1 INFO-VAX	Mon, 22 Jul 2002	Volume 2002 : Issue 401       Contents:( "no privilege" to view a job queue entry RE: CSA -> DSPP % Re: Cygwin/xfree86 and VMS/DecWindows % Re: Cygwin/xfree86 and VMS/DecWindows $ Re: filename argument in java on VMS HSV , Snaps and older databases # Re: HSV , Snaps and older databases # Re: HSV , Snaps and older databases # Re: HSV , Snaps and older databases # Re: HSV , Snaps and older databases & HSV Adaptive caching - how is it done?> Re: HSV snapshots vs. Spiralog ? (Was: CLI access to HSV's...) Re: I wanna buy this book... RE: INFO-VAX 2002.396 + Re: Itanic2 - the cHumPaq spin continues... + Re: Itanic2 - the cHumPaq spin continues... + Re: Itanic2 - the cHumPaq spin continues...  Laramie, anyone? OpenVMS webring:  Where's Mike? / Re: Powerstorm 300/350 cards and OpenGL on OVMS  SMISERVER Bytlim problem Re: SMTP question  RE: SMTP question  Re: Terminal input from DCL ( Re: Upgrading from 7.2-1 and latest ECOs Re: VMS commitment/ VMS Mail problems Internal/External receipient. C Re: WATCHER (was Re: Command Procedure to logoff inactive users...) ! Re: [Hobbyist] User account setup ! Re: [Hobbyist] User account setup   F ----------------------------------------------------------------------  % Date: Mon, 22 Jul 2002 15:17:40 +0200 8 From: "Tomasz Dryjanski" <tdryjanski.nospam@hotmail.com>1 Subject: "no privilege" to view a job queue entry , Message-ID: <ahh0ll$fq$1@news2.ipartners.pl>   Hi,   H Perhaps this is a lame question, but when I try to list jobs in a  batch queue, I receive the following:   & Batch queue SYS$BATCH, idle, on NODE::   Entry Jobname Username Status    ----- ------- -------- ------   % 2 RESTORE_STBY no privilege File: ...   J I have all system privileges granted (both authorized and default), BYPASS+ is among them. The queue is set as follows:   & Batch queue SYS$BATCH, idle, on NODE::  / /BASE_PRIORITY=4 /JOB_LIMIT=100 /OWNER=[SYSTEM]  /PROTECTION=(S:M,O:D,G:R,W:S)   
 i.e. default.    What can cause this message?  + The system in question is OpenVMS V7.2-1H1.        TIA -    T. D.    ------------------------------  % Date: Mon, 22 Jul 2002 10:06:46 -0400 * From: WILLIAM WEBB <WWEBB1@email.usps.gov> Subject: RE: CSA -> DSPP- Message-ID: <0033000073419225000002L052*@MHS>   B =0ASo did you drop the webminions an email asking them to do this?   WWWebb   -----Original Message-----/ From: Info-VAX-Request@Mvb.Saic.Com at INTERNET # Sent: Friday, July 19, 2002 9:59 AM B To: Webb, William W Raleigh, NC; Info-VAX@Mvb.Saic.Com at INTERNET Subject: RE: CSA -> DSPP     <conspiracy>  > http://h21007.www2.hp.com/partner/technologies/1,2610,,00.html   contains NOTHING about OpenVMS.     H Why would HP want their partners to know anything about OpenVMS? But th= enH again, this is an HP site, not a Compaq site.  Seems to me that having = a F link to existing OpenVMS resources on www.compaq.com is something thatC should be done. It would take the webmaster all of 2 minutes to do.    </conspiracy>=   ------------------------------    Date: 22 Jul 2002 05:03:51 -0700# From: tadamsmar@aol.com (Tom Adams) . Subject: Re: Cygwin/xfree86 and VMS/DecWindows= Message-ID: <793af3df.0207220403.7bc1238d@posting.google.com>   ~ Jordi Guillaumes i Pons <jguilla-at-attglobal-dot-net@nospam.please> wrote in message news:<3D387A58.2030004@nospam.please>... > Tom Adams wrote:4 > > Will CYGWIN/FREE86 work with VMS and DecWindows?. > > How can they be configured to communicate?G > > Can it be done with and without Pathworks? How is it done each way?  > > Thanks for any input.  > ? > It will work, although some DEC fonts doesn't have a XFree86  7 > equivalent,and some things will not look very pretty.  > I > My recipe is as follows (I'm sure it can be enhanced in a milion ways):  > - > 1) Open two CygWin windows under windows ;)  > 2 > 2) In one of them use the following enchantment: >  > $ XWin -ac > H > 3) In the other, telnet to your VMS machine. Now you have two chances: > 9 > a) Do a SET DISPLAY/CREATE/NODE=yourpcname/TRANSP=tcpip F > b) $ CREATE/TERMINAL/DETACH at your discretion, or MCR DECW$VUE, or . > whatever Decwindows program you want to run. >  > -OR- > # > a) Login to the VMS box as SYSTEM 8 > b) $SET DISPLAY/CREATE/NODE=yourpcname/TRANSPORT=tcpip > c) $ MCR DECW$STARTLOGIN >   E Thank you, dear wizard. I had overlooked the TRAN qualifier during my / own perusal of the tomes, leaving me befuddled.   D But isn't an advance offering required so that the system will deignF to recognize the node's name "youpcname" when it arrives at the gates.C  If I am not mistaken, I was told it in response to the chant "HELP F SET DISPLAY".  But, it was in a language that I little comprehend, andF I did not understand how to cast the spell that causes the node's nameB to materialize in some network database on my VMS system.  The our! local wizard says that NCP is not & involved, since it only covers DECnet.  K > Using the first method you won't have any window manager (you can enable  I > the default XFree86 twm using the usual "unix" way). Using the second,  @ > you'll get the DecWindows login widget and you'll have a full H > DECWindows/CDE desktop. The window manager will be running on the VMS I > side, so the widgets will be managed thru the network (I've been doing  N > this using a uVAX 3300 over 10mbps ethernet and the performance is not bad). >  > Hope that this helps.  >  > JG   ------------------------------  + Date: Mon, 22 Jul 2002 14:11:34 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk. Subject: Re: Cygwin/xfree86 and VMS/DecWindows+ Message-ID: <ahh3qm$e56$1@aquila.mdx.ac.uk>   c In article <793af3df.0207220403.7bc1238d@posting.google.com>, tadamsmar@aol.com (Tom Adams) writes:  >Jordi Guillaumes i Pons <jguilla-at-attglobal-dot-net@nospam.please> wrote in message news:<3D387A58.2030004@nospam.please>...  >> Tom Adams wrote: I >> 3) In the other, telnet to your VMS machine. Now you have two chances:  >>  : >> a) Do a SET DISPLAY/CREATE/NODE=yourpcname/TRANSP=tcpipG >> b) $ CREATE/TERMINAL/DETACH at your discretion, or MCR DECW$VUE, or  / >> whatever Decwindows program you want to run.  >>   >> -OR-  >>  $ >> a) Login to the VMS box as SYSTEM9 >> b) $SET DISPLAY/CREATE/NODE=yourpcname/TRANSPORT=tcpip  >> c) $ MCR DECW$STARTLOGIN  >>   > F >Thank you, dear wizard. I had overlooked the TRAN qualifier during my0 >own perusal of the tomes, leaving me befuddled. > E >But isn't an advance offering required so that the system will deign G >to recognize the node's name "youpcname" when it arrives at the gates. D > If I am not mistaken, I was told it in response to the chant "HELPG >SET DISPLAY".  But, it was in a language that I little comprehend, and G >I did not understand how to cast the spell that causes the node's name C >to materialize in some network database on my VMS system.  The our " >local wizard says that NCP is not' >involved, since it only covers DECnet.  >   M Since your using TCPIP translating the "yourpcname" to the correct IP address ; is handled as always by either DNS or the local hosts file.   G If the VMS system is setup to use DNS and "yourpc.domain" is setup in a L properly configured accessible DNS server somewhere on the internet then you are all set.I If not then you need to get "yourpc.domain" set up in the local host file H on the VMS system. With DEC TCPIP Services / UCX  this is done using the+ UCX set host/address=x.x.x.x yourpc.domain   command.  M This should probably be enough to get it working with most PC based X-servers K since they generally come with no security restrictions restricting who can K connect. Most/all do though come with the ability to close down this gaping M security hole - but you'll need to see the documentation for the PCs X server  package.   J VMS workstations in contrast default to not allowiung any remote X windows3 connections - you have to specifically enable them.       
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------    Date: 22 Jul 2002 08:06:32 -0600- From: koehler@encompasserve.org (Bob Koehler) - Subject: Re: filename argument in java on VMS 3 Message-ID: <u0LrVCW$lpVC@eisner.encompasserve.org>   X In article <sb_Z8.39372$Z6.469626@nnrp1.uunet.ca>, "June Young" <jy@dymaxion.ca> writes: > " > I also set up the java$classpath > sh log java$classpath ? >    "JAVA$CLASSPATH" = "DYM$DISK:[DYMAX.LUCENE]LUCENE-1_2.JAR"  > (LNM$PROCESS_TABLE) 9 >         = "DYM$DISK:[DYMAX.LUCENE]LUCENE-DEMOS-1_2.JAR"  >         = "[]" > M > I know the class path should be right,because it works on NT, but on vms, I  > get: >   E    Most likely your .jar files are not in stream-lf.  Nothing will be +    found in them if they are not stream-lf.   K    The usual process for getting .jar files (or .zip files with classes in  "    them) from another platform is:         1) FTP using binary mode  C       2) set file xxxxx.jar /attributes=(rfm:stmlf,mrs:0,lrl:32767)    ------------------------------    Date: 22 Jul 2002 09:33:41 -0600+ From: young_r@encompasserve.org (Rob Young) ( Subject: HSV , Snaps and older databases3 Message-ID: <Ebp3tbVW0SqY@eisner.encompasserve.org>   P In article <eesgju4hbir9vgidm4fpteld2pqmomrual@4ax.com>, jlsue <jlsuexxxz@screa 8 >> On Fri, 19 Jul 2002 01:53:43 GMT, "David J. Dachtera" >>   >>H >>IMO, the point is that the loss if CLI and/or virtual console terminal& >>will "break" many sites' procedures. > @ >First off, there SSSU for HSV scripting.  I believe it's on the >platform kit for the OS.  >  >>H >>A former site of mine has backup procedures that are 100% dependent onG >>being able to command HSx's from within an OpenVMS batch job to split J >>mirror-sets. They can't afford the performance degradation of shadow-setJ >>merging (50+ shadow-sets, can't guarantee they'll always mini-merge), soC >>this was the only way to guarantee that would never happen again.  > D >Second, the HSV is VIRTUAL storage.... you don't split-off mirrors. >It just doesn't work this way.  > D >Unfortunately we don't have EVM support for the HSV yet, as this is: >how most of our EBS solutions integrate with the process. > B >You can write host-based scripts using SSSU, but in spite of whatF >Andrew's been trying to imply, I don't necessarily recommend it since5 >HP can't support all of the issues that may come up.  > F >On the other hand, I just implemented some backups for a customer whoC >insisted on using it (from VMS no less).  It worked.  We'd quiesce C >their application, then run a script to take all of the snapshots, E >present them to the backup host, and then use SYSMAN IO AUTO to make ( >them available for mounting and backup.  B 	The issue of course is older database technology whereby you must7 	quiesce the volume, "break off" a copy and back it up.   = 	While EVM is the recommended solution, from its description:   @ http://www.compaq.com/products/quickspecs/10432_na/10432_na.html  I Enterprise Volume Manager (EVM) is a web-enabled application program that J enables users to create and execute jobs that use storage controller-based snapshots and clones.   A 	It doesn't appear that runs out of a VMS batch queue.  So if the + 	following line up (end-user requirements):     2 			1)  No Volume Shadowing as HSV does not support 				mini-merge8 			2)  Heritage database product that requires quiescing& 				volumes and "break out" (or snap).2 			3)  EVM supported "clone" or "snapclone" method  > 	Seems 3) means there is an operator sitting there clicking on/ 	something and we all know how well that works.    	What am I missing?   I 	What is the best solution if a backup is being driven from VMS and 1) +  
 	2) are true?    				Rob    ------------------------------  + Date: Mon, 22 Jul 2002 15:50:13 +0000 (UTC)  From: david20@alpha1.mdx.ac.uk, Subject: Re: HSV , Snaps and older databases+ Message-ID: <ahh9jl$g1b$1@aquila.mdx.ac.uk>   a In article <Ebp3tbVW0SqY@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: Q >In article <eesgju4hbir9vgidm4fpteld2pqmomrual@4ax.com>, jlsue <jlsuexxxz@screa  9 >>> On Fri, 19 Jul 2002 01:53:43 GMT, "David J. Dachtera"  >>>  >>> G >>On the other hand, I just implemented some backups for a customer who D >>insisted on using it (from VMS no less).  It worked.  We'd quiesceD >>their application, then run a script to take all of the snapshots,F >>present them to the backup host, and then use SYSMAN IO AUTO to make) >>them available for mounting and backup.  > C >	The issue of course is older database technology whereby you must 8 >	quiesce the volume, "break off" a copy and back it up. > > >	While EVM is the recommended solution, from its description: > A >http://www.compaq.com/products/quickspecs/10432_na/10432_na.html  > J >Enterprise Volume Manager (EVM) is a web-enabled application program thatK >enables users to create and execute jobs that use storage controller-based  >snapshots and clones.   > B >	It doesn't appear that runs out of a VMS batch queue.  So if the, >	following line up (end-user requirements): >  > 3 >			1)  No Volume Shadowing as HSV does not support  >				mini-merge 9 >			2)  Heritage database product that requires quiescing ' >				volumes and "break out" (or snap). 3 >			3)  EVM supported "clone" or "snapclone" method  > ? >	Seems 3) means there is an operator sitting there clicking on 0 >	something and we all know how well that works. >  >	What am I missing? > J >	What is the best solution if a backup is being driven from VMS and 1) +  >	2) are true? >  >				Rob >   N Unless you have a logging file system and flush out the caches I don't see how@ you can safely clone a volume without quiescing the application.  F Am I missing something here ?  Cloning at the controller level withoutM quiescing the application or flushing OS file system caches seems to me to be E running the risk of capturing the data in the middle of transactions. O eg To complete a transaction you need to update two files. You have updated one A file but then before you update the second the clone is produced. N A logging file system like Tru64s ADVFS will help to some extent with this butL that doesn't help if you were using UFS. Also although Oracle uses extensiveO logs in a similar way to a logging file system it requires you to have produced K a checkpoint (which amongst other things flushes out caches) before you can  take an online backup.  D Hence I don't think it is just "heritage" applications that would be potentially at risk here.       
 David Webb VMS and Unix team leader CCSS Middlesex University     ------------------------------    Date: 22 Jul 2002 11:36:19 -0600+ From: young_r@encompasserve.org (Rob Young) , Subject: Re: HSV , Snaps and older databases3 Message-ID: <88WphCoBv34W@eisner.encompasserve.org>   L In article <ahh9jl$g1b$1@aquila.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:c > In article <Ebp3tbVW0SqY@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes: R >>In article <eesgju4hbir9vgidm4fpteld2pqmomrual@4ax.com>, jlsue <jlsuexxxz@screa : >>>> On Fri, 19 Jul 2002 01:53:43 GMT, "David J. Dachtera" >>>>   >>>>H >>>On the other hand, I just implemented some backups for a customer whoE >>>insisted on using it (from VMS no less).  It worked.  We'd quiesce E >>>their application, then run a script to take all of the snapshots, G >>>present them to the backup host, and then use SYSMAN IO AUTO to make * >>>them available for mounting and backup. >>D >>	The issue of course is older database technology whereby you must9 >>	quiesce the volume, "break off" a copy and back it up.  >>? >>	While EVM is the recommended solution, from its description:  >>B >>http://www.compaq.com/products/quickspecs/10432_na/10432_na.html >>K >>Enterprise Volume Manager (EVM) is a web-enabled application program that L >>enables users to create and execute jobs that use storage controller-based >>snapshots and clones.  >>C >>	It doesn't appear that runs out of a VMS batch queue.  So if the - >>	following line up (end-user requirements):  >> >>4 >>			1)  No Volume Shadowing as HSV does not support >>				mini-merge: >>			2)  Heritage database product that requires quiescing( >>				volumes and "break out" (or snap).4 >>			3)  EVM supported "clone" or "snapclone" method >>@ >>	Seems 3) means there is an operator sitting there clicking on1 >>	something and we all know how well that works.D >> >>	What am I missing?d >>K >>	What is the best solution if a backup is being driven from VMS and 1) + n >>	2) are true?  >>	 >>				Robe >> > P > Unless you have a logging file system and flush out the caches I don't see howB > you can safely clone a volume without quiescing the application. > H > Am I missing something here ?  Cloning at the controller level withoutO > quiescing the application or flushing OS file system caches seems to me to be.G > running the risk of capturing the data in the middle of transactions.mQ > eg To complete a transaction you need to update two files. You have updated onenC > file but then before you update the second the clone is produced.dP > A logging file system like Tru64s ADVFS will help to some extent with this butN > that doesn't help if you were using UFS. Also although Oracle uses extensiveQ > logs in a similar way to a logging file system it requires you to have produced:M > a checkpoint (which amongst other things flushes out caches) before you cane > take an online backup. > F > Hence I don't think it is just "heritage" applications that would be > potentially at risk here.  >   C 	I had this lengthy reply and realized I was in several directions.i  > 	While what you bring up is correct, I believe you would be in1 	the same spot with HSJ/HSG.  The wrinkle is thateF 	you would probably always have situations where you have inconsistentD 	views of volumes.  As a for instance (from Encompasserve discussionA 	a year or two ago), you could be in the middle of a bucket split-F 	when backup occurs.  Something as simple as VMS mail bucket splittingD 	at the point in time that MAIL.MAI was backup scanned, you won't be 	getting a consistent restore.  ? 	While what you bring up is of course a concern, I am trying toIE 	drill down to a specific case that touches many of us (I'm certain), A 	i.e. databases that need to be quiesced and busted out.  What isM 	the solution?   				Rob    ------------------------------  + Date: Mon, 22 Jul 2002 17:03:11 +0000 (UTC)0 From: david20@alpha1.mdx.ac.uk, Subject: Re: HSV , Snaps and older databases+ Message-ID: <ahhdsf$hb0$1@aquila.mdx.ac.uk>   a In article <88WphCoBv34W@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:-M >In article <ahh9jl$g1b$1@aquila.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:2d >> In article <Ebp3tbVW0SqY@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:S >>>In article <eesgju4hbir9vgidm4fpteld2pqmomrual@4ax.com>, jlsue <jlsuexxxz@screa t; >>>>> On Fri, 19 Jul 2002 01:53:43 GMT, "David J. Dachtera"  >>>>>  >>>>>t >> wQ >> Unless you have a logging file system and flush out the caches I don't see how C >> you can safely clone a volume without quiescing the application.  >> oI >> Am I missing something here ?  Cloning at the controller level without P >> quiescing the application or flushing OS file system caches seems to me to beH >> running the risk of capturing the data in the middle of transactions.R >> eg To complete a transaction you need to update two files. You have updated oneD >> file but then before you update the second the clone is produced.Q >> A logging file system like Tru64s ADVFS will help to some extent with this butoO >> that doesn't help if you were using UFS. Also although Oracle uses extensive R >> logs in a similar way to a logging file system it requires you to have producedN >> a checkpoint (which amongst other things flushes out caches) before you can >> take an online backup.a >> XG >> Hence I don't think it is just "heritage" applications that would be  >> potentially at risk here. >> t >nD >	I had this lengthy reply and realized I was in several directions. >w? >	While what you bring up is correct, I believe you would be in: >	the same spot with HSJ/HSG.  h  K No because you would quiesce the applications which were updating the disksE  before you split the mirrorsets.  N That was the point I was trying to get across that this needs to be controlled from the OS.H Databases such as oracle are the prime example but this is more general.  L Sure for many purposes a less than perfect backup is "good enough". If thats. the case then you can use backup/ign=interlock  K However sometimes thats not good enough. Splitting shadowsets or mirrorsets I isn't much better than backup/ign=interlock unless you have quiesced youre
 applications.i       > The wrinkle is thataG >	you would probably always have situations where you have inconsistentsE >	views of volumes.  As a for instance (from Encompasserve discussion B >	a year or two ago), you could be in the middle of a bucket splitG >	when backup occurs.  Something as simple as VMS mail bucket splittingoE >	at the point in time that MAIL.MAI was backup scanned, you won't beo >	getting a consistent restore.s >C  L And if you were really concerned about this you would overcome it by loggingI everybody who might be using mail off and stopping any SMTP mail deliverynE processes before you split the disks for the backups ie quiescing the  application.    @ >	While what you bring up is of course a concern, I am trying toF >	drill down to a specific case that touches many of us (I'm certain),B >	i.e. databases that need to be quiesced and busted out.  What is >	the solution?t >  >				Rob >     O Fine. I was just put off by the use of the phrase "heritage applications" whichUP make it sound as though these applications are old fashioned and doing something1 which a "modern" application wouldn't need to do.   
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------    Date: 22 Jul 2002 12:18:31 -0600+ From: young_r@encompasserve.org (Rob Young)s, Subject: Re: HSV , Snaps and older databases3 Message-ID: <iq1Lkg0P$f05@eisner.encompasserve.org>   L In article <ahhdsf$hb0$1@aquila.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:c > In article <88WphCoBv34W@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:rN >>In article <ahh9jl$g1b$1@aquila.mdx.ac.uk>, david20@alpha1.mdx.ac.uk writes:e >>> In article <Ebp3tbVW0SqY@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:CT >>>>In article <eesgju4hbir9vgidm4fpteld2pqmomrual@4ax.com>, jlsue <jlsuexxxz@screa < >>>>>> On Fri, 19 Jul 2002 01:53:43 GMT, "David J. Dachtera" >>>>>> B >>>>>> >>> R >>> Unless you have a logging file system and flush out the caches I don't see howD >>> you can safely clone a volume without quiescing the application. >>> J >>> Am I missing something here ?  Cloning at the controller level withoutQ >>> quiescing the application or flushing OS file system caches seems to me to behI >>> running the risk of capturing the data in the middle of transactions.VS >>> eg To complete a transaction you need to update two files. You have updated onevE >>> file but then before you update the second the clone is produced. R >>> A logging file system like Tru64s ADVFS will help to some extent with this butP >>> that doesn't help if you were using UFS. Also although Oracle uses extensiveS >>> logs in a similar way to a logging file system it requires you to have produced O >>> a checkpoint (which amongst other things flushes out caches) before you can  >>> take an online backup. >>> H >>> Hence I don't think it is just "heritage" applications that would be >>> potentially at risk here.o >>>  >>E >>	I had this lengthy reply and realized I was in several directions.y >>@ >>	While what you bring up is correct, I believe you would be in  >>	the same spot with HSJ/HSG.   > M > No because you would quiesce the applications which were updating the diskse" > before you split the mirrorsets. > P > That was the point I was trying to get across that this needs to be controlled > from the OS.J > Databases such as oracle are the prime example but this is more general. > N > Sure for many purposes a less than perfect backup is "good enough". If thats0 > the case then you can use backup/ign=interlock > M > However sometimes thats not good enough. Splitting shadowsets or mirrorsets K > isn't much better than backup/ign=interlock unless you have quiesced your@ > applications." >   9 	Right... and that is exactly what I am highlighting whena 	I said at the outset:  B 	The issue of course is older database technology whereby you must7 	quiesce the volume, "break off" a copy and back it up.D  ; 	How do you do that with HSV?  You don't apparently (or youn2 	do it but it is unsupported as jlsue points out).   >  >  >> The wrinkle is thatH >>	you would probably always have situations where you have inconsistentF >>	views of volumes.  As a for instance (from Encompasserve discussionC >>	a year or two ago), you could be in the middle of a bucket splitnH >>	when backup occurs.  Something as simple as VMS mail bucket splittingF >>	at the point in time that MAIL.MAI was backup scanned, you won't be  >>	getting a consistent restore. >> > N > And if you were really concerned about this you would overcome it by loggingK > everybody who might be using mail off and stopping any SMTP mail deliverytG > processes before you split the disks for the backups ie quiescing the  > application. >   % 	Sure... that is how you would do it.>   >  >>A >>	While what you bring up is of course a concern, I am trying tonG >>	drill down to a specific case that touches many of us (I'm certain),rC >>	i.e. databases that need to be quiesced and busted out.  What is  >>	the solution? >> >  > L > Fine. I was just put off by the use of the phrase "heritage applications"  > whichtI > make it sound as though these applications are old fashioned and doing r > somethinga3 > which a "modern" application wouldn't need to do.' >   = 	Put off?  Okay.  But I guess if something is 20 years old oro? 	older, it is safe to say it is heritage.  Anyhow, I was tryingiG 	to highlight the fact that there are a number of database technologies E 	that are older that don't have the luxury of being put in hot backuptB 	mode whereby transactions are logged to archive logs, i.e. Oracle@ 	and/or RDB and of course the language may be munged but you get 	the point.A   				Robn   ------------------------------    Date: 22 Jul 2002 12:03:39 -0600+ From: young_r@encompasserve.org (Rob Young) / Subject: HSV Adaptive caching - how is it done?k3 Message-ID: <Lglqt+vNiz4u@eisner.encompasserve.org>M  ? http://www.compaq.com/products/quickspecs/11008_na/11008_na.pdfn  , Read-Ahead and Adaptive Read Caching Support  L Read caching increases performance by satisfying host read requests from theF controller cache memory instead of from the physical disks. Enterprise. read cache has the following three properties:  J   Normal read caching - data read from physical disks and stored in cache.N   Read-ahead - controller anticipates host need for sequential data and stores" it in cache in advance of request.I   Adaptive read caching - adaptive caching algorithms determine whether atL certain I/O stream is likely to benefit from caching in order to efficiently manage the cache resource.    = 	As the subject line suggests, what is adaptive caching?  Cans= 	anyone provide an example of where adaptive caching would benB 	a "win" whereas read-ahead wouldn't?  Is it recognizing a pattern< 	of block hits and pre-fetching based on a pattern?  Thanks.   				Robo   ------------------------------  # Date: Mon, 22 Jul 2002 06:32:52 GMTr? From: Jim.Johnson@software-exploration.nospam.com (Jim Johnson)hG Subject: Re: HSV snapshots vs. Spiralog ? (Was: CLI access to HSV's...)k. Message-ID: <3d3ba3f9.392254@news.demon.co.uk>  E On 22 Jul 2002 05:01:01 +0800, Paul Repacholi <prep@prep.synonet.com>  wrote:  B >Jim.Johnson@software-exploration.nospam.com (Jim Johnson) writes: >aH >> On 20 Jul 2002 12:08:30 +0800, Paul Repacholi <prep@prep.synonet.com>	 >> wrote:b > , >> >jlsue <jlsuexxxz@screaminet.com> writes: >sF >> >> Now it is being done by the hardware and any OS can benefit from	 >> >> it.. > ? >> >> (Note, I do not mean to imply that the same work done for E >> >> host-based snapshots was migrated to hardware, I'd have no ideae! >> >> if that actually happened.)  >nG >> >So how does the HSV ensure that *at the `split time'* all of the on D >> >disk data is consistant? That there are no suprises in the block
 >> >cache? > D >> Without knowing how it was actually built, but working from first >> principles: > G >> How does the HSV ensure that the data is consistent at the time of a  >> power outage? >eL >It doesn't! It just hangs on to what ever it has. If you run out of batteryA >and the caches and memory hit the floor, well, enjoy the wait...c  ) Still working from first principles here:s  B It should be pretty clear that the battery is not going to run outF during a successful snapshot operation.  Loss of battery life is a red herring for this discussion.   So, let me re-phrase my point>  B In the absence of losing the battery power, an HSV presents a diskE that is consistent with the I/Os done to it up to a specific point in D time.  Ignore where it is getting that data -- be it from its cache,? the on-disk cache, or the oxide.  The point is that the HSV canuB always, during operation, present a wholly consistent view of that disk.e  D It is therefore possible, reasonable, and probably required that anyA snapshot facility operate against that presented view of the diskoB contents, rather than some lower (inconsistent) state of the data.E Doing so means that a snapshot operation will present a snapshot thathE matches the I/Os done to a disk up to a specific point in time.  ThatnD is sufficient for a consistent snapshot.  No host operations need be	 involved.v  D Does it work this way?  I don't know, but I'd be willing to bet that0 it either does, or it uses something very close.     >  >> It should be the same logic.  >T >-- = >Paul Repacholi                               1 Crescent Rd.,e8 >+61 (08) 9257-1001                           Kalamunda.A >                                             West Australia 6076 / >Raw, Cooked or Well-done, it's all half baked.fG >EPIC, The Architecture of the future, always has been, always will be.a   Jim Johnsoni Software Exploration, Ltd.) (remove '.nospam' from the reply address)m   ------------------------------    Date: 22 Jul 2002 02:29:18 -0700) From: jbrankin@ntlworld.com (Jim Brankin)"% Subject: Re: I wanna buy this book...o= Message-ID: <863f19d6.0207220129.25a652f0@posting.google.com>e   These people have it.   9 http://www.pcbooks.co.uk/searchresults.aspx?txtsearch=Vax0    g "Brian Tillman" <tillman_brian@notnoone.notnohow.com> wrote in message news:<3d386e1d$1@news.si.com>...  > >I want buy this book: > >r@ > > - WRITING VAX/VMS APPLICATIONS USING PASCAL - THEOS DE KLERK > N > http://www.amazon.com/exec/obidos/ASIN/155558067X/qid=1027108336/sr=1-1/ref= > sr_1_1/002-9385472-6435254   ------------------------------  % Date: Mon, 22 Jul 2002 09:11:44 -0500x$ From: "Art Beane" <beane@petris.com> Subject: RE: INFO-VAX 2002.3967 Message-ID: <003501c23189$b5a44f00$342810ac@petris.com>   & >Date: Sat, 20 Jul 2002 01:27:43 +0200. >From: Didier Morandi <Didier.Morandi@Free.fr>E >Subject: Re: Tops-10/Tops-20 features not in VMS ?, was: Re: Lookingp for terminal' >Message-ID: <3D38A06C.7191D2E@Free.fr>e >cC >Art, please do not post in HTML, this is not Netiquette-compliant.  >n >:-) >b >D..   VERY sorry.   G The wrong draft got posted, anyway. I had meant to recount some ancientvE history. When Digital first started the AI product group, we tried tooF get a port of Interlisp-10 to VAX to buy time-to-market for the CommonA LISP standards activity and its VAX LISP implementation. DWIM wasa@ considered to be a feature many users would want, and there wereH considerable discussions (arguments?) on whether to include it in CommonC LISP and in what form. It had devolved to a simple spelling checkera before it was dropped.   ------------------------------  # Date: Mon, 22 Jul 2002 06:49:55 GMTe* From: "Bill Todd" <billtodd@metrocast.net>4 Subject: Re: Itanic2 - the cHumPaq spin continues...C Message-ID: <nWN_8.214427$iX5.11311035@bin3.nnrp.aus1.giganews.com>M  , "sasadmin" <jec@nospam.net> wrote in message, news:87vissba5y.fsf@Alethion.systasis.net...   ...i  ? > I would ask two things intended to encourage you to keep your> > reader in mind as you write: >         1) Provide urls.  J I usually try to the first time or two that I introduce a reference (e.g.,H as I have in past posts for the McKinley time-line through 2005), and inJ this case indicated spec.org as the source of most of the numbers I quotedC (navigating their site is so straight-forward that I did not botherr% referring to the specific page:  it'stH http://www.spec.org/osg/cpu2000/results/cpu2000.html ).  But I forgot to$ include the URL for the paper itselfL  http://h21007.www2.hp.com/dspp/files/unprotected/itanium2.pdf - and while IK was able to track it down in my history folder I'm damned if I can rememberoI how I got there, though it was most likely via a reference from somethingm
 else I read).h    I know google, spec &c areyD >         available. However, your criticisms require specific urls,   And I gave it:  spec.org.y    and5 >         good writing style includes a bibliography.m  H I'm not particularly wedded to 'good writing style':  while I do make anI effort to be coherent, I don't place a high priority on retaining readersl. who aren't sufficiently interested to keep up.   > E >         2) The topics you discuss are complex, and it's easy to get F >         lost in relative comparisons. Your style combines subjective? >         and objective argument techniques; which style can beiB >         difficult to parse.  You dilute your presentation in theH >         following sentence: "... accepted by SPEC it's *well under 50%A >         higher* than the 750 MHz 8700's SPECint base" (emphasissG >         added). Is the number higher or lower? I will usually re-read D >         such sentences to understand the relative magnitude of theD >         comparison. Employing "under ... higher" obscures the true >         relationship.t  J Sorry for any lack of clarity there.  I don't see a parsing ambiguity, but5 can see how parsing it might at least give one pause.    >-D >         An example of the subjective/objective conflation is: "AndG >         McKinley's integer performance is shown as *noticeably higherrD >         than* the coming PA8800, in contrast to the estimates I'veD >         seen (which *if anything appear conservative*) that PA8800F >         will debut with a *score of about 900*..." I would encourage@ >         you to reduce subjective comments in favor of absolute >         numbers.  K When one is dealing with estimates, absolute numbers are somewhat difficulttJ to come by.  This particular estimate came from a c.o.v. post by Ken GreenB on July 10th.  And the feeling that an estimate may be if anythingJ conservative is even more resistant to quantification - but still entirelyJ legitimate to express (in this case, my feeling was that the PA8800 shrinkH from the 8700's 180 nm process to 130 nm might well bump SPECint up moreI than just 200 points from the roughly 700 - itself a SWAG - value for thet new 875 MHz 8700).  -  The subjective case blunts the force of yours@ >         argument when you can obviously present the numbers as >         objective evidence.f  H I'm not in the business of convincing people, just of laying out data ofI varying (and where possible specified) degrees of solidity and suggesting>B conclusions that seem reasonable to draw from them.  If it prompts? discussion on points people would like further clarified, fine.r  K While I do feel it's important to provide reality-checks in response to theeL garbage cHumPaq continues to dish out, both directly and via its sycophants,K my commitment stops well short of becoming the kind of careful, calculating G politician/orator it would take to be optimally effective in convincingvD those uninterested in forming judgements based on the actual merits.   - bill   ------------------------------  + Date: Mon, 22 Jul 2002 11:12:26 +0000 (UTC)i From: david20@alpha2.mdx.ac.uk4 Subject: Re: Itanic2 - the cHumPaq spin continues...+ Message-ID: <ahgpaq$agj$1@aquila.mdx.ac.uk>   l In article <40ac1f47.0207211403.44ef226@posting.google.com>, loopnz@hotmail.com (loopnz@hotmail.com) writes:v >"Bill Todd" <billtodd@metrocast.net> wrote in message news:<LPt_8.198334$iB1.10663413@bin4.nnrp.aus1.giganews.com>... >oE >> But we do glean the new nugget of information on page 29 that "ThemJ >> performance of the Itanium 2 processor's IA-32 engine is expected to beN >> comparable with a 300 MHz Pentium Pro."  Wow - that'll sure cut Hammer IA32N >> performance down to size!  Especially since if one runs a code mix heavy onK >> the IA32 side the IA32 code will consume most of the available processoreF >> time even though only running at about 15% of native performance... >eG >Why again are we VMS users concerned about Hammer? Oh right, Elvis wasbC >sighted in one of the VMS labs doing a secret VMS port to Hammer - E >after he lost his job in accounting trying to cook the books to showgG >that VMS was making billions of dollars in profits and that Compaq x86l* >servers were dragging the company down...  O We are concerned with Hammer because if Hammer creams Itanic then the future ofo! VMS becomes even more uncertain.  B Itanics future is tied to how Intel is forced to react to Hammer. * VMSs future is (currently) tied to Itanic.  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Mon, 22 Jul 2002 14:44:12 GMT # From: "John Smith" <a@nonymous.com> 4 Subject: Re: Itanic2 - the cHumPaq spin continues...J Message-ID: <0TU_8.169478$WJf1.72829@news01.bloor.is.net.cable.rogers.com>  5 "Bob Ceculski" <bob@instantwhip.com> wrote in message-7 news:d7791aa1.0207211632.466c7a3f@posting.google.com...m >oF > hp has not stopped selling vms, they have stated they are selling toH > target markets, probably healthcare, finance and govt., and ecommerce,D > and outside of that anyone else who is smart enough to want it ...I > and benchmarks are run by more test houses than hp, and time will tell,vF > no one will be able to fudge anything ... but intel cannot afford toH > pick the wrong processor ... remember, the alpha goodies don't kick inG > till chivano, and those may very well suprise you if epic doesn't cutaE > the mustard ... intel will not fail, one way or the other they willcG > produce a viable chip, either an alpha enhanced epic one or switch to>G > an ev8-ev9 design, one way or the other, they will produce, they have  > to!h   Bob,  G If you read my comments, nowhere would you find a statement that HP hast@ stopped selling VMS. What HP has stated is that all new businessJ opportunities will be steered to HP-UX, except in the case of the existingL customer base, which will be offered VMS if they ask for it. In other words,I HP isn't going out of its way to solicit VMS business from new customers.dL The URL's to the HP documents containing statements to that effect have been2 posted in c.o.v. numerous times since May 2, 2002.  J My point as to the benchmark tests is that most code for the benchmarks isL carefully crafted by expert programmers, not your average programmer workingL for a Fortune 1000 company, government department, or a 100-person business.I A more useful benchmark for most customers is what performance an averageaF programmer could extract from a machine, not what the top 0.01% of allL programmers could do. Not all programmers deal with word-alignment properly,J not all programmers know how to unroll a loop properly...in fact, probablyJ more don't than do. Most companies don't go through the exercise of formalJ code reviews, so they typically would not find these and other horrors. InK other words, they have typical code. What's so bad about having a benchmark L written by typical programmers? It would surely stop the fingerpointing backH to the hardware manufacturers when a customer finds out that they need aH system that costs an extra $1MM to get the customer's typical production& code working to the speed they expect.  J As to what Intel will or will not incorporate of Alpha in future products,- that's merely WAG'ing  (wild-assed guessing).    ------------------------------    Date: 22 Jul 2002 11:57:55 -0600- From: koehler@encompasserve.org (Bob Koehler)t Subject: Laramie, anyone?i3 Message-ID: <PcHhz2v0XUDT@eisner.encompasserve.org>   H    One the way home from a Las Vegas DECUS, I once ran into someone fromE    the University of Wyoming (Laramie) on the plane.  Is there anyoneu,    here still conneted to/located at U. Wyo?   ------------------------------  % Date: Mon, 22 Jul 2002 13:55:12 -0400w; From: "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov>e( Subject: OpenVMS webring:  Where's Mike?K Message-ID: <BA52530E3149734A9BAABDBBFA808E4901479EF4@rlghncst964.usps.gov>m  ' Does anybody have information as to thec& whereabouts of Mike Taylor, who is the" ringmaster of record for the ring?  ' I've sent email to tetranet and gtw.neto$ to see if they have any information.   ========================  William W. Webb / DSSC/RLM, USPS OpenVMS Support Services& 4924 Green Road Raleigh, NC 27616-2800 919.874.3043   ------------------------------  % Date: Mon, 22 Jul 2002 08:29:40 -0700b' From: David Mathog <mathog@caltech.edu> 8 Subject: Re: Powerstorm 300/350 cards and OpenGL on OVMS+ Message-ID: <3D3C24E4.EE5C3DC6@caltech.edu>r   Robert Deininger wrote:l > F > In article <f_FZ8.41$Ry3.811158@news.cpqcorp.net>, "Fred Kleinsorge"& > <kleinsorge@star.zko.dec.com> wrote: > M > >The 7500 is the Rv200, which we will be shipping when Marvel FRS's.  It is J > >an AGP card, so that isn't very interesting to most people with currentJ > >Alphas.  ATI is building a PCI card (well, they already had one, but itI > >doesn't work in hot swap slots) and we won't see them until this fall.  > I > One model of the ES45 has an AGP slot as well.  I think that's the onlye5 > opportunity for VMS users to use AGP at the moment.h  ? Well, maybe.  That depends on whether or not there is currentlyn an AGP driver in. VMS that supports AGP on that machine.  Fred?    Regards,   David Mathog mathog@caltech.edu   ------------------------------  % Date: Mon, 22 Jul 2002 08:33:36 -0400i0 From: "Alan Boyles" <alan.boyles@mindspring.com>! Subject: SMISERVER Bytlim problem0/ Message-ID: <ujnugbaio34hd5@corp.supernews.com>u  F I have a 2-GS80's running 7.2-1H1 that is experiencing an intermittentK problem with the SMISERVER.  When I try to issue $MOUNT/SYSTEM/SHAD  inside_K SYSMAN I am getting an EXCEEDED BYTLIM  error message.  A command procedureiI in batch on one node goes into SYSMAN and sets the environment on the 2ndlK node and then issues the $MOUNT command to bring a third disk into a shadowsJ set.  There are 13 shadow sets which we issue this command for inside thisL procedure, and the current MAX COPY is set to 6.  After about the 4th or 5thH command issuance we get the EXCEEDED BYTLIM error and then it will occurL every other issue of the MOUNT command.  We have raised BYTLM and JTQUOTA inG both the process used for SYSMAN and the PQL_  parameters in SYSGEN andyI restarted the SMISERVER but have seen it happen again .  Anyone else seena$ this kind of problem with SMISERVER.   Thanks,o   Alan   ------------------------------  % Date: Mon, 22 Jul 2002 09:07:33 +0200c1 From: Franz-Josef Fornefeld <jo.fornefeld@gmx.de>  Subject: Re: SMTP question( Message-ID: <ahgi0m.162.1@jo.dyndns.org>   Tom Linden wrote:   9 > tcpip5.1 on Alpha 7.3 pop client Outlook on W2K server.n >s9 > Every now and then (couple of months) I can't send maila( > because of timeout to the smtp server.  % Who times out? Outlook or the server?>  & Does the IP connection still exist[1]?   >  Still continue receiveingG > mail, however.  The problem is (temporarily) resolved by stopping andi > starting SMTP. >l > Anybody else experience this?o  & Not in this configuration but general:H It is a known problem that Outlook terminates connections to its serversG without QUITting. Servers handle this normally analog to QUIT (which isN the correct behavior).  ) [1] check the IP ports on your W2K box[2]b# [2] I suggest using Active Ports[3]r [3] http://www.ntutility.com/t   ------------------------------  % Date: Mon, 22 Jul 2002 09:00:21 -0700s# From: "Tom Linden" <tom@kednos.com>d Subject: RE: SMTP question9 Message-ID: <CIEJLCMNHNNDLLOOGNJIGENLFGAA.tom@kednos.com>k  1 I think there was a fatal error on the SMTP side,r1 because TCPIP ANAL MAIL/REPAIR fixed the problem.t   >-----Original Message-----h9 >From: Franz-Josef Fornefeld [mailto:jo.fornefeld@gmx.de]n% >Sent: Monday, July 22, 2002 12:08 AMu >To: Info-VAX@Mvb.Saic.Com >Subject: Re: SMTP questione >  >o >Tom Linden wrote: > : >> tcpip5.1 on Alpha 7.3 pop client Outlook on W2K server. >>: >> Every now and then (couple of months) I can't send mail) >> because of timeout to the smtp server.f >s& >Who times out? Outlook or the server? > ' >Does the IP connection still exist[1]?n >o >>  Still continue receiveingHH >> mail, however.  The problem is (temporarily) resolved by stopping and >> starting SMTP.i >>  >> Anybody else experience this? > ' >Not in this configuration but general:hI >It is a known problem that Outlook terminates connections to its servers H >without QUITting. Servers handle this normally analog to QUIT (which is >the correct behavior).b >r* >[1] check the IP ports on your W2K box[2]$ >[2] I suggest using Active Ports[3] >[3] http://www.ntutility.com/ >r >---' >Incoming mail is certified Virus Free.a; >Checked by AVG anti-virus system (http://www.grisoft.com).a@ >Version: 6.0.373 / Virus Database: 208 - Release Date: 7/1/2002 >  ---f& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.377 / Virus Database: 211 - Release Date: 7/15/2002   ------------------------------  % Date: Mon, 22 Jul 2002 19:29:17 +0200T- From: Didier Morandi <Didier.Morandi@Free.fr> $ Subject: Re: Terminal input from DCL' Message-ID: <3D3C40ED.3AF8DF00@Free.fr>.  O This raises a different issue. DCL is not a high-level programming language. ItoP cannot perform ASTs (except the read/timeout/err), it cannot perform QIO-nowait,H not computer GOTOs. If this automatic validation field feature is reallyL requested, yes it has to be programmed with a 3rd level language (or FMS :-)   D.   "David J. Dachtera" wrote: ../..aE > My point was that without being able to trap individual keystrokes,oI > there's no practical way to limit the count of characters entered so ase8 > to limit entry into a field to a predetermined length. > B > You can do that in BASIC with INKEY$, but DCL has no equivalent.   ------------------------------   Date: 22 JUL 2002 16:09:15 GMT4 From: karcher@thuria.waisman.wisc.edu (Carl Karcher)1 Subject: Re: Upgrading from 7.2-1 and latest ECOsa6 Message-ID: <22JUL02.16091596@thuria.waisman.wisc.edu>  ? In a previous article, "Main, Kerry" <Kerry.Main@hp.com> wrote:-  J ->Is the issue you are describing different than what was described in the ->V7.3 cover letter?  F Yes. I can confirm the behavior that Lyndon mentions. I justed appliedE VMS721_SYS-V1200 to a 7.2-1 system and the following entries exist ine& sys$update:vms$remedial_old_files.txt:  @ [SYSEXE]APB.EXE                                 VMS721_SYS-V1200@ [SYSEXE]DEBUG_APB.EXE                           VMS721_SYS-V1200  H They SHOULD be suffixed with _OLD (otherwise the 7.3 upgrade will deleteE them). This is exactly the problem that VMS721_RENAME_OLD-V0100 fixedm8 but with different patch kits affecting different files.  G A big thanks to Lyndon for pointing this out (it would have burned me).m  / Lyndon Bartels <lbartels@pressenter.com> wrote:t  G ->Came across a interesting little issue when trying to upgrade to v7.3 H ->last week. I had installed the very latest ECOs to v7.2-1 for a coupleF ->reasons. Then, a couple weeks later, I upgraded to v7.3. The upgrade; ->went "According to Hoyle" At least it looked like it did.- -> -G ->The first time the computer tried to boot, it failed. Turns out, that I ->the latest "sys" eco, and "amacro" eco, improperly write entries in the.H ->sys$update:vms$remedial_old_files.txt file.  Namely an entry for "apb"I ->and "apb_debug" (from "sys" eco) and "starlet.mlb" (from "amacro" eco).n -> aD ->Any subsequent upgrade would "remove" the ECOs and hence the files ->mentioned in this file.=20 -> aG ->This has happened before and the "rename old" eco fixed those issues.m: ->This is the same issue. But with new and different ECOs. -> oI ->I logged a call, and got the system up and running on v7.3. The guy whop; ->handled the call says it's been elevated to engineering..t -> hF ->We'll probably see a new "rename old" ECO, or some such pretty soon.   --G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madison 9 --                      karcher.nospam@waisman.wisc.edu  t   ------------------------------  % Date: Mon, 22 Jul 2002 10:01:16 -0400a5 From: "Sue Skonetski" <susan.skonetski@hp.nospam.com>  Subject: Re: VMS commitment * Message-ID: <ahh36n$bjp$1@web1.cup.hp.com>   Folks,  L Thank you for thinking of me.  Yes I have seen this message and have been inH contact with Ed and hopefully sent him what he needed.  Ed's meeting was; moved to today (Monday) so hopefully we will get an update.r  
 Warm Regards,  Suec  ? "Lucas, Edward A (SAIC)" <Edward.Lucas@bp.com> wrote in messaget9 news:EF1DC894691AD5118AF000508BB85FDE034CC70F@AMCLVX11...e > Hello everyone,s >iH > The stupid company I work for has decided that VMS is going away.  The. > merger with HP signals the way for UNIX. : (G > This week I need to start a project called "VAX Elimination project".  > (dingle berries) >nB > Can someone please send me the link where HP commits to OpenVMS.J > I need some fire power.  These dingle berries deciding that VMS is goingH > away and the costs are too high need to find another job.  I have beenK > working with VMS for over 16 years and I am so sick and tired of everyonesK > blaming the VAX's/Alpha's for every little problem, and each time we have  atG > problem its always another platform. ( You think I would become numb)  > G > Please help. We have the opening meeting Thursday, and at that time I  wouldn+ > like to present all information possible.tB > These dingle berries feel they well be able to eliminate the VMS environmente4 > for around 2.5 mil. (daaaaaa, maybe around 9 mil.) >m1 > To give an example of how stupid management is;nJ > They feel having a DR site is not necessary.  So I have purchased backupK > systems.  My recommendation was to put the systems in Ohio or anywhere on J > the east cost.  The felt it was best to put them on another floor in theK > same fricken building.  (The Building is in Tornado alley) They STATE theiK > building will holdup during a tornado. (Ok, lets see the fricken buildingt > stand during a level 4 or 5) >p >  > Edward A. Lucasl" > Sr. VAX/VMS System Administrator > SAIC > Phone:  (216) 525-7492 > Email:   Lucaea@bp.com >o >a   ------------------------------  % Date: Mon, 22 Jul 2002 16:44:03 +0800e) From: Kiasu Surfer <verykiasu@hotpop.com>88 Subject: VMS Mail problems Internal/External receipient.8 Message-ID: <gehnjug3kuhtvup7kcnh5f8eqe607be9kc@4ax.com>  1 Can someone help me with this mail setup problem?b     OpenVMS 7.1 with UCX 4.1 ECO 9 ============================== UCX>show service smtph  uL Service             Port  Proto    Process          Address            State   O SMTP                  25  TCP      UCX$SMTP         0.0.0.0             Enableds   UCX> show conf smtpt SMTP ConfigurationI                                                                   OptionspL Initial interval:   0 00:30:00.00       Address_max:    16       NOEIGHT_BITF Retry interval:     0 01:00:00.00       Hop_count_max:  16       RELAYJ Maximum interval:   3 00:00:00.00                                NOHEADERSG Timeout             Initial       Mail    Receipt       Data  TerminatetG   Send:                   5          5          5          3         10i   Receive:                5 ) Alternate gateway:  exchange.mydomain.com- General gateway:    not defined2 Substitute domain:  not defined & Zone:               dec02.mydomain.com Postmaster:         UCX_SMTP? Log file:           SYS$SPECIFIC:[UCX_SMTP]UCX$SMTP_LOGFILE.LOGn =======<  H With this setting, I am send mails to our internal Exchange mail users. J But not external email internet users. The remote mail server will reject ; our mails due to unreachable hosts (machine not reachable)..6 Only our domain is a qualified registered domain name.  K With the setting of SET CONFIG SMTP /SUBSTITUTE=(hidden,name=mydomain.com),eK I am able to send mails to external internet parties, but no mails send to  D internal exchange mail users. I had a hard time debugging the cause.  , Can someone help me to resolve this problem?  
 Thank you.       Regards,  	 Kevin Laii   **************** ** SPAM BLOCK ** ****************? REPLACE "verykiasu" with "iamverykiasu" at verykiasu@hotpop.com    ------------------------------  % Date: Mon, 22 Jul 2002 11:34:50 -0400b From: norm.raphael@metso.comL Subject: Re: WATCHER (was Re: Command Procedure to logoff inactive users...)? Message-ID: <OF23636A4F.E9B46A52-ON85256BFE.0054A3A4@metso.com>C  + / Please respond to mckinneyj@cpva.saic.comd   / To:  Info-VAX@Mvb.Saic.Com / cc:bE / Subject:  Re: WATCHER (was Re: Command Procedure to logoff inactiver        users...) /c /uJ / > ($FORCEX does not guarantee that a success status has forced an exit.) /nJ / The success status returned by $forcex indicates the successful queueingI / of the AST requesting that the process exit the running image. There is:I / no guarantee that the AST will/can be delivered (or that the image wille / exit). / 6 That's what I said (although without the explanation). /l      + / Please respond to mckinneyj@cpva.saic.com@   / To:  Info-VAX@Mvb.Saic.Com / cc:sE / Subject:  Re: WATCHER (was Re: Command Procedure to logoff inactives        users...) /h /hA / In article <OF315413A4.46722D0B-ON85256BF4.0056813F@metso.com>,r  / norm.raphael@metso.com writes: / > ' / > I use WATCHER and it calls $FORCEX.aI / > I have found that one of my applications requires that I call $FORCEX " / > twice to get processes to DCL.J / > Recently I tried this on the whole user population of that application =p / > and.J / > found that one of the processes was still not at DCL after this double =y / > hit.3 / > Is there any way to figure out what's going on. J / > ($FORCEX does not guarantee that a success status has forced an exit.)	 / > -Norma /hF / Could it be that the target process that requires the double $forcexI / is executing a command procedure where once the image exits a 2nd imagen / is activated?f  ( That is certainly possible, even likely.  < / Or maybe a lengthy exit handler in the image is in invoked% / in response to the initial $forcex?h    Less likely, but still possible.  , / Do you have the application's source code?    & No, nor any possibility of getting it.   /oG / What does SDA say about the process after the delivery of the initial ) / $forcex? Is there a force exit pending?e /   I I am not sufficently proficent to check this.  I suppose I could set up ahH process and issue a single $forcex, the go look with SDA if I knew whereF to look and what to look at/for.  Is this an easy sequence to detail a procedure for?  I In the case where two doesn't do it, would three be a reasonable thing totJ try?  Or maybe a second pass with two at the processes remaining after theK first pass, if any?  Or a longer wait between the first and second $forcex?m   -Normw   ------------------------------  % Date: Mon, 22 Jul 2002 09:59:02 -0400 - From: "Peter Weaver" <peter.weaver@stelco.ca>o* Subject: Re: [Hobbyist] User account setup5 Message-ID: <ahh337$sna04$1@ID-141708.news.dfncis.de>   : "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message- news:LmSh0u5l+Y6E@eisner.encompasserve.org... 4 > In article <3D391411.6000309@spammotel.com>, Alder$ <PGDEHMKOKIMD@spammotel.com> writes: >...L > > EUREKA!  The user account logged in!  So now we revoke privileges one by0 > > one?  Or is there a more dignified approach? >oJ > Enable auditing for "successful use of privilege", which also might haveK > to be done separately for file access (I forget).  Most likely it is filee0 > access, and auditing will tell you which file. > ...l  6 Right, these commands should do what Larry recommends;  4 $ ! Using your SYSTEM (or another privileged account $ reply/enableL $ set aud/enab=(access=(readall:all,grpprv:all,sysprv:all,bypass:all))/alarmJ $ ! log in using your privileged account on another terminal (or using SETL HOST) and watch for messages on the terminal you get on the terminal you did the REPLY/ENABLE onaE $ ! Once you have logged in and looked at the messages then enter theC commands $ setsG aud/disab=(access=(readall:all,grpprv:all,sysprv:all,bypass:all))/alarm3   $ reply/disable8  L If these do not tell you what is going on, then let us know and someone willJ post the commands to set an audit alarm for failure, then you will have toJ modify you account to remove all the privileges you gave it before and tryD again. But I'm hoping this will tell you the file that has the wrong/ protection (or has been installed incorrectly).N  ? BTW: If anyone else gives you the advice of either rebooting oreF re-installing VMS, please don't. It is so seldom that VMS breaks it is@ interesting to have some practice trouble-shooting a problem. :)   -- Peter WeaverL Opinions are my own, and do not reflect the opinions of my employer, nor theK company that it sub-contracts to, nor the company that it sub-contracts to.c   ------------------------------  % Date: Mon, 22 Jul 2002 19:26:05 +0200r- From: Didier Morandi <Didier.Morandi@Free.fr>'* Subject: Re: [Hobbyist] User account setup' Message-ID: <3D3C402C.5ED51DAC@Free.fr>d  W http://www.openvms.compaq.com:8000/73final/6346/6346pro_021.html#attaching_security_aceh   "David J. Dachtera" wrote: > D > *SIGH* Someone else suggested setting an Audit ACE (Access Control@ > Entry) in the ACL (Access Control List) on both of DCL.EXE andH > DCLTABLES.EXE. That's kind of heavy stuff for a newbie, but maybe that, > resulting data would yield a further clue. > G > There's not enough info. in HELP to determine the exact the syntax of-I > the appropriate ACE, but if I can review the on-line doc.'s and come upoC > with it, I'll post it later. Perhaps someone who does such things : > "everyday" (Larry K.?) can post it from memory. Dunno...  	 Opzatelps    D.   ------------------------------   End of INFO-VAX 2002.401 ************************