Cannot openI > SSLSessionCache DBM file '/apache$root/logs/ssl_scache' for scanning ",o< > followed by "System: no such file or directory (errno:2)". >PI >     This does not appear to be affecting operations, but seems strange.U >"M >     It seems to create a file in the [logs] area called SSL_SCACHE.DIR withMA > zero blocks. When the server is shutdown, this file is deleted.D >g >oG > Can anybody help with what I have setup wrong, or explain the quirks.1 >5	 > Thanks,  > Chris Barratt  >a   --
 Rick Barry  3 Compaq Secure Web Server (CSWS) - Development Group  Compaq Computer Corporationo
 Nashua, NH   ------------------------------   Date: 5 Jan 2001 16:18:56 GMTi2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)" Subject: Re: CSWS Newbie Questions6 Message-ID: <934s5g$cgp$1@mailint03.im.hou.compaq.com>   In article <07103702F27FD411ACA30000F808545257C2E5@sagemshs001.fmc.sa.gov.au>, "Barratt, Chris (FMC)" <Chris.Barratt@fmc.sa.gov.au> writes:  ..L :1.  The time being displayed in the log files is one hour different to that :from a $SHOW TIME at DCL.E :     Where does CSWS get it's time from ?    (I am expecting this isr- :something to do with an incorrect NTP setup)r ..  J   Ayup.  Please see the OpenVMS FAQ for details on how to set your system I   time, and (particularly in this case) how to set daylight savings time.7  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Fri, 5 Jan 2001 09:02:07 +0100> From: "Jean-Franois Marchal" <jean-francois.marchal@x9000.fr>O Subject: Re: DCL Temp Files Was: DCL: Retrieving file version from OPEN commande. Message-ID: <933usi$ph5$1@reader1.imaginet.fr>  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in messagee- news:87pui24ypa.fsf_-_@k9.prep.synonet.com...r6 > hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes: >w >d$ > How about a wish list item for DCL > . >              $open/write/temporary/delete...    ! I would prefer $ close/delete ...f   Jean-Franois Marchalr X9000 - LYON (FR)a   ------------------------------   Date: 5 Jan 2001 15:48:32 GMT-2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)O Subject: Re: DCL Temp Files Was: DCL: Retrieving file version from OPEN commandO, Message-ID: <934qcg$keu@gap.cco.caltech.edu>  b In article <87pui24ypa.fsf_-_@k9.prep.synonet.com>, Paul Repacholi <prep@prep.synonet.com> writes: >t8 >Open an file for tempory use, deleting it when the .com0 >exits. Like a open tmp and mark for delete QIO. >uC Well ok, but also fix the behavior with SORT regarding these files,TI at least under 7.2-1 Alpha.  If you exceed diskquota while sorting a file H SORT bombs with the expected diskquota message, but leaves the temporaryE files in place!  It may actually be callable SORT that does this, themF program I see it with invokes a SORT and I'm not sure if it's by SPAWNE or through the library calls.  Anyway, this has the exceedingly nastymH effect of both failing the present operation and screwing up the account@ until such time as the "temporary" files are removed - manually.     David Mathog mathog@seqaxp.bio.caltech.eduu? Manager, sequence analysis facility, biology division, Caltech e   ------------------------------  % Date: Fri, 05 Jan 2001 08:21:29 -0500t4 From: "Bochnik, William J" <BochnikWJ@bernstein.com>; Subject: RE: DCL: Retrieving file version from OPEN commandaJ Message-ID: <2B37459189B0D211BE710000F8EF9D8508908AE0@nts0147.beehive.com>  J This message is in MIME format. Since your mail reader does not understand< this format, some or all of this message may not be legible.  ' ------_=_NextPart_001_01C0771A.65ED0AA4R Content-Type: text/plain;a 	charset="iso-8859-1"   ) >In article <932nni$9fe$1@acme.gcfn.org>, + >  kennsmit@gcfn.org (Kenneth Smith) wrote:e >>( >> I am looking for a method to delete a+ >> file that I am creating and closing withs( >> OPEN and CLOSE.  I have the file name* >> from the filespec name, but do not have* >> the version number.  I would prefer not, >> to use a wildcard for the version number. >>* >> Feel free to e-mail with any questions.
 >> Thanks, >> ks. >> kennsmit @ gcfn.org >> >> Sample Code:t >>, >> $       OPEN /WRITE OUTPUT_FILE ERROR.TXT4 >> $       WRITE OUTPUT_FILE "An error has occurred" >> $       CLOSE OUTPUT_FILE, >> $       ... do something with OUTPUT_FILE >> $	... delete OUTPUT_FILE  >> > - >You've hard-coded the file name, so just use" >m >$ DELETE ERROR.TXT; >d= >to delete the file. When you use the semicolon by itself, ittH >automatically refers to the most current version, which in this case isF >the one you just created via OPEN/WRITE. There is no need to know theF >actual version number. Normally, you don't even need the ";" to referB >to the most current version, but with the DELETE command you mustD >explicitly specify the version number, and using ";" is sufficient. >zF >Whether this works for your full code depends, of course, on what "do  >something with OUTPUT_FILE" is.    I If there is a chance that more than one of these can be running at a timeeE (implied by his not wanting to use wildcards) it'll break.  He shouldtI consider creating a "random" filename out of the pid, date/time, and some I fixed part - and check that the file doesnt exist.  That should guaranteewL that he won't have collisions.  He can then do the wildcard delete (or the ; delete)m   $ PARSED_NAME = F$FAO("!%T",0)2 $ PARSED_NAME = "''PARSED_NAME'" - ":" - ":" - "." $ PROC_ID = F$GETJPI("","PID")= $ U_NAME = "sys$login:AA_TEMP''PARSED_NAME'"+"''PROC_ID'.log"     ' ------_=_NextPart_001_01C0771A.65ED0AA4A Content-Type: text/html; 	charset="iso-8859-1" + Content-Transfer-Encoding: quoted-printable   1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">s <HTML> <HEAD>9 <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =  charset=3Diso-8859-1">@ <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
 5.5.2651.65">aA <TITLE>RE: DCL: Retrieving file version from OPEN command</TITLE>  </HEAD>C <BODY>  " <P><FONT SIZE=3D2>&gt;In article =* &lt;932nni$9fe$1@acme.gcfn.org&gt;,</FONT>A <BR><FONT SIZE=3D2>&gt;&nbsp; kennsmit@gcfn.org (Kenneth Smith) =h
 wrote:</FONT>l" <BR><FONT SIZE=3D2>&gt;&gt;</FONT>A <BR><FONT SIZE=3D2>&gt;&gt; I am looking for a method to delete =T a</FONT>A <BR><FONT SIZE=3D2>&gt;&gt; file that I am creating and closing =a with</FONT>lC <BR><FONT SIZE=3D2>&gt;&gt; OPEN and CLOSE.&nbsp; I have the file =a name</FONT>o@ <BR><FONT SIZE=3D2>&gt;&gt; from the filespec name, but do not = have</FONT>iF <BR><FONT SIZE=3D2>&gt;&gt; the version number.&nbsp; I would prefer =
 not</FONT>? <BR><FONT SIZE=3D2>&gt;&gt; to use a wildcard for the version =c number.</FONT>" <BR><FONT SIZE=3D2>&gt;&gt;</FONT>: <BR><FONT SIZE=3D2>&gt;&gt; Feel free to e-mail with any = questions.</FONT>v* <BR><FONT SIZE=3D2>&gt;&gt; Thanks,</FONT>% <BR><FONT SIZE=3D2>&gt;&gt; ks</FONT> 6 <BR><FONT SIZE=3D2>&gt;&gt; kennsmit @ gcfn.org</FONT>" <BR><FONT SIZE=3D2>&gt;&gt;</FONT>/ <BR><FONT SIZE=3D2>&gt;&gt; Sample Code:</FONT>e" <BR><FONT SIZE=3D2>&gt;&gt;</FONT>H <BR><FONT SIZE=3D2>&gt;&gt; $&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; OPEN =# /WRITE OUTPUT_FILE ERROR.TXT</FONT> I <BR><FONT SIZE=3D2>&gt;&gt; $&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WRITE = 4 OUTPUT_FILE &quot;An error has occurred&quot;</FONT>I <BR><FONT SIZE=3D2>&gt;&gt; $&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; CLOSE =a OUTPUT_FILE</FONT>G <BR><FONT SIZE=3D2>&gt;&gt; $&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ... =p$ do something with OUTPUT_FILE</FONT>< <BR><FONT SIZE=3D2>&gt;&gt; $&nbsp;&nbsp;&nbsp; ... delete = OUTPUT_FILE</FONT>" <BR><FONT SIZE=3D2>&gt;&gt;</FONT> <BR><FONT SIZE=3D2>&gt;</FONT>A <BR><FONT SIZE=3D2>&gt;You've hard-coded the file name, so just =i
 use</FONT> <BR><FONT SIZE=3D2>&gt;</FONT>1 <BR><FONT SIZE=3D2>&gt;$ DELETE ERROR.TXT;</FONT>i <BR><FONT SIZE=3D2>&gt;</FONT>G <BR><FONT SIZE=3D2>&gt;to delete the file. When you use the semicolon =I by itself, it</FONT>A <BR><FONT SIZE=3D2>&gt;automatically refers to the most current = % version, which in this case is</FONT>SG <BR><FONT SIZE=3D2>&gt;the one you just created via OPEN/WRITE. There =  is no need to know the</FONT>lH <BR><FONT SIZE=3D2>&gt;actual version number. Normally, you don't even =& need the &quot;;&quot; to refer</FONT>I <BR><FONT SIZE=3D2>&gt;to the most current version, but with the DELETE =M command you must</FONT> I <BR><FONT SIZE=3D2>&gt;explicitly specify the version number, and using = # &quot;;&quot; is sufficient.</FONT>h <BR><FONT SIZE=3D2>&gt;</FONT>G <BR><FONT SIZE=3D2>&gt;Whether this works for your full code depends, =e" of course, on what &quot;do</FONT>B <BR><FONT SIZE=3D2>&gt;something with OUTPUT_FILE&quot; is.</FONT> </P> <BR>  H <P><FONT SIZE=3D2>If there is a chance that more than one of these can =D be running at a time (implied by his not wanting to use wildcards) =E it'll break.&nbsp; He should consider creating a &quot;random&quot; =0E filename out of the pid, date/time, and some fixed part - and check = G that the file doesnt exist.&nbsp; That should guarantee that he won't =dE have collisions.&nbsp; He can then do the wildcard delete (or the ; =  delete)</FONT></P>  C <P><FONT SIZE=3D2>$ PARSED_NAME =3D F$FAO(&quot;!%T&quot;,0)</FONT>rC <BR><FONT SIZE=3D2>$ PARSED_NAME =3D &quot;''PARSED_NAME'&quot; - = 4 &quot;:&quot; - &quot;:&quot; - &quot;.&quot;</FONT>" <BR><FONT SIZE=3D2>$ PROC_ID =3D =- F$GETJPI(&quot;&quot;,&quot;PID&quot;)</FONT>i! <BR><FONT SIZE=3D2>$ U_NAME =3D = I &quot;sys$login:AA_TEMP''PARSED_NAME'&quot;+&quot;''PROC_ID'.log&quot;</=  FONT>- </P>   </BODY>- </HTML> ) ------_=_NextPart_001_01C0771A.65ED0AA4--    ------------------------------  # Date: Fri, 05 Jan 2001 14:45:17 GMTL* From: Alan E. Feldman <alan48@my-deja.com>; Subject: RE: DCL: Retrieving file version from OPEN commandn) Message-ID: <934mls$hvf$1@nnrp1.deja.com>e  
 In article? <2B37459189B0D211BE710000F8EF9D8508908AE0@nts0147.beehive.com>,f7   "Bochnik, William J" <BochnikWJ@bernstein.com> wrote: A > This message is in MIME format. Since your mail reader does note
 understand> > this format, some or all of this message may not be legible. >a) > ------_=_NextPart_001_01C0771A.65ED0AA4- > Content-Type: text/plain;a > 	charset="iso-8859-1"m >a+ > >In article <932nni$9fe$1@acme.gcfn.org>,u- > >  kennsmit@gcfn.org (Kenneth Smith) wrote:f > >>* > >> I am looking for a method to delete a- > >> file that I am creating and closing with2* > >> OPEN and CLOSE.  I have the file name, > >> from the filespec name, but do not have, > >> the version number.  I would prefer not. > >> to use a wildcard for the version number. > >>, > >> Feel free to e-mail with any questions. > >> Thanks, > >> ksf > >> kennsmit @ gcfn.org > >> > >> Sample Code:d > >>. > >> $       OPEN /WRITE OUTPUT_FILE ERROR.TXT6 > >> $       WRITE OUTPUT_FILE "An error has occurred" > >> $       CLOSE OUTPUT_FILE. > >> $       ... do something with OUTPUT_FILE > >> $	... delete OUTPUT_FILEm > >> > >f/ > >You've hard-coded the file name, so just usey > >t > >$ DELETE ERROR.TXT; > >m? > >to delete the file. When you use the semicolon by itself, iteG > >automatically refers to the most current version, which in this case. isH > >the one you just created via OPEN/WRITE. There is no need to know theH > >actual version number. Normally, you don't even need the ";" to referD > >to the most current version, but with the DELETE command you mustF > >explicitly specify the version number, and using ";" is sufficient. > > H > >Whether this works for your full code depends, of course, on what "do" > >something with OUTPUT_FILE" is. >oF > If there is a chance that more than one of these can be running at a timeG > (implied by his not wanting to use wildcards) it'll break.  He shouldA  G Well, it would help if the orignial poster were a little more specific.nF I knew it. I should have answered: "Could you be a little more vague?"  G Furthermore, if he had several of these running, his "do something withdB OUTPUT_FILE" would also break because all his concurrent processesA would attempt to access the most recent version of ERROR.TXT, andrD possible crash, regardless of which process made that ERROR.TXT, and= therefore his procedure would *already* be broken, or at besti unreliable.   F Other interpretations are possible. Perhaps he runs it several times aG day, and only during some of those he wants to delete it, in which caseU my answer is fine.  F > consider creating a "random" filename out of the pid, date/time, and someA > fixed part - and check that the file doesnt exist.  That shouldd	 guaranteeeD > that he won't have collisions.  He can then do the wildcard delete	 (or the ;e	 > delete)n >e  > $ PARSED_NAME = F$FAO("!%T",0)4 > $ PARSED_NAME = "''PARSED_NAME'" - ":" - ":" - "."  > $ PROC_ID = F$GETJPI("","PID")? > $ U_NAME = "sys$login:AA_TEMP''PARSED_NAME'"+"''PROC_ID'.log"m >e) > ------_=_NextPart_001_01C0771A.65ED0AA4e > Content-Type: text/html; > 	charset="iso-8859-1"t- > Content-Transfer-Encoding: quoted-printablei >b3 > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">, > <HTML> > <HEAD>  " Please turn off MIME mode. Thanks.  " > <META HTTP-EQUIV=3D"Content-Type   [...tons of mime stuff deleted]n   --F NOTE: If you wish to e-mail me, please do NOT use the deja address. ItE is broken. Instead, use one of the addresses below, removing the longd wrong part first. Thanks.W   Disclaimer: JMHO Alan E. Feldman  &-)+ w: afeldman@gfigroup.ButItSaidItPrinted.coma5 h: alan48@dellnet.YouCantBelieveEverythingYouRead.com-     Sent via Deja.coma http://www.deja.com/   ------------------------------   Date: 5 Jan 2001 13:00:00 -0500:' From: kennsmit@gcfn.org (Kenneth Smith)s; Subject: RE: DCL: Retrieving file version from OPEN command0( Message-ID: <935230$1u3$1@acme.gcfn.org>   Alan:u  % My apologies.  Yes, it was too vague.g, This is what happens when you get frustrated$ and post something without thinking.  ) Let me backtrack and propose the originalp( design problem, with the key information I originally left out:  ' Create an alert message ON ERROR from aU% batch procedure.  Transfer this alertw$ to another (non-VMS) server via FTP,! where it is processed and logged.r  ( Currently I am hitting the posted sample# script in an error handling section ) of the code (i.e. ON ERROR GOTO error_1)."% The intent of the sample script is tos. write the output to a text file, then kick-off( another procedure (FTP.COM) that handles* the FTP process and then deletes the file.' The file name is being passed via P1 tol& FTP.COM.  This would be implemented in$ multiple batch procedures, so having( one procedure handling the FTP minimizes" code and allows a central point of administration for FTP changes.o  r& The problem I ran into was getting the$ version number for the file that the% OPEN /WRITE statement is creating, son% it could be deleted from the original" procedure or even FTP.COM.  & As it has been pointed out, creating a- unique string for the file name and including & the semicolon in DELETE command is one& alternative.  As Alan indicated, it is( possible that multiple versions could be( created before the deletion takes place.% By including the PID, date and time I># suspect this will be unlikely here.t  % Unless another solution is presented,>$ possibly a different solution to the# design problem all together, I planh to pursue the path above.s  $ Again, my apoligies to all for being$ so vague.  Hopefully this clears the air.   Thanks,@ ks kennsmit @ gcfn.org-  + > >In article <932nni$9fe$1@acme.gcfn.org>,1- > >  kennsmit@gcfn.org (Kenneth Smith) wrote:n > >>* > >> I am looking for a method to delete a- > >> file that I am creating and closing withA* > >> OPEN and CLOSE.  I have the file name, > >> from the filespec name, but do not have, > >> the version number.  I would prefer not. > >> to use a wildcard for the version number. > >>, > >> Feel free to e-mail with any questions. > >> Thanks, > >> ks  > >> kennsmit @ gcfn.org > >> > >> Sample Code:a > >>- > >> $      OPEN /WRITE OUTPUT_FILE ERROR.TXT>5 > >> $      WRITE OUTPUT_FILE "An error has occurred"u > >> $      CLOSE OUTPUT_FILEp- > >> $      ... do something with OUTPUT_FILEu > >> $ ... delete OUTPUT_FILEs > >> > >o/ > >You've hard-coded the file name, so just use  > >  > >$ DELETE ERROR.TXT; > >w? > >to delete the file. When you use the semicolon by itself, ittG > >automatically refers to the most current version, which in this caset isH > >the one you just created via OPEN/WRITE. There is no need to know theH > >actual version number. Normally, you don't even need the ";" to referD > >to the most current version, but with the DELETE command you mustF > >explicitly specify the version number, and using ";" is sufficient. > >>H > >Whether this works for your full code depends, of course, on what "do" > >something with OUTPUT_FILE" is. >pF > If there is a chance that more than one of these can be running at a timeG > (implied by his not wanting to use wildcards) it'll break.  He shoulde   I Well, it would help if the orignial poster were a little more specific. I>E knew it. I should have answered: "Could you be a little more vague?" N  UG Furthermore, if he had several of these running, his "do something with H OUTPUT_FILE" would also break because all his concurrent processes wouldD attempt to access the most recent version of ERROR.TXT, and possibleI crash, regardless of which process made that ERROR.TXT, and therefore his / procedure would *already* be broken, or at besth unreliable.>   F Other interpretations are possible. Perhaps he runs it several times aJ day, and only during some of those he wants to delete it, in which case my answer is fine.   >F > consider creating a "random" filename out of the pid, date/time, and someA > fixed part - and check that the file doesnt exist.  That should-	 guarantee D > that he won't have collisions.  He can then do the wildcard delete	 (or the ;t	 > delete)c >c  > $ PARSED_NAME = F$FAO("!%T",0)4 > $ PARSED_NAME = "''PARSED_NAME'" - ":" - ":" - "."  > $ PROC_ID = F$GETJPI("","PID")? > $ U_NAME = "sys$login:AA_TEMP''PARSED_NAME'"+"''PROC_ID'.log"h > ) > ------_=_NextPart_001_01C0771A.65ED0AA4  > Content-Type: text/html; >  charset="iso-8859-1"P- > Content-Transfer-Encoding: quoted-printable1 >93 > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">  > <HTML> > <HEAD>  O" Please turn off MIME mode. Thanks.   " > <META HTTP-EQUIV=3D"Content-Type    [...tons of mime stuff deleted]c  e --F NOTE: If you wish to e-mail me, please do NOT use the deja address. ItE is broken. Instead, use one of the addresses below, removing the long- wrong part first. Thanks.r   Disclaimer: JMHO Alan E. Feldman  &-)+ w: afeldman@gfigroup.ButItSaidItPrinted.comc5 h: alan48@dellnet.YouCantBelieveEverythingYouRead.com    ------------------------------  $ Date: Fri, 5 Jan 2001 10:53:54 -0600+ From: "Cliff Squidlow" <csquid@hotmail.com>A Subject: DTR install problem+ Message-ID: <E_m56.3$kU.3296@nnrp2.sbc.net>l  L I am trying to install Datatrieve on a Microvax 3100/10e (all latest versionK software) and get a LINK error (can't find VAXC$CRTL_INIT) which aborts theNL install.  Any ideas?  I played with the options and also got a LINK error onI a reference to FORMS$something.  I'm guessing something didn't make it toCL IMAGELIB or STARLET, but didn't see anything in the release notes or iguide.  G Also, I see that CDD isn't part of the package anymore, something aboutoL Oracle taking it over - do you need it / can you get it ?  I'd only be usingC RMS files, and haven't done the Wombat thing for quite a while now..   ------------------------------   Date: 5 Jan 2001 18:26:25 GMTe2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)  Subject: Re: DTR install problem6 Message-ID: <9353kh$dof$1@mailint03.im.hou.compaq.com>  Y In article <E_m56.3$kU.3296@nnrp2.sbc.net>, "Cliff Squidlow" <csquid@hotmail.com> writes:1M :I am trying to install Datatrieve on a Microvax 3100/10e (all latest versionRL :software) and get a LINK error (can't find VAXC$CRTL_INIT) which aborts the	 :install.   I   Um, your definition of "latest" might differ from mine or from those ofeK   another site -- the explicit specification of the version of OpenVMS VAX rK   and of Datatrieve would be useful.  As of this writing, OpenVMS VAX V7.2  I   and Datatrieve V7.2A would be good choices, but of course SDK2/EFT2 of i6   OpenVMS VAX V7.3 (G7.3) is also currently available.  $   What is the exact LINK error seen?  4   Are any LNK* logical names defined on this system?  ?   Which C compilers (and which versions) are installed, if any?   F   I'll also assume that the MicroVAX 3100 model 10e either has a smallF   system disk, or it has the ROM upgrade to permit correct operations G   with larger system disks -- please see the OpenVMS FAQ for details...s  C :Any ideas?  I played with the options and also got a LINK error ondJ :a reference to FORMS$something.  I'm guessing something didn't make it toM :IMAGELIB or STARLET, but didn't see anything in the release notes or iguide.e     FDV$?  Is FMS installed?  5   FORM$ would imply DECforms.  Is DECforms installed?o   (Try reinstallation?)p   H :Also, I see that CDD isn't part of the package anymore, something aboutM :Oracle taking it over - do you need it / can you get it ?  I'd only be usingiD :RMS files, and haven't done the Wombat thing for quite a while now.  B   Datatrieve circa V7.2 can operate with Oracle CDD/Repository, orE   without it.  Earlier versions can operate only with CDD/Repository.h  F   You might also want to contact the customer support center, as this <   might take a little digging to figure out what happened...  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Fri, 05 Jan 2001 11:15:17 +0000-8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>, Subject: RE: Ebay dead as a post - yet againL Message-ID: <35666012DF4CD411BE940090279FA24011201A@ppnt41.physics.ox.ac.uk>  K After reading the description of the eBay problem on their website it would.L not have surprised me if a VMS cluster had behaved in a similar fashion. TheL initial problem was a hardware failure on one node of the cluster. I've beenL running clusters for years and have experienced similar scenarios to the oneI described. A hardware problem anywhere in a cluster can often lead to the E whole cluster becoming unusable through processes hanging and devices>H becoming inaccessible. Bringing another node (the "third backup system")9 into the messed-up cluster simply compounded the problem.e  L While it's sometimes possible to localise a problem in a cluster, especiallyE if detected at an early stage, problems often rapidly get out of handbF leading to what can only be described as a mess. The best option in myF experience is usually to bring down the whole cluster, fix/isolate the problem and reboot.N  C The problem with the eBay setup may have been partly caused by some I deficiency in the SUN system but the major fault at eBay appears to me to/L have been caused by some lack of experience on the part of those running theK system at the time. If they had shut down the whole system when the problemFG first appeared the outage time may have been minutes rather than hours.b   John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)t   ------------------------------  % Date: Fri, 05 Jan 2001 12:09:20 +0000B8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>, Subject: RE: Ebay dead as a post - yet againL Message-ID: <35666012DF4CD411BE940090279FA24011201D@ppnt41.physics.ox.ac.uk>  G > VMS clusters can be configured to have no single-point-of-failure.=20>  G The Queue manager database file has to reside on a disk and when that =F diskE is faulty the cluster fairly quickly becomes unusable especially whenX> there's a large amount of  e-mail traffic using the queueing = mechanisms.t  	 This is a    John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukC Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 =/ 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)r     -----Original Message-----3 From: Arne Vajh=F8j [mailto:arne.vajhoej@gtech.com]o' Sent: Friday, January 05, 2001 11:52 AM; To: John Macallister, Subject: Re: Ebay dead as a post - yet again     John Macallister wrote:MG > After reading the description of the eBay problem on their website it  would C > not have surprised me if a VMS cluster had behaved in a similar =  fashion. TheFF > initial problem was a hardware failure on one node of the cluster. = I've beenH > running clusters for years and have experienced similar scenarios to = theL onetI > described. A hardware problem anywhere in a cluster can often lead to =e the G > whole cluster becoming unusable through processes hanging and devicessC > becoming inaccessible. Bringing another node (the "third backup =t system"); > into the messed-up cluster simply compounded the problem.r >=20C > While it's sometimes possible to localise a problem in a cluster,h
 especiallyG > if detected at an early stage, problems often rapidly get out of hand G > leading to what can only be described as a mess. The best option in =% myF > experience is usually to bring down the whole cluster, fix/isolate = thet > problem and reboot.  >=20E > The problem with the eBay setup may have been partly caused by someEG > deficiency in the SUN system but the major fault at eBay appears to =3 me toqD > have been caused by some lack of experience on the part of those = running- the>E > system at the time. If they had shut down the whole system when the- problem-D > first appeared the outage time may have been minutes rather than = hours.  ' As I see it, then there are two issues:   ' 1)  How often does Ebay have problems ?   D     It is my impression that Ebay has problems more often than otherA     internet sites and more often than what we in the VMS world =l consider     normal.t  0     Yes - I think Ebay could do better with VMS.  ( 2)  How well does Ebay handle problems ?  A     If the backup and the backup for the backup can not take over E     because of the same failure that hit the first system, then thereaG     are a design problem. There are a single-point-of-failure, so there>*     are not full redundancy in the system.  @     It could have been designed much better. A VMS cluster couldD     be configured to do a much better job. Even SUN systems could be"     configured to do a better job.  A     And I do not understand your remark about VMS clusters. VMS =  clustersE     do not gurantee against single-point-of-failure. VMS clusters are  usedF     for many other purposes than redundancy. I would even think that = the A     majority of clusters exist for easy sharing of user database,$ sharingIF     of files, sharing of printer-queues and shareing of system disk. = VMS$F     clusters can be configured to have no single-point-of-failure. A = VMS>C     cluster not configured to have no single-point-of-failure couldDB     exhibit similar problems to Ebay. Single-point-of-failure is aE     generic problem not specific for any OS or redundancy technology.o  G     Knowing that they do not have full redundancy, then your suggestionrB     about rebooting everything instead of trying to let the backup     take over also makes sense.    Arne   ------------------------------  % Date: Fri, 05 Jan 2001 12:14:20 +0000p8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>, Subject: RE: Ebay dead as a post - yet againL Message-ID: <35666012DF4CD411BE940090279FA24011201E@ppnt41.physics.ox.ac.uk>  E > VMS clusters can be configured to have no single-point-of-failure.    J The Queue manager database file has to reside on a disk and when that diskE is faulty the cluster fairly quickly becomes unusable especially when H there's a large amount of  e-mail traffic using the queueing mechanisms.  K This is a single point of failure in a  VMS CLUSTER which is hard to avoid.F   John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)r   ------------------------------  % Date: Fri, 05 Jan 2001 12:52:18 +0100i= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>o, Subject: Re: Ebay dead as a post - yet again) Message-ID: <3A55B572.45286BB5@gtech.com>o   John Macallister wrote:fM > After reading the description of the eBay problem on their website it would N > not have surprised me if a VMS cluster had behaved in a similar fashion. TheN > initial problem was a hardware failure on one node of the cluster. I've beenN > running clusters for years and have experienced similar scenarios to the oneK > described. A hardware problem anywhere in a cluster can often lead to theRG > whole cluster becoming unusable through processes hanging and devicesyJ > becoming inaccessible. Bringing another node (the "third backup system"); > into the messed-up cluster simply compounded the problem.. > N > While it's sometimes possible to localise a problem in a cluster, especiallyG > if detected at an early stage, problems often rapidly get out of handyH > leading to what can only be described as a mess. The best option in myH > experience is usually to bring down the whole cluster, fix/isolate the > problem and reboot.m > E > The problem with the eBay setup may have been partly caused by someaK > deficiency in the SUN system but the major fault at eBay appears to me tooN > have been caused by some lack of experience on the part of those running theM > system at the time. If they had shut down the whole system when the problemrI > first appeared the outage time may have been minutes rather than hours..  ' As I see it, then there are two issues:s  ' 1)  How often does Ebay have problems ?   D     It is my impression that Ebay has problems more often than otherH     internet sites and more often than what we in the VMS world consider     normal.   0     Yes - I think Ebay could do better with VMS.  ( 2)  How well does Ebay handle problems ?  A     If the backup and the backup for the backup can not take over E     because of the same failure that hit the first system, then thereeG     are a design problem. There are a single-point-of-failure, so therep*     are not full redundancy in the system.  @     It could have been designed much better. A VMS cluster couldD     be configured to do a much better job. Even SUN systems could be"     configured to do a better job.  H     And I do not understand your remark about VMS clusters. VMS clustersE     do not gurantee against single-point-of-failure. VMS clusters aref usedH     for many other purposes than redundancy. I would even think that theA     majority of clusters exist for easy sharing of user database,d sharingeH     of files, sharing of printer-queues and shareing of system disk. VMSH     clusters can be configured to have no single-point-of-failure. A VMSC     cluster not configured to have no single-point-of-failure couldsB     exhibit similar problems to Ebay. Single-point-of-failure is aE     generic problem not specific for any OS or redundancy technology.w  G     Knowing that they do not have full redundancy, then your suggestionaB     about rebooting everything instead of trying to let the backup     take over also makes sense.,   Arne   ------------------------------   Date: 5 Jan 2001 08:10:35 -0500 9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)$, Subject: RE: Ebay dead as a post - yet again+ Message-ID: <igVpYz3Y$3hN@eisner.decus.org>k   In article <35666012DF4CD411BE940090279FA24011201E@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> writes:eF >> VMS clusters can be configured to have no single-point-of-failure.  > L > The Queue manager database file has to reside on a disk and when that diskG > is faulty the cluster fairly quickly becomes unusable especially when>J > there's a large amount of  e-mail traffic using the queueing mechanisms. > M > This is a single point of failure in a  VMS CLUSTER which is hard to avoid.L  G By "disk", I presume you mean a hardware failure rather than a softwareTG failure.  Host-based volume shadowing should defend against such a disk  problem.  G Of course if it is the Queue Manager _software_ that fails, VMS driversiH will happily let it scribble false data all over the file, but I presume% that is not what you meant by "disk".a  N ==============================================================================N Great Inventors of our time: Al Gore -> Internet; Sun Microsystems -> ClustersN ==============================================================================   ------------------------------  % Date: Fri, 05 Jan 2001 14:22:21 +0100n( From: Etienne Vogt <vogt@alice.obspm.fr>, Subject: Re: Ebay dead as a post - yet again. Message-ID: <3A55CA8D.1CBD2035@alice.obspm.fr>   John Macallister wrote: L > The Queue manager database file has to reside on a disk and when that diskG > is faulty the cluster fairly quickly becomes unusable especially whenlJ > there's a large amount of  e-mail traffic using the queueing mechanisms.M > This is a single point of failure in a  VMS CLUSTER which is hard to avoid.s   Volume shadowing ?   -- r# 		Etienne Vogt (vogt @ na.astro.it)s   ------------------------------  % Date: Fri, 05 Jan 2001 08:15:43 -0500w0 From: David Beatty <David.Beatty.NOSPAM@sas.com>, Subject: Re: Ebay dead as a post - yet again2 Message-ID: <Z8hVOmAkzMKSYKpB8q4qkJQxFr5j@4ax.com>  @ Do you mean faulty as in hardware failure?  How about a mirrored disk?t  < You can also have multiple queue manager databases, although, I've never had the need to use this feature.  ? You can get around the Queue Manager database; it just requiresh a little work.   David R. Beatty   4 On Fri, 05 Jan 2001 12:14:20 +0000, John Macallister( <J.Macallister1@physics.ox.ac.uk> wrote:  F >> VMS clusters can be configured to have no single-point-of-failure.  >EK >The Queue manager database file has to reside on a disk and when that diskiF >is faulty the cluster fairly quickly becomes unusable especially whenI >there's a large amount of  e-mail traffic using the queueing mechanisms.  >tL >This is a single point of failure in a  VMS CLUSTER which is hard to avoid. >D >John. >EC >Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.uk I >Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKeB >Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax) >s >t   ------------------------------  % Date: Fri, 05 Jan 2001 08:34:43 -0500r4 From: "Bochnik, William J" <BochnikWJ@bernstein.com>, Subject: RE: Ebay dead as a post - yet againJ Message-ID: <2B37459189B0D211BE710000F8EF9D8508908AE1@nts0147.beehive.com>   -----Original Message----- >From: John Macallister L >After reading the description of the eBay problem on their website it wouldI >not have surprised me if a VMS cluster had behaved in a similar fashion.s TheDH >initial problem was a hardware failure on one node of the cluster. I've beenI >running clusters for years and have experienced similar scenarios to the  onedJ >described. A hardware problem anywhere in a cluster can often lead to theF >whole cluster becoming unusable through processes hanging and devicesI >becoming inaccessible. Bringing another node (the "third backup system")D: >into the messed-up cluster simply compounded the problem.  J The clusters that you're familar with - were they shared local storage, or" using some sort of hsc/hsz?  With J properly configured hsc configurations, we never ran into any problem like this.  Shadowed disks prevented@K a single disk failure from bringing down the system, dual hsc's prevented ag( failure of them from an outage, and the L 3 system cluster allowed us to bring down any 1 system w/o harming the other 2 (aside from 10-30 seconds ofH a cluster state transition).  We very rarely needed to reboot the entireI cluster (at least when it was VMS; DUnix and NT are different stories) to  clear a problem.   >rB >While it's sometimes possible to localise a problem in a cluster,
 especiallyF >if detected at an early stage, problems often rapidly get out of handG >leading to what can only be described as a mess. The best option in mynG >experience is usually to bring down the whole cluster, fix/isolate the  >problem and reboot. >sD >The problem with the eBay setup may have been partly caused by someJ >deficiency in the SUN system but the major fault at eBay appears to me toI >have been caused by some lack of experience on the part of those running  theVL >system at the time. If they had shut down the whole system when the problemH >first appeared the outage time may have been minutes rather than hours.  L I agree that a deficienty in the SUN system is to blame, but I think that it' is the major contributor, as well as a OL lack of understanding on the part of the people who set the system up.  Just= rebooting will get the system back up (usually) quickly, but  I 1 - if the problem is not machine related, but database related this willa not helpL 2 - rebooting can destroy any footprint of the problem, thus the problem can come back again if not fixed      3 These opinions are of course mine, not my employer.n   William J. Bochnik   ------------------------------  % Date: Fri, 05 Jan 2001 08:36:52 -0500e4 From: "Bochnik, William J" <BochnikWJ@bernstein.com>, Subject: RE: Ebay dead as a post - yet againJ Message-ID: <2B37459189B0D211BE710000F8EF9D8508908AE2@nts0147.beehive.com>   -----Original Message----- From: John Macallister    F >> VMS clusters can be configured to have no single-point-of-failure.  >oK >The Queue manager database file has to reside on a disk and when that diskeF >is faulty the cluster fairly quickly becomes unusable especially whenI >there's a large amount of  e-mail traffic using the queueing mechanisms.e    * use shadowed disks or mirrored raid disks.   ------------------------------  % Date: Fri, 05 Jan 2001 08:39:39 -0500t2 From: norm lastovica <norman.lastovica@oracle.com>, Subject: Re: Ebay dead as a post - yet again* Message-ID: <3A55CE9B.7B6D7EAB@oracle.com>  B I believe that technically this isn't a true statement.  there canC be multiple queue managers in the cluster (one per node).  as well,aC you'd probably be using shadowing to protect from disk & controllerrB (and perhaps site) failure.  I'd also imagine that for many shops,B the inability to print reports doesn't amount to the cluster being 'unusable'.i   John Macallister wrote:( > F > > VMS clusters can be configured to have no single-point-of-failure. > L > The Queue manager database file has to reside on a disk and when that diskG > is faulty the cluster fairly quickly becomes unusable especially whenfJ > there's a large amount of  e-mail traffic using the queueing mechanisms. >  > This is a  >  > John > D > Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukJ > Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKC > Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)e >  > -----Original Message-----3 > From: Arne Vajhj [mailto:arne.vajhoej@gtech.com] ) > Sent: Friday, January 05, 2001 11:52 AMm > To: John Macallister. > Subject: Re: Ebay dead as a post - yet again >  > John Macallister wrote: I > > After reading the description of the eBay problem on their website itu > would L > > not have surprised me if a VMS cluster had behaved in a similar fashion. > ThetK > > initial problem was a hardware failure on one node of the cluster. I'vee > beenL > > running clusters for years and have experienced similar scenarios to the > one M > > described. A hardware problem anywhere in a cluster can often lead to the0I > > whole cluster becoming unusable through processes hanging and devicesrL > > becoming inaccessible. Bringing another node (the "third backup system")= > > into the messed-up cluster simply compounded the problem.l > > E > > While it's sometimes possible to localise a problem in a cluster,t > especiallyI > > if detected at an early stage, problems often rapidly get out of handnJ > > leading to what can only be described as a mess. The best option in myJ > > experience is usually to bring down the whole cluster, fix/isolate the > > problem and reboot.r > >iG > > The problem with the eBay setup may have been partly caused by somemM > > deficiency in the SUN system but the major fault at eBay appears to me tooL > > have been caused by some lack of experience on the part of those running > theoG > > system at the time. If they had shut down the whole system when thee	 > problem K > > first appeared the outage time may have been minutes rather than hours.o > ) > As I see it, then there are two issues:u > ) > 1)  How often does Ebay have problems ?  > F >     It is my impression that Ebay has problems more often than otherJ >     internet sites and more often than what we in the VMS world consider
 >     normal.  > 2 >     Yes - I think Ebay could do better with VMS. > * > 2)  How well does Ebay handle problems ? > C >     If the backup and the backup for the backup can not take overtG >     because of the same failure that hit the first system, then there I >     are a design problem. There are a single-point-of-failure, so there , >     are not full redundancy in the system. > B >     It could have been designed much better. A VMS cluster couldF >     be configured to do a much better job. Even SUN systems could be$ >     configured to do a better job. > J >     And I do not understand your remark about VMS clusters. VMS clustersG >     do not gurantee against single-point-of-failure. VMS clusters aree > usedJ >     for many other purposes than redundancy. I would even think that theC >     majority of clusters exist for easy sharing of user database,w	 > sharingrJ >     of files, sharing of printer-queues and shareing of system disk. VMSJ >     clusters can be configured to have no single-point-of-failure. A VMSE >     cluster not configured to have no single-point-of-failure couldmD >     exhibit similar problems to Ebay. Single-point-of-failure is aG >     generic problem not specific for any OS or redundancy technology.d > I >     Knowing that they do not have full redundancy, then your suggestionbD >     about rebooting everything instead of trying to let the backup! >     take over also makes sense.t >  > Arne   -- a> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------   Date: 5 Jan 2001 08:54:55 -0500d* From: young_r@eisner.decus.org (Rob Young), Subject: RE: Ebay dead as a post - yet again+ Message-ID: <2YyQuFmofzl5@eisner.decus.org>r   In article <35666012DF4CD411BE940090279FA24011201D@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> writes:pH >> VMS clusters can be configured to have no single-point-of-failure.=20 > I > The Queue manager database file has to reside on a disk and when that =r > diskG > is faulty the cluster fairly quickly becomes unusable especially whene@ > there's a large amount of  e-mail traffic using the queueing =
 > mechanisms.  >   9 	You can set anything up for failure.  Putting QMAN files : 	on a single disk is one way.  My preferred way of setting= 	up crummy clusters includes that technique.  But when I flipbD 	open my "Preferred Way To Set Up High Availability Clusters" manual: 	I see I want my QMAN files on mirrored disks, preferrably< 	mirrored and THEN shadowed just in case a controller hangs., 	Disks are cheaper and cheaper all the time.   				Rob-   ------------------------------  % Date: Fri, 05 Jan 2001 15:36:23 +0000o8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>, Subject: RE: Ebay dead as a post - yet againL Message-ID: <35666012DF4CD411BE940090279FA24011201F@ppnt41.physics.ox.ac.uk>  H While it's possible to have multiple queue managers, place QMAN files onE multiple disks and to have mirroring/shadowing with or without raid af@ particular queue has to use a file on some disk. The disk may beL mirrored/shadowed to provide some protection against disk failure but if theK hardware or software providing the mirroring/shadowing fails the queue will 
 misbehave.  F While VMS was the main OS here we never had the financial resources toK implement shadowing or any form of disk mirroring. Now that disks are cheap1E we do have raid on our critical systems but they're now WNT or UNIX. -   John  -B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)M   ------------------------------  # Date: Fri, 05 Jan 2001 15:49:08 GMTn4 From: "Terry C. Shannon" <terryshannon@mediaone.net>, Subject: Re: Ebay dead as a post - yet again< Message-ID: <U1m56.33059$1M.7688398@typhoon.ne.mediaone.net>  E "John Macallister" <J.Macallister1@physics.ox.ac.uk> wrote in message F news:35666012DF4CD411BE940090279FA24011201A@ppnt41.physics.ox.ac.uk...   >lE > The problem with the eBay setup may have been partly caused by some K > deficiency in the SUN system but the major fault at eBay appears to me toeJ > have been caused by some lack of experience on the part of those running thepE > system at the time. If they had shut down the whole system when thea problem I > first appeared the outage time may have been minutes rather than hours.  >h  K Operator error or inexperience makes sense. After all, there are a *lot* of K Sun systems out there (Sun shipped ~750 StarFires in 3Q00 alone) and if therJ problem was truly chronic and pandemic I am sure we'd have heard about it, NDA or no NDA.   ------------------------------  % Date: Fri, 05 Jan 2001 15:10:55 +0000e- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>e, Subject: Re: Ebay dead as a post - yet again) Message-ID: <3A55E3FF.DF193E1D@bbc.co.uk>l   John Macallister wrote:    >s >pE > The problem with the eBay setup may have been partly caused by someIK > deficiency in the SUN system but the major fault at eBay appears to me to N > have been caused by some lack of experience on the part of those running theM > system at the time. If they had shut down the whole system when the problemlI > first appeared the outage time may have been minutes rather than hours.  >y >n  L Interesting analysis John. I wonder if the Ebay IT people have an test/develP system similar if not identical  to the live one that they can make and break to theireJ hearts content to become familiar with recovering from problem situations.N Or do they just panic when it goes down and attempt to learn while their users wait.t  P Isn't that what ones backup datacentre should be used for? Do they have a backup  ( data centre? It doesn't appear so to me.   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukn  A I speak for myself only and my views in no way represent those ofr MedAS or the BBC.e   ------------------------------  % Date: Fri, 05 Jan 2001 15:14:33 +0000g- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>g, Subject: Re: Ebay dead as a post - yet again) Message-ID: <3A55E4D9.754D6618@bbc.co.uk>g   John Macallister wrote:   F > > VMS clusters can be configured to have no single-point-of-failure. >iL > The Queue manager database file has to reside on a disk and when that diskG > is faulty the cluster fairly quickly becomes unusable especially whenoJ > there's a large amount of  e-mail traffic using the queueing mechanisms. >s  J So, use a host-based-shadowed disk for cluster common files. Ok, it costs,: but you will be resilient to single disk or node failures.  4 Now, if each new disk came with a VOLSHAD licence...    --D6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uke  A I speak for myself only and my views in no way represent those ofe MedAS or the BBC.    ------------------------------  % Date: Fri, 05 Jan 2001 17:34:43 +0100t2 From: "Thomas H. Pauli" <thomaspauli@arcormail.de>, Subject: Re: Ebay dead as a post - yet again+ Message-ID: <3A55F7A3.4070008@arcormail.de>   E Well, in our 24-7-365-environment we even shadow our page/swap disks oF (host based) and got a complete hangup of our four database instances E running on four cluster nodes due to a loose SCSI cable at which the uF page/swap disks of one node were attached. This node slowed extremely D down, but not enough to be declared 'dead' and be thrown out of the G cluster. Our ORACLE database simply blocked all transactions, in a way =H sensing (presumably by the lock manager) that there is something wrong. E Simply pressing the OFF button on this node put life again into this -B cluster and the database did her work again as planned. We had to B install an additional SCSI host controller to eliminate that SOP,  overlooked in the design phase.rI This is the worst example of a VMS cluster failure in more than 14 years yI of VMS practice. There were more failures which but cannot be attributed rI to the clusters itself. Mostly it's been bad software or bad procedures. nJ If a cluster installation was designed properly you can rely on it! We do.   Thomas  ? ---------------------------------------------------------------a9 *** Please notice! New eMail address from 12-dec-2000 on:E thomaspauli@arcormail.de? ---------------------------------------------------------------t9 Thomas H. Pauli, Hammersteinstr.19, 14199 Berlin, Germany0   ------------------------------  % Date: Fri, 05 Jan 2001 12:52:46 -0500c2 From: norm lastovica <norman.lastovica@oracle.com>, Subject: Re: Ebay dead as a post - yet again* Message-ID: <3A5609EE.744DCB1E@oracle.com>  ? I would argue that this is stand is true for any number of datarA files (not simply the queue file).  The SYSUAF file (for example) < exhibits the same exposure.  As does a system paging file.    B Personally, I've been marginally involved with a number of systemsA using cluster technologies to reduce downtime/dataloss risk.  The:C "best" of these (in terms of highest likelyhood of risk avoidance) uA shadowed all modification transaction events between two seperateaA clusters.  Each cluster was multiple nodes (VAX 8650 if I recall)nB with multiple network interfaces, multiple CIs with multiple HSCs D with wide-spread shadowing (also called mirroring or raid1 by some).  @ However, in practicle terms, most people (as you indicate) don't@ bother to spend the dollars on hardware redundancy to avoid even the simplest of failures.    John Macallister wrote:o > J > While it's possible to have multiple queue managers, place QMAN files onG > multiple disks and to have mirroring/shadowing with or without raid a:B > particular queue has to use a file on some disk. The disk may beN > mirrored/shadowed to provide some protection against disk failure but if theM > hardware or software providing the mirroring/shadowing fails the queue willJ > misbehave. > H > While VMS was the main OS here we never had the financial resources toM > implement shadowing or any form of disk mirroring. Now that disks are cheap F > we do have raid on our critical systems but they're now WNT or UNIX. >  > John > D > Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukJ > Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKC > Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)a   -- c> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Fri, 05 Jan 2001 10:45:26 -0800 ! From: Shane.F.Smith@Healthnet.comw, Subject: Re: Ebay dead as a post - yet againD Message-ID: <OFA51EBC79.D720D443-ON882569CB.00646827@foundation.com>  I I'd say we don't know enough yet. Let's wait for an official explainationg/ to analyse  before we start blaming the E10K's.a   Shaner          F Curtis Rempel <vmsguy.no.spam.here@home.com> on 01/04/2001 06:18:40 PM  > Please respond to Curtis Rempel <vmsguy.no.spam.here@home.com>   To:   Info-VAX@Mvb.Saic.Com  cc:i  - Subject:  Re: Ebay dead as a post - yet againe      9 "Bob Koehler" <koehler@eisner.decus.org> wrote in messaget% news:WAhuFHFoKpCc@eisner.decus.org...iJ > In article <3A546D87.A3970877@gtech.com>, Arne =?iso-8859-1?Q?Vajh=F8j?=  <arne.vajhoej@gtech.com> writes: > > Anders wrote:yG > >> Well it has been dead for some time now, I wonder who will get thee blameo > >> this time.e > >s@ > > Maybe their machine-romm are too hot. Or have to much static > > electricity.F > > Or the humidity is wrong. Or the power floctates too much. Or .... > >c >cJ > It was the guy selling the antique watch.  Turns out it had radium paintI > on the hands.  Cache memory chips started imagining the radiation hits.s > H > ----------------------------------------------------------------------A > Bob Koehler                     | Computer Sciences Corporation.? > NASA GSFC Flight Software       | Federal Sector, Civil GroupIG >                                 | please remove ".aspm" when replyinga  C Andrew is missing his cue ... must still be fitting his mask on ...   1 <background music swells to "Send in the Clowns">    ------------------------------  % Date: Fri, 05 Jan 2001 15:03:23 +0000 % From: Alan Greig <agreig@my-deja.com>  Subject: Re: Happy New Years8 Message-ID: <v9ab5tk0m3fkoi8f55je9v44lsla94grsh@4ax.com>  . On Thu, 04 Jan 2001 20:39:57 +0000, John Laird* <john@laird-towers.freeserve.co.uk> wrote:  D >On Wed, 03 Jan 2001 13:43:25 +0000, Alan Greig <agreig@my-deja.com> >wrote:< >h >>And we still have E >>to explain that exact 9/5 + 32 conversion factor whcih follows fromD% >>the 180 degrees liquid water range.h >.I >It may not have been exact, but was redefined to be exact at some point.t > / >The inch became 2.54cm in just such a fashion.u  F Yep, that's quite possible. Although 180 is a lot nicer a number to be close to by chance than 2.54  @ Btw (as far as I know) legally an inch in the UK is not 2.54cm -A that's a US definition although for most practical purposes it is F close enough. In reality anyone still using inches in the UK is likelyC to be calibrated against the 2.54cm per inch standard though unless<> using a very old measure.  CNC controllers operating in inches) certainly use the 2.54 conversion factor.o     >	John   --
 Alan Greig   ------------------------------   Date: 5 Jan 2001 09:16:13 GMTu) From: leslie@clio.rice.edu (Jerry Leslie)a& Subject: Re: How can I setup ucx$ntpd?' Message-ID: <9343ct$71t$1@joe.rice.edu>a  + Chin-Yung (chinyung@mail2000.com.tw) wrote:tF : I can startup ntpd without any error, but I do not know how to setupK : UCX$NTP_TZ ( My server is in taiwan.) and configure my vms server to havecI : time synchronization with the other time server in internet? Is there ae# : ntp.conf file, like unix, in vms?s :n  L The instructions on how to create the SYS$SPECIFIC:[TCPIP$NTP]TCPIP$NTP.CONF file are at:  1   http://www.openvms.compaq.com/doc/72final/6526/u.   6526profile_contents_002.html#toc_chapter_10  6 The URL aove is split into two lines for this posting.  4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------   Date: 5 Jan 2001 09:35:04 GMTe) From: leslie@clio.rice.edu (Jerry Leslie) & Subject: Re: How can I setup ucx$ntpd?' Message-ID: <9344g8$71t$2@joe.rice.edu>   * Jerry Leslie (leslie@clio.rice.edu) wrote:- : Chin-Yung (chinyung@mail2000.com.tw) wrote:dH : : I can startup ntpd without any error, but I do not know how to setupM : : UCX$NTP_TZ ( My server is in taiwan.) and configure my vms server to haveeK : : time synchronization with the other time server in internet? Is there a % : : ntp.conf file, like unix, in vms?o : :   C My reply was for TCP/IP Services For OpenVMS 5.0 or newer. You needhC to tell which version of UCX you're running, using the DCL command:r       $ UCX SHOW VERSION   You create the file:  %    SYS$SPECIFIC:[UCX$NTP]UCX$NTP.CONFa   from:g  )    SYS$SPECIFIC:[UCX$NTP]UCX$NTP.TEMPLATEg  " specifying your NTP sources; e.g.:      peer 192.6.38.127    peer 128.249.1.10    peer 165.91.52.110p  @ NTP_TZ is for VMS 5.X and older systems. At one time, there was  a DSNLink article:  ;   [DEC TCP/IP] How to Setup and Use NTP Servers and Clientso   That described NTP for UCX.h    4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------  % Date: Fri, 05 Jan 2001 12:59:29 +0100l= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>  Subject: Re: RMS for unix?) Message-ID: <3A55B721.52F1C709@gtech.com>e   "David L. Nicol" wrote:bP > Is there a rms package which will work on top of the unix filesystem interface > (seek, read, write) anywhere?r > I > I am porting a vms/rms application to portable unix code and if I couldwR > simply link the thing against a local rms library I think that might make things, > easier than rewriting it all in terms like > : >    lseek(datafile, getpositionfromname(name), SEEK_SET);2 >    read(datafile, &datastructure, sizeof(datum)) > D > I know that this question violates the spirit of RMS being tied toC > disk geometry and page size and so forth as tight as the wires inbF > a twisted pair cable, responses reinforming me of the superiority ofD > VMS over UNIX w/o providing directions towards valid (incomplete?)J > libraries will be printed out and lateralyl inserted into the bird cage.  5 Certain companies do deliver "VMS packages" for Unix.i  1 Try and check what the following companies offer:c    http://www.accelr8.com/    http://www.bosbc.com/    http://www.sector7.com/   Arne  * PS: And shame on you for moving from VMS !   ------------------------------   Date: 5 Jan 2001 16:01:51 GMT 2 From: mathog@seqaxp.bio.caltech.edu (David Mathog) Subject: Re: RMS for unix?, Message-ID: <934r5f$keu@gap.cco.caltech.edu>  i In article <3A55B721.52F1C709@gtech.com>, Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> writes:  > + >PS: And shame on you for moving from VMS !f   Don't blame the messenger.    J Shame on Digital/Compaq for doing such a crappy job of marketing the VMS. 3 And Tru64.  And Linux/Alpha.  And the Alpha.  Sigh.o  S I predict that either AMD or Intel will own most of the 64 bit market in 3 years.  mH It's just plain stupid of Compaq to let that happen, but they will.  Or  rather, they are.    Regards,   David Mathog mathog@seqaxp.bio.caltech.edu ? Manager, sequence analysis facility, biology division, Caltech t   ------------------------------  # Date: Fri, 05 Jan 2001 14:32:00 GMTp1 From: "Mark D. Jilson" <jilly@clarityconnect.com>  Subject: Re: Source dor TSM021?d2 Message-ID: <3A55DBEE.60B4DC53@clarityconnect.com>  F Well we have found a whole class of customers who will not install andH UNZIP software on their systems that demanded a way to get patch kits inE a compressed format.  This would all be moot if the power that be had1G simply used an extension of .EXE :*(  I have no way to make this changelC and it has been reported numerous times to the powers that be.  I'm G hoping that someday an UNZIP utility will become a standard part of thefF VMS distribution, living in the SYS$ETC directory possibly but I'm not holding my breath.   Paul Repacholi wrote:l > , > Terry Kennedy <terry@gate.tmk.com> writes: > M > >   No, my analysis was correct - the server is screwed up. Here's the log:e > >t" > > prompt> telnet www.dnpg.com 80 > > Trying 63.105.58.26... > > Connected to www.dnpg.com. > > Escape character is '^]'. 7 > > GET /dnpg/dr/servers/tsm/TSM021.DCX_AXPEXE HTTP/1.0l > >t > > HTTP/1.1 200 OKI' > > Date: Thu, 04 Jan 2001 20:04:47 GMTkJ > > Server: Apache/1.3.6 (Unix) mod_perl/1.21 mod_ssl/2.2.8 OpenSSL/0.9.2b0 > > Last-Modified: Wed, 16 Jun 1999 20:55:05 GMT" > > ETag: "1ed803-47911e-37680f29" > > Accept-Ranges: bytes > > Content-Length: 4690206  > > Connection: closea > > Content-Type: text/plain > >:* > >   Note the "Content-type: text/plain". > >  > 0 > Terry is correct. Lynx also downloads in text. > < > BTW, you must get the file to suit the machine you want to > decompress on... > < > Why the F they didn't go to zip files, and get MUCH better+ > compression. And, shock horror, standard!  >  > Nights who say NIH.  >  > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda.fB >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.   -- nD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Fri, 05 Jan 2001 15:20:44 +0000n% From: Alan Greig <agreig@my-deja.com>c Subject: Re: Source dor TSM021?-8 Message-ID: <vdpb5ts0ho5hvtp25jbop74iqd6neg3vhq@4ax.com>  < On 4 Jan 2001 12:22:04 PST, Fairfield@SLAC.Stanford.EDU (Ken4 Fairfield; SLAC: 650-926-2924; FAX: 926-3515) wrote:   >In article <sbWDyF14Tzl0@mccdev.slac.stanford.edu>, Fairfield@SLAC.Stanford.EDU (Ken Fairfield; SLAC: 650-926-2924; FAX: 926-3515) writes:1 >[..]9K >>         If anyone out there  has  copies  of  the TSM021 savesets or DCX K >>     files,  at  least  for  Alpha VMS if not both,  and  can  make  themt< >>     available, I'd be most appreciative and in your debt! > I >        Thanks to everyone who  responded,  both  here and privately.  I 7 >    now have a "good" copy of the TSM021 distribution.r >eI >        Just to be  sure,  I  tried  downloading  from  DNPG  again thissI >    morning  and  even  after several attempts, none  of  the  downloadsaI >    completed successfully.  It may be some network issue between us and)I >    them.  Therefore, it may not be so unlikely that  there  was  actualaI >    corruption  during my download attempts yesterday.  Furthermore, the,I >    file attributes being stream-lf was  a  red herring since I was ableeI >    to download the DCX kit from another kind soul this morning, it cameEF >    in with stream-lf records, but it ran and decompressed correctly.  F Co-incidentally someone here downloaded TSM021 from the DNPG site just@ a few days ago. He had no problems after fixing the record size.    I >        One other point that was  unclear  from  the DNPG web site.  ThemI >    TSM021.A  installation saveset installs on _both_ VAX and Alpha,  soiI >    you  only   need   to   download   one   of   TSM021.DCX_AXPEXE   or I >    TSM021.DCX_VAXEXE,   not  both.   Choose  which  ever  one  is  moret >    convenient for you. > 1 >        Again, thanks to everyone for your help!r >  >            -Kenc   --
 Alan Greig   ------------------------------  % Date: Fri, 05 Jan 2001 12:39:40 -0500n2 From: rdeininger@mindspring.com (Robert Deininger)4 Subject: Re: Suggestion for improvement for DECterm.L Message-ID: <rdeininger-0501011239400001@user-2iveak3.dialup.mindspring.com>  ^ In article <9324tk$r18$1@mailint03.im.hou.compaq.com>, hammond@not@peek.ppb.cpqcorp.net wrote:  G > The following kludge is undoubtedly incomplete and imperfect, but it iF > is in my LOGIN.COM and seems to work.  Offered for what it is worth.( > [actually modified a bit for clarity.] >  > ..$ > $ IF F$GETDVI("SYS$COMMAND","TRM") > $ THENC > $     DEVNAM = F$GETDVI("SYS$COMMAND","TT_PHYDEVNAM") - "_" - "_"a& > $     DEVNAM = F$EXTRACE(0,2,DEVNAM)1 > $     IF DEVNAM .EQS. "WT" THEN GOTO IS_DECTERMr1 > $     IF DEVNAM .EQS. "TW" THEN GOTO IS_DECTERMa1 > $     IF DEVNAM .EQS. "FT" THEN GOTO IS_DECTERMa1 > $     IF DEVNAM .EQS. "RT" THEN GOTO IS_DECTERMt2 > $     IF DEVNAM .EQS. "TN" THEN GOTO IS_DECTERM 	 > $ ENDIFl > $ GOTO NOT_DECTERM    I FTAn devices are also created by the Session Support Utility (SSU).  Does  a VT420 count as a decterm?    -- r Robert Deininger rdeininger@mindspring.com    ------------------------------   Date: 5 Jan 2001 18:29:49 GMTn2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)4 Subject: Re: Suggestion for improvement for DECterm.6 Message-ID: <9353qt$dof$2@mailint03.im.hou.compaq.com>   In article <rdeininger-0501011239400001@user-2iveak3.dialup.mindspring.com>, rdeininger@mindspring.com (Robert Deininger) writes:A! :Does a VT420 count as a decterm?-  L   A VT420 is not a DECterm, and a DECterm (at least under DECwindows Motif) J   cannot specifically emulate a VT420 -- it can emulate various terminals D   from VT100 through VT340, and it can identify itself as a DECterm.  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Fri, 5 Jan 2001 11:28:17 -0500% From: "Islandco" <sales@islandco.com> * Subject: Used DS20E Special - One Only !!!/ Message-ID: <t5bt07pu2nkp7c@corp.supernews.com>l  K We are getting rid of this system - it was a rental and now is up for grabs) very cheaply   Configured as follows:   DS20e 6/500F 512mbn 9Gb Disk S3Trio64 2mb Video Cardh KZPBA-CA UW SCSI controllere Thin CD and Floppy DE500-BA 10/100 Ethernet 3 Power supplies Keyboard and Mouse Includedm Pedestal System   G We only have ONE used up for grabs - the others are new and refurbishedr   Your for only $64954* Or $ 10,899 with NEW VMS EIP license kit !     -- Island Computers US Corporation  2700 Gregory Streetf	 Suite 150' Savannah GA 31404  Tel: 912 447 6622n Fax: 912 201 0096a sales@islandco.com www.islandco.com  C This message and any files transmitted with it are confidential and-J may be privileged and/or subject to the provisions of privacy legislation.H They are intended solely for the use of the individual or entity to whomE they are addressed. If the reader of this message is not the intendede
 recipient,G please notify Island Computers US Corp immediately and then delete this0 message.I You are notified that reliance on, disclosure of, distribution or copying  of this message is prohibited.   ------------------------------  % Date: Fri, 05 Jan 2001 08:14:30 +0100r2 From: Nicholas Barnes <nicholas-za.barnes@ubs.com>M Subject: Using Telnet create_session (was: Re: Can DCL talk to TCP/IP ports?)s' Message-ID: <3A557456.57C15B52@ubs.com>e   Hi,n  B Thanks to everybody for their useful (and sometimes not so useful)E replies. Unfortunately, Perl and Kermit are out of the question. It'se+ got to be done from DCL (and maybe Telnet).a   JF Mezei wrote:n > 3 > $TELNET/CREATE_SESSION/PROT=NONE host_name port ss > $OPEN/READ/WRITE mylink TNAs:i  > $write mylink "a line of text" > $close mylinkt > $TELNET/DELETE_SESSSION s   9 I've been having a play with this and have hit a problem.   B Consider the following script (anonymised for public consumption):  ? $ telnet /create_session/prot=none smtpgate.somewhere.com 25 21v $ open/read/write mylink TNA21: 
 $loop_top:! $ read/time_out=2 mylink responsei $ write sys$output responsen $ goto loop_top   C There seems to be some sort of loop going on here, as the output iso (again, anonymised):   $ @linktest G 220 smtpgate.somewhere.com ESMTP sendmail 8.1.1 : XYZ satellite unifieds configB 500 Command unrecognized: "Data:  220 smtpgate.somewhere.com ESMTP sendmail 8.1.1 :"s9 500 Command unrecognized: " UBS satellite unified config"MH 500 Command unrecognized: "Data:   500 Command unrecognized: "Data:  220 smtpgate.somewhere.com ESMTP "- 500 Command unrecognized: "sendmail 8.1.1 :""aB 500 Command unrecognized: "Data:   500 Command unrecognized: " XYJ satellite unified config""E 500 Command unrecognized: "Data:   500 Command unrecognized: "Data:   & 500 Command unrecognized: "Data:  220"; 500 Command unrecognized: " smtpgate.somewhere.com ESMTP ""c< 500 Command unrecognized: "Data:   500 Command unrecognized:G "sendmail500 Command unrecognized: "Data:   500 Command unrecognized: "e. Data:   500 Command unrecognized: " XYZ satel"   etc. etc. etc. etc.   D Which looks to me as though the contents of the symbol 'response' isH being sent back to the TNA device. I've checked all logicals and I can't@ see any obvious reason for this to be happening. If the "$ writeD sys$output response" line is commented out, CTL+T still shows the IO count increasing very quickly.  
 Any ideas?   Nick.    ------------------------------  $ Date: Fri, 5 Jan 2001 12:26:42 -05002 From: "John Gemignani, Jr." <john@ossc.DELETE.net>Q Subject: Re: Using Telnet create_session (was: Re: Can DCL talk to TCP/IP ports?)e+ Message-ID: <3a5603e0$1@newsfeed.vitts.com>i  I The problem here is that the terminal has to be set NOECHO.  It's echoing I the SMTP responses back as command input.  Sorry that I forgot to mention F that, but I thought that I turned off echo on outgoing connections ... apparently I am mistaken.    -Johno    ? "Nicholas Barnes" <nicholas-za.barnes@ubs.com> wrote in messageh! news:3A557456.57C15B52@ubs.com...t >s > Hi,c >OD > Thanks to everybody for their useful (and sometimes not so useful)G > replies. Unfortunately, Perl and Kermit are out of the question. It'sn- > got to be done from DCL (and maybe Telnet).- >- > JF Mezei wrote:- > >-5 > > $TELNET/CREATE_SESSION/PROT=NONE host_name port s ! > > $OPEN/READ/WRITE mylink TNAs:." > > $write mylink "a line of text" > > $close mylinki > > $TELNET/DELETE_SESSSION s  > ; > I've been having a play with this and have hit a problem.  >sD > Consider the following script (anonymised for public consumption): >iA > $ telnet /create_session/prot=none smtpgate.somewhere.com 25 21-! > $ open/read/write mylink TNA21:  > $loop_top:# > $ read/time_out=2 mylink responsej > $ write sys$output responseg > $ goto loop_topo >5E > There seems to be some sort of loop going on here, as the output isL > (again, anonymised): >e
 > $ @linktestrI > 220 smtpgate.somewhere.com ESMTP sendmail 8.1.1 : XYZ satellite unified> > configD > 500 Command unrecognized: "Data:  220 smtpgate.somewhere.com ESMTP > sendmail 8.1.1 :"s; > 500 Command unrecognized: " UBS satellite unified config"oJ > 500 Command unrecognized: "Data:   500 Command unrecognized: "Data:  220  > smtpgate.somewhere.com ESMTP "/ > 500 Command unrecognized: "sendmail 8.1.1 :""1D > 500 Command unrecognized: "Data:   500 Command unrecognized: " XYJ > satellite unified config""E > 500 Command unrecognized: "Data:   500 Command unrecognized: "Data:t( > 500 Command unrecognized: "Data:  220"= > 500 Command unrecognized: " smtpgate.somewhere.com ESMTP "" > > 500 Command unrecognized: "Data:   500 Command unrecognized:I > "sendmail500 Command unrecognized: "Data:   500 Command unrecognized: "s0 > Data:   500 Command unrecognized: " XYZ satel" >n > etc. etc. etc. etc.- >-F > Which looks to me as though the contents of the symbol 'response' isJ > being sent back to the TNA device. I've checked all logicals and I can'tB > see any obvious reason for this to be happening. If the "$ writeF > sys$output response" line is commented out, CTL+T still shows the IO  > count increasing very quickly. >t > Any ideas? >g > Nick.s   ------------------------------  # Date: Fri, 05 Jan 2001 17:01:23 GMTt From: stevep3989@my-deja.com+ Subject: uVAX 3100 model80/90 hardware infot) Message-ID: <934uks$pog$1@nnrp1.deja.com>I   Hi  : I recently acquired a microVAX 3100 model 80 and model 90.; Unfortunately the previous owner of the systems didn't keepp9 the hardware documentation/owner's manual. Anyone know ift< these manuals are available anywhere on the web? (I couldn't find them at compaq.com.)-  ) One specific problem I'm trying to solve:   > The 3100-90 has an internal RRD42 cdrom drive (which is listed from >>> show dev ).  < I want install VMS from cd, but haven't managed to load a cdA into the RRD42. There is a button on the front panel (with a bent = arrow symbol), but pressing that doesn't eject the cd tray. It5 can't see any other way to get a disc into the drive.gB I'm wondering if the drive is faulty? - the green led on the front? doesn't flicker at all during system power-up, or subsequently.xE Is there any way of testing/exercising the drive from the >>> prompt?    Grateful for any pointers,  ( Regards, Steve Percival (Cambridge, UK).     Sent via Deja.comC http://www.deja.com/   ------------------------------  % Date: Fri, 05 Jan 2001 10:33:35 -0700S From: Kevin Handy <kth@srv.net>L/ Subject: Re: uVAX 3100 model80/90 hardware infop' Message-ID: <3A56056F.9F571527@srv.net>    stevep3989@my-deja.com wrote:  >  > Hi > < > I recently acquired a microVAX 3100 model 80 and model 90.= > Unfortunately the previous owner of the systems didn't keep ; > the hardware documentation/owner's manual. Anyone know ifi> > these manuals are available anywhere on the web? (I couldn't > find them at compaq.com.)2 > + > One specific problem I'm trying to solve:r > @ > The 3100-90 has an internal RRD42 cdrom drive (which is listed > from >>> show dev ). > > > I want install VMS from cd, but haven't managed to load a cdC > into the RRD42. There is a button on the front panel (with a bentw? > arrow symbol), but pressing that doesn't eject the cd tray. I 7 > can't see any other way to get a disc into the drive.oD > I'm wondering if the drive is faulty? - the green led on the frontA > doesn't flicker at all during system power-up, or subsequently.eG > Is there any way of testing/exercising the drive from the >>> prompt?   B The RRD40's require a CD caddy (or whatever it is called) in orderB to load a CD. This looks like a specialized case for a CD that youB slide into the drive (with the CD inside of the caddy).  The driveB doesn't contain a tray. You should check with whoever you got the C system from if he still has a caddy. They were common in the first  % generation of Computer CD-ROM drives.d   > Grateful for any pointers,   --  @ The people who vote do not determine the outcome of an election,6 it is the people who count the votes  -- Joseph Stalin  ? The counting isn't done, until the Democrats have won -- Algore-   ------------------------------  # Date: Fri, 05 Jan 2001 08:13:09 GMTg From: Dirk Munk <munk@home.nl>! Subject: Re: VAX to NT migration? ' Message-ID: <3A558215.1F37F1CE@home.nl>d  G There is a Vax/VMS emulation that runs on NT. It gets you MVII speed onjH a fast intel box. I haven't seen it, but maybe you can use it instead of  translating your assembler code.  G I doubt very much if a translation tool exists. Remember that assembleriH instructions have a direct relationship with processor instructions, andG a Vax processor differs enormously from a intel processor. Otherwise itaE might have been possible. For instance there are cross assemblers fori Vax > Motorola 68000.a   Regards,   Dirk      sol gongola wrote: > 7 > You can't easily translate the complex VAX (or alpha)o0 > instruction set to the relatively simple intel8 > platform instructions. Even less possible would be the5 > man-handling of the VMS system calls into somethingS; > for windows/NT, assuming VAX meant VMS although ultrix isn > still possible.  > 
 > sol gongolac >  > Stegner wrote: > >nL > > Might anyone know of a utility (either free or fee) to convert/translateE > > VAX Assembler (Macro?) to Windows/NT assembler?   I'm thinking ofh" > > migrating some applications... > >E > > thanks,e > > TJ   ------------------------------  % Date: Fri, 05 Jan 2001 14:59:49 +0000 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>i! Subject: Re: VAX to NT migration? ) Message-ID: <3A55E165.BC833161@bbc.co.uk>T   Stegner wrote:  J > Might anyone know of a utility (either free or fee) to convert/translateC > VAX Assembler (Macro?) to Windows/NT assembler?   I'm thinking ofc  > migrating some applications... >t  J How about kidnapping VAXman and Paul Sture and Arne  and locking them in a6 deep dark dungeon until they have the job done :-) :-)  F Seriously, this ain't gonna be easy, usually people write in assembler
 either forD performance or to access hardware. You will probably need to reverse
 engineer theno reimplement the app.  K Then again, I could be wrong as I've never tried this particular excercise.    Regards1   >6	 > thanks,  > TJ   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those oft MedAS or the BBC.n   ------------------------------  % Date: Fri, 05 Jan 2001 15:06:29 +0000g- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>r! Subject: Re: VAX to NT migration?N) Message-ID: <3A55E2F5.8A63AEA9@bbc.co.uk>    Dirk Munk wrote:  I > There is a Vax/VMS emulation that runs on NT. It gets you MVII speed on J > a fast intel box. I haven't seen it, but maybe you can use it instead of" > translating your assembler code. > I > I doubt very much if a translation tool exists. Remember that assemblerYJ > instructions have a direct relationship with processor instructions, andI > a Vax processor differs enormously from a intel processor. Otherwise it-G > might have been possible. For instance there are cross assemblers form > Vax > Motorola 68000.e >c  O A minor nit pick. A cross assembler is not a code translator. A cross-assembler3P as I understand it allows one to build code for one platform (eg 68k) on another  L platform (eg VMS). It does not allow one to convert eg VAX code to 68k code.  K That said, having written VAX and 68k assembler (a long while back) the two9P do have similarities (32 bit CISC etc) and in that particlar case maybe tools do  @ exist to convert the assembler from one architecture to another.   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukt  A I speak for myself only and my views in no way represent those ofe MedAS or the BBC.b   ------------------------------   Date: 5 Jan 2001 08:37:17 -0500g, From: koehler@eisner.decus.org (Bob Koehler)/ Subject: Re: voor de belgische Alpha gebruikerst+ Message-ID: <nEs+x7WIEbew@eisner.decus.org>   h In article <OFFB0B8EB5.16FCFFFF-ON882569CA.0064B152@foundation.com>, Shane.F.Smith@Healthnet.com writes: > M > Oh great. Now it's only a matter of time before someone posts in Klingon...  >  > Shaneg >   8 I take it you don't use EISNER (or you don't use NOTES).  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationl= NASA GSFC Flight Software       | Federal Sector, Civil Group E                                 | please remove ".aspm" when replyinga   ------------------------------  % Date: Fri, 05 Jan 2001 15:15:05 +0000 % From: Alan Greig <agreig@my-deja.com>a/ Subject: Re: voor de belgische Alpha gebruikersn8 Message-ID: <nnob5t8ec2gsmf4kfpj7h5fhtl7lo5m7jr@4ax.com>  A On Thu, 04 Jan 2001 22:49:46 GMT, Dirk Munk <munk@home.nl> wrote:o   >. > # >Shane.F.Smith@Healthnet.com wrote:f >> uN >> Oh great. Now it's only a matter of time before someone posts in Klingon... >>   >> Shane >>   > ! >Heh heh heh, it aint that bad...g >mI >Please remember that at the time the founding fathers of the U.S. had tooE >choose a national language fot their new nation, Dutch missed out byp >only 1 vote .... E >Maybe if they had high courts or recounts at that time, things might. >have been different now.   B Whover designed the Florida State election web site must have been= psychic because it contained a large section dealing with theiF importance of voting in very close elections and gave examples of suchC from US History. The national language was one example although I'mCE fairly sure it said the choice was between English and German.  MightVD have been Dutch though. I think it also claimed that some historians: say the vote never actually took place. Isn't history fun.     --
 Alan Greig   ------------------------------  $ Date: Fri, 5 Jan 2001 13:11:31 +0100- From: "Roland Hauk" <roland.hauk@indramat.de>N& Subject: ZIP and UNZIP for VAX-VMS 7.2- Message-ID: <934db0$p925@sunny.mannesmann.de>i   Hallo,  : Who knows, where I can find newest zipper for VAX-VMS7.2 .  5 I'm looking for files such as ZIP*.exe and UNZIP*.exe-   Seasons Greetings  Roland Hauk    ------------------------------  % Date: Fri, 05 Jan 2001 17:14:51 +0100e= From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com>a* Subject: Re: ZIP and UNZIP for VAX-VMS 7.2) Message-ID: <3A55F2FB.78C46C54@gtech.com>u   Roland Hauk wrote:< > Who knows, where I can find newest zipper for VAX-VMS7.2 . > 7 > I'm looking for files such as ZIP*.exe and UNZIP*.exe   D ftp://ftp.wku.eu/vms/fileserv/ have them with source and object-code, (so that you can link without a C compiler).   Arne   ------------------------------   Date: 5 Jan 2001 16:20:39 GMT-2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)* Subject: Re: ZIP and UNZIP for VAX-VMS 7.26 Message-ID: <934s8n$cgp$2@mailint03.im.hou.compaq.com>  ] In article <934db0$p925@sunny.mannesmann.de>, "Roland Hauk" <roland.hauk@indramat.de> writes:a; :Who knows, where I can find newest zipper for VAX-VMS7.2 .!6 :I'm looking for files such as ZIP*.exe and UNZIP*.exe  E   Please check the OpenVMS FAQ for pointers to the available download)E   sites -- I've updated the version of zip and unzip included on the i#   next OpenVMS Freeware, as well...A  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  # Date: Fri, 05 Jan 2001 16:18:59 GMT - From: goathunter@goatley.com (Hunter Goatley) * Subject: Re: ZIP and UNZIP for VAX-VMS 7.2/ Message-ID: <3a55f397.7055034@swen.process.com>e  0 On Fri, 5 Jan 2001 13:11:31 +0100, "Roland Hauk"  <roland.hauk@indramat.de> wrote:   >Hallo,y >l; >Who knows, where I can find newest zipper for VAX-VMS7.2 .e >n6 >I'm looking for files such as ZIP*.exe and UNZIP*.exe >y: You can always find the latest versions on FTP.WKU.EDU and its mirrors.  ! http://www2.wku.edu/www/fileserv/   % ftp://ftp.wku.edu/vms/unzip.alpha_exe # ftp://ftp.wku.edu/vms/unzip.vax_exep& ftp://ftp.wku.edu/vms/fileserv/zip.zip( ftp://ftp.wku.edu/vms/fileserv/unzip.zip  ; The .ZIP files include binaries; just UnZip and do @LINK ini: [.VMS-BINARIES] and all the .EXEs will be created for you.   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/u9 goathunter@goatley.com     http://www.goatley.com/hunter/o   ------------------------------   End of INFO-VAX 2001.010 ************************