1 INFO-VAX	Sat, 19 Jun 2004	Volume 2004 : Issue 338       Contents:0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.0 Re: COPY VMS Save Sets to CD on WINDOWS machine.- DECnet-Plus V7.3 ECO 4 for VAX breaks routing  Re: delete directory recursive. How to detect user idle time for DECUS FINGER?8 Re: Secure Web Server Version 2.0 available for download8 Re: Secure Web Server Version 2.0 available for downloadP Re: sftp and non-stream_lf files? (was: Re: Secure Web Server Version 2.0 availa$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file???$ Re: Split I/Os to contiguous file??? Re: SWS-2.0 (Apache) & CGI ??? Re: SWS-2.0 (Apache) & CGI ??? Re: SWS-2.0 (Apache) & CGI ???& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.& Re: VMS Gets Short Shrift at HP World.2 Re: Why does idle telnet session get disconnected?2 Re: Why does idle telnet session get disconnected?  F ----------------------------------------------------------------------  + Date: Fri, 18 Jun 2004 19:44:47 +0000 (UTC) & From: News Reader <nomail@noemail.vms>9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. $ Message-ID: <cavgne$pmn$1@online.de>  E On a related note about using the backup command to backup the system ! disk to multiple smaller disks...   D There's a section covering Backing Up the System Disk to a Disk, butD it assumes the recipient disk is large enough on its own.  There's aC section on Backing Up the System Disk to Tape and the example shows A the splitting across several tapes by specifiying a single output @ device and pausing as you change the media in the output device:  F $ BACKUP/IMAGE/VERIFY DUA1: MUA0:DEC_31_BACKUP.BCK/REWIND/LABEL=WKY101    But when I try it, this happens:  K $ backup /image /verify dka0: dad10:jne_18_backup.bck /rewind /label:backup D %BACKUP-F-IMGFILSPE, /IMAGE specification must have only device name   Also tried:   M $ backup /image /verify dka0: dad10:jne_18_backup.bck /save_set /label:backup 8 %BACKUP-F-NOHOMEBLK, no valid home block found on DAD10:  E It grinded for an hour before producing that error.  It never paused.   H If you can only specify one output device, then what I need is for it toG pause so as to give me an opportunity to change the media in the output B device.  Alternatively, what I need is to specify a list of outputE devices so it can write to them in turn to produce an image backup of % one disk split across several others.    ------------------------------    Date: 18 Jun 2004 15:41:27 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) 9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. 3 Message-ID: <IWlXTHax0eT5@eisner.encompasserve.org>   M In article <cavgne$pmn$1@online.de>, News Reader <nomail@noemail.vms> writes: G > On a related note about using the backup command to backup the system # > disk to multiple smaller disks...  > F > There's a section covering Backing Up the System Disk to a Disk, butF > it assumes the recipient disk is large enough on its own.  There's aE > section on Backing Up the System Disk to Tape and the example shows C > the splitting across several tapes by specifiying a single output B > device and pausing as you change the media in the output device: > H > $ BACKUP/IMAGE/VERIFY DUA1: MUA0:DEC_31_BACKUP.BCK/REWIND/LABEL=WKY101  E To do the corresponding thing to disk is called the "sequential disk" F feature of VMS backup, introduced sometime _after_ VMS V4.0 but likely prior to VMS 5.5.   " > But when I try it, this happens: > M > $ backup /image /verify dka0: dad10:jne_18_backup.bck /rewind /label:backup F > %BACKUP-F-IMGFILSPE, /IMAGE specification must have only device name  D Presuming that is an Infoserver, you cannot count on the disks thereA to behave in all respects like a local VMS disk.  Even if it is a ? tape (and I have never used Infoserver tapes) the same applies.   
 > Also tried:  > O > $ backup /image /verify dka0: dad10:jne_18_backup.bck /save_set /label:backup : > %BACKUP-F-NOHOMEBLK, no valid home block found on DAD10:  I My recollection is that Sequential Disk backups involve a /FOREIGN mount.    ------------------------------  % Date: Fri, 18 Jun 2004 16:57:50 -0400 < From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. + Message-ID: <2jh3afF10bvh9U1@uni-berlin.de>    News Reader wrote:G > On a related note about using the backup command to backup the system # > disk to multiple smaller disks...  >...A > $ backup /image /verify dka0: dad10:jne_18_backup.bck /save_set H > /label:backup %BACKUP-F-NOHOMEBLK, no valid home block found on DAD10: >...  F I do not have a system here I can test this on right now, but it looksG like you mounted DAD10 as a Files-11 device, if you change the mount to   MOUNT/FOR that you should be OK.   --   Peter Weaver Weaver Consulting Services Inc.  Canadian VAR for CHARON-VAX  www.weaverconsulting.ca    ------------------------------    Date: 18 Jun 2004 16:07:59 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. 3 Message-ID: <LmeC$I2jkBfP@eisner.encompasserve.org>   M In article <cavgne$pmn$1@online.de>, News Reader <nomail@noemail.vms> writes:  > H > $ BACKUP/IMAGE/VERIFY DUA1: MUA0:DEC_31_BACKUP.BCK/REWIND/LABEL=WKY101 > " > But when I try it, this happens: > M > $ backup /image /verify dka0: dad10:jne_18_backup.bck /rewind /label:backup F > %BACKUP-F-IMGFILSPE, /IMAGE specification must have only device name  J    IIRC to backup to a disk-based saveset the output disk must be labeled !    and mounted as a labeled disk.      J > If you can only specify one output device, then what I need is for it toI > pause so as to give me an opportunity to change the media in the output D > device.  Alternatively, what I need is to specify a list of outputG > devices so it can write to them in turn to produce an image backup of ' > one disk split across several others.   E    You can specify multiple devices in the standard DCL list format.  .    I've done this with tape drives many times:  F    $backup dka0:/image/verify mua0:stuff.bck/rewind/density=1600,mua1:   ------------------------------  + Date: Fri, 18 Jun 2004 22:11:45 +0000 (UTC) & From: News Reader <nomail@noemail.vms>9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. $ Message-ID: <cavpb1$htt$1@online.de>  M On 18 Jun 2004 15:41:27 -0600, Kilgallen@SpamCop.net (Larry Kilgallen) wrote:   N >In article <cavgne$pmn$1@online.de>, News Reader <nomail@noemail.vms> writes:H >> On a related note about using the backup command to backup the system$ >> disk to multiple smaller disks... >>F >To do the corresponding thing to disk is called the "sequential disk"G >feature of VMS backup, introduced sometime _after_ VMS V4.0 but likely  >prior to VMS 5.5.  D Ah thanks.  Saw the term somewhere but it didn't go on to define it.  N >> $ backup /image /verify dka0: dad10:jne_18_backup.bck /rewind /label:backupG >> %BACKUP-F-IMGFILSPE, /IMAGE specification must have only device name  > E >Presuming that is an Infoserver, you cannot count on the disks there 1 >to behave in all respects like a local VMS disk.   E No, but VMS does know the size of the media to stop when it fills it.    >Even if it is a@ >tape (and I have never used Infoserver tapes) the same applies.  F Whoever got a hold of the tape function access disk is keeping schtum;6 probably planning to take it with them to their grave.  P >> $ backup /image /verify dka0: dad10:jne_18_backup.bck /save_set /label:backup; >> %BACKUP-F-NOHOMEBLK, no valid home block found on DAD10:  > J >My recollection is that Sequential Disk backups involve a /FOREIGN mount.  A Thanks, though I did actually do a mount /foreign /system /write.   I I thought perhaps my mistake was specifying a label in the backup command D when the drive was mounted foreign.  Also, I thought that perhaps myF mistake was the /rewind.  I think it should have been /init used, as I$ had done, in conjunction with /save.  G $ backup /image dka0: dad10:jne_18_backup.bck /init /label:backup /save   G That merrily trundled away and did pause.  Without off-lining the drive 0 I copied it to CD then continued, but it failed:  * %BACKUP-I-OPREPLY, operator reply is "yes"9 %BACKUP-I-NEXTVOL, volume 2 mounted - operation continues L %BACKUP-F-OPENOUT, error opening DAD10:[000000]JNE_18_BACKUP.BCK;1 as output' -SYSTEM-W-NOTQUEUED, request not queued    ------------------------------  + Date: Fri, 18 Jun 2004 22:17:53 +0000 (UTC) & From: News Reader <nomail@noemail.vms>9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. $ Message-ID: <cavpmh$htt$2@online.de>  9 On Fri, 18 Jun 2004 16:57:50 -0400, "Peter Weaver" wrote:  >News Reader wrote: H >> On a related note about using the backup command to backup the system$ >> disk to multiple smaller disks... >>B >> $ backup /image /verify dka0: dad10:jne_18_backup.bck /save_setI >> /label:backup %BACKUP-F-NOHOMEBLK, no valid home block found on DAD10:  >>G >I do not have a system here I can test this on right now, but it looks H >like you mounted DAD10 as a Files-11 device, if you change the mount to! >MOUNT/FOR that you should be OK.   E Well thanks for that and indeed I did make that mistake initially and E promptly unmounted it and started again.  The error you see above did B actually follow a mount /foreign /system /write.  So there must be something else I'm doing wrong.    ------------------------------  + Date: Fri, 18 Jun 2004 22:39:13 +0000 (UTC) & From: News Reader <nomail@noemail.vms>9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. $ Message-ID: <cavquh$lud$1@online.de>  2 On 18 Jun 2004 16:07:59 -0600 (Bob Koehler) wrote:N >In article <cavgne$pmn$1@online.de>, News Reader <nomail@noemail.vms> writes: >>N >> $ backup /image /verify dka0: dad10:jne_18_backup.bck /rewind /label:backupG >> %BACKUP-F-IMGFILSPE, /IMAGE specification must have only device name  > J >   IIRC to backup to a disk-based saveset the output disk must be labeled" >   and mounted as a labeled disk.  H Hmm, but wouldn't that mean that I'd have to format the disk in order toD label it, and then wouldn't that mean re-mounting it as non-foreign? Its all jolly confusing to me.  1 Would your conjecture account for this behaviour:   G $ backup /image dka0: dad10:jne_18_backup.bck /init /label:backup /save  ...  [ much trundling ] ...  Request 1, from user SYSTEM ; %BACKUP-I-READYWRITE, mount volume 2 on _DAD10: for writing  Enter "YES" when ready:  ... 3 [ copy the drive to cd, now ready to overwrite it ]  ...  $ reply /to:1 "yes"  yes 6 22:37:23.85, request 1 was completed by operator OPA0::  Saveset volume:2, saveset block:16606 (32256 byte blocks)* %BACKUP-I-OPREPLY, operator reply is "yes"9 %BACKUP-I-NEXTVOL, volume 2 mounted - operation continues L %BACKUP-F-OPENOUT, error opening DAD10:[000000]JNE_18_BACKUP.BCK;1 as output' -SYSTEM-W-NOTQUEUED, request not queued    ------------------------------  % Date: Fri, 18 Jun 2004 22:15:11 -0700  From: Z <z@no.spam> 9 Subject: Re: COPY VMS Save Sets to CD on WINDOWS machine. 0 Message-ID: <10d7iu9p86hdd6b@corp.supernews.com>  
 Tim wrote:H > Our ALPHA server systems do not have write-able CD drives. I attemptedG > to put VMS save sets on a RW_CD by ftp'ing the save-sets to a WINDOWS G > Machine,and then transfering the save sets from the DOS Hard Drive to  > the CD-RW disk.  > E > Unfortunately, the ALPHA machines are unable to read save sets from  > the MS created disks.  > E > Any suggestions on workarounds or freeware is greatly apprectiated.   & $! RESET_BACKUP_SAVESET_ATTRIBUTES.COM $!1 $! P1  is the specification of the BACKUP saveset  $!5 $! This procedure resets the record format and record 4 $! length attributes of a BACKUP saveset -- savesets0 $! can get "broken" during certain sorts of file6 $! transfers -- such as FTP.  This procedure reads the4 $! (undocumented) saveset record attributes directly $! out of the target file. $!4 $! First render the saveset readable, and implicitly $! check that the file exists. $! $ Set File -=      /Attributes=(RFM:FIX,MRS:512,LRL=512,ORG=SEQ,RAT=NONE) - 	      'p1'  $ ( $ Open/Error=whoops/Read BckSaveset 'p1'0 $ Read/Error=whoops/End=whoops BckSaveset Record $ Close/Nolog BckSaveset $ / $! Find the blocksize from within the record...  $  $ BlockSize = 0  $ BBH_L_BLOCKSIZE = %x28*81 $ BlockSize = F$CVUI(BBH_L_BLOCKSIZE, 32, Record) 2 $ If BlockSize .lt. 2048 .or. BlockSize .gt. 65535 $ Then, $   Write sys$output "Unexpected block size" $   Goto whoops  $ Else4 $   Set File /Attributes=(RFM:FIX,LRL='BlockSize', -#         MRS='BlockSize',RAT=none) -          'p1' $ endif  $ exit $WHOOPS: $ Write sys$output "Error" $ exit   ------------------------------  % Date: Sat, 19 Jun 2004 01:30:23 -0400   From: John Santos <JOHN@egh.com>6 Subject: DECnet-Plus V7.3 ECO 4 for VAX breaks routing4 Message-ID: <1040619011517.404C-100000@Ives.egh.com>  E Installed DNVOSIECO04 on two VAX systems.  On the 1st it worked fine. E On the 2nd, it crashed repeatedly while booting as soon as it enabled < the second circuit circuit (near as I can narrow it down) in$ sys$manager:net$routing_startup.ncl.  6 The 1st system was an end node (ROUTING TYPE ENDNODE).  4 The 2nd system was a router (ROUTING TYPE L2ROUTER).  6 The crash is in sys$loadable_images:net$routing_is.exe  = I replaced it with the version from ECO02 (the previous ECO),  and it now works again.   > The thing about this crash is it looks a lot like one in V7.1,4 that we reported and received a corrected version of; net$routing_is.exe to fix.  The corrected version turned up ? in ECO 1 for V7.1, and routing has worked fine in V7.2 and V7.3 8 up til now, but I'm wondering if they regressed the fix?  = Anyway, beware of this ECO if you are using DECnet routing on  VAXes.  ( I'll be submitting this to HP on Monday.   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------    Date: 18 Jun 2004 15:18:18 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)' Subject: Re: delete directory recursive = Message-ID: <b096a4ee.0406181418.56af7b37@posting.google.com>   v koehler@eisner.nospam.encompasserve.org (Bob Koehler) wrote in message news:<G2nnbpuq1MaQ@eisner.encompasserve.org>...p > In article <b096a4ee.0406171804.24515e74@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes: > > F > > This can also fail if there are multiple versions of the same .DIR3 > > file, in which case repeating ,;* would fix it.  > @ >    No.  Legitimate directory files must be version ;1.  If youC >    have multiple versions the contents of the other versions will 6 >    never be deleted until you correct the situation.  D Oops, you're right! Well, at least the first half of my sentence was$ right. Thanks for pointing that out.   ------------------------------  + Date: Fri, 18 Jun 2004 19:55:58 +0000 (UTC) * From: bleau@UMTOF.UMD.EDU (Lawrence Bleau)7 Subject: How to detect user idle time for DECUS FINGER? 0 Message-ID: <cavhce$k3g$2@grapevine.wam.umd.edu>  F This could be easy, or it could be something of an internals question.  J I'm running OpenVMS 7.3-2 AXP and an old version of DECUS FINGER, V51.1.31I with a few local mods.  I just got it working okay; there were a few bugs G in it.  Something that still doesn't work is the /IDLE option, which is K supposed to report a user's idle time.  Does anyone have a version of DECUS B FINGER that works on an Alpha and correctly reports the idle time?  D I poked around the code some, and FINGER determines the idle time byE calling the routine TERM_INFO, which is in TERMINAL.MAR (yes, a Macro I routine).  The arguments are a character string that is the terminal name I and an I*4 variable, whose value is returned.  TERM_INFO goes into kernel H mode, locates the UCB for the device name it was given, verifies it is aK terminal device, then retrieves the field UCB$L_DUETIM.  This, according to F an old I&D book, is the system uptime at which the current i/o requestH expires.  If this value is nonzero, TERM_INFO then subtracts from it theJ quantity in EXE$GL_ABSTIM, which is the current system uptime.  Here's the code fragment:  4 25$:    MOVL    UCB$L_DUETIM(R1),R0     ; get duetimE         BEQL    26$                     ; if zero, save it as zero... 5         SUBL3   R0,G^EXE$GL_ABSTIM,R0   ; save duetim  26$:    MOVL    R0,(R10)+   E It seems to me that the difference it gets is the amount of time left D before the i/o expires, which isn't very helpful, unless I'm missingH something important here.  Anyways, from Finger's behavior, I infer thatM the value returned is always zero.  Finger - actually, the routine Get_Idle - G returns all blanks if the returned idle time is zero.  Every time I run H Finger on my system, the Idle column is always blank, no matter how long the user has been "idle".   & Any hints on this?  Thanks in advance.   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.edu    ------------------------------    Date: 18 Jun 2004 14:08:06 -0700" From: horn@shsu.edu (James T Horn)A Subject: Re: Secure Web Server Version 2.0 available for download = Message-ID: <843706dc.0406181308.10a9a6e6@posting.google.com>   B In all this discussion, we are finding that using sftp to transferE files between systems is having problems with the non-streamlf files. E We are having to convert files created using cobol to streamlf before ! we us sftp to transfer the files.   A Anyone having the same kind of problem with sftp and non-streamlf  files?   ------------------------------  % Date: Fri, 18 Jun 2004 19:19:45 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> A Subject: Re: Secure Web Server Version 2.0 available for download : Message-ID: <qMKAc.39736$nY.1291816@news20.bellglobal.com>  / "James T Horn" <horn@shsu.edu> wrote in message 7 news:843706dc.0406181308.10a9a6e6@posting.google.com... D > In all this discussion, we are finding that using sftp to transferG > files between systems is having problems with the non-streamlf files. G > We are having to convert files created using cobol to streamlf before # > we us sftp to transfer the files.  > C > Anyone having the same kind of problem with sftp and non-streamlf  > files?  J I'm no expert but I've only ever encountered three types of problems using any kind of FTP:  H 1. newbees prefer to use BINARY transfers for everything. As I said in aH previous posting, when sending TEXT you must use ASCII mode so that textJ stored a certain way on one system is converted to the expected format forG the other system. (Some systems end a line of text with <cr>, some with L <lf>, some with <cr><lf>, one with <ctrl-z>, and OpenVMS with nothing at allE because the lines are usually stored in RMS as records that appear to G resemble Pascal strings). BTW, don't take my word for any of this, just  check the RFCs.   J 2. some systems treat BINARY and IMAGE as synonyms while others treat themH differently. I'm currently using TCPware's FTP to send binary files to aI UNIX system. I must use IMAGE to send files to UNIX while the UNIX system H must send them back using BINARY (not all UNIX systems support the IMAGEL command). BTW, TCPware's help states that BINARY mode should only be used to* send certain files like ".OBJ" and ".STB".  J 3. FTPing a file with a non standard block size (like a VMS save set) willH almost always change the block size to 512 bytes. So if you need to do aJ restore, you either FTP or copy the saveset back to OpenVMS and then run aL little tool to correct the block size. BTW, this is officially known as "FTP: Induced Save Set Corruption". Click here for more details:6 http://h71000.www7.hp.com/openvms/freeware/backup.html    
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  # Date: Fri, 18 Jun 2004 22:47:56 GMT # From: hoff@hp.nospam (Hoff Hoffman) Y Subject: Re: sftp and non-stream_lf files? (was: Re: Secure Web Server Version 2.0 availa 2 Message-ID: <wiKAc.4286$S44.1517@news.cpqcorp.net>  b In article <843706dc.0406181308.10a9a6e6@posting.google.com>, horn@shsu.edu (James T Horn) writes:C :In all this discussion, we are finding that using sftp to transfer F :files between systems is having problems with the non-streamlf files.  C   Details of what you mean by "having problems" would be useful, as D   would details of the (OpenVMS?) host, the IP stack, the command(s)   used, errors, etc.   ..B :Anyone having the same kind of problem with sftp and non-streamlf :files?      Guessing at the problem...  E   ftp file "corruptions" are quite common (particularly when working  D   with a mixture of operating systems), hence the associated sectionE   in the FAQ around recovering corrupted BACKUP savesets, and various %   discussions here in the newsgroups.   D   When using ftp or sftp, I generally zip the files first, and then    transfer them.  D   If this is TCP/IP Services, there may well be related information C   on the handling of non-stream_lf files in the release notes, and  C   specifically associated with a known limitation of sftp and scp.  B   Check the release notes for whichever version and ECO is in use.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------   Date: 18 Jun 04 12:18:46 PST From: mckinneyj@cpva.saic.com - Subject: Re: Split I/Os to contiguous file??? ( Message-ID: <sVqr6swsNkm5@cpva.saic.com>  = In article <bdc65a53.0406180924.49bc3df0@posting.google.com>, '  gspamtackett@yahoo.com (Galen) writes: @ > My application performs write QIOWs to a single multi-gigabyteF > contiguous file (created contiguous-best-try but actually contiguous? > since it has only one mapping pointer in its header). MONITOR H > PROCESS/TOPDIO shows that it is the only process doing any significant
 > direct I/O.  > H > When the application is running, MONITOR IO shows a high rate of split: > I/Os, about 200-300 per second. Using some browser-based@ > instrumentation built into one of my disk controllers (a KZPDCE > SmartArray 5304A) I can see that the largest I/O being performed at D > the controller is only 256 blocks, while my application issues its0 > QIOs for multiple thousands of blocks per I/O. > D > This same behavior apparently occurs using an identical file via aF > vanilla internal builtin SCSI controller and drive. (I can't observeF > the actual I/O sizes for this hardware, but I can see that the levelF > of split I/Os and the DIO rate both are similar to what I see on the	 > KZPDC.)  > E > What is causing the I/Os to be split, when the file is contiguous?? B > Does the "magic" 256-block maximum actual I/O provide any clues? > A > This is an OpenVMS V7.3-2 EV6 500MHz ES40 with 1 GB RAM that is F > current on ECOs, except that my FIBRE_SCSI V2 may not be the newest.B > (The release notes for the newest FIBRE_SCSI don't appear to sayH > anything that might be related to this issue.) It is not clustered andF > I am the only user, running the only "user" application (the rest is5 > HP TCP/IP, DECnet and Web-Based Management Agents.)   ? I recall that the SCSI disk drivers used to limit the size of a > disk transfer - result being split I/Os when you exceeded this@ limit. Don't recall what the limit is/was - but your descriptionB suggests its still there with a value of 256 blocks for DKDRIVER -4 perhaps someone else will follow-up to confirm this.   ------------------------------  # Date: Fri, 18 Jun 2004 20:03:54 GMT 1 From: "konabear" <konabearg-newsgroups@yahoo.com> - Subject: Re: Split I/Os to contiguous file??? A Message-ID: <KUHAc.25599$eH1.11951219@newssvr28.news.prodigy.com>   C All disks (devices?)  have a max transfer size.  It's listed in the   UCB$L_MAXBCNT of the UCB.  So...  
 $ ANAL/SYSTEM  SDA> READ SYSDEF SDA> show dev dsa0   I/O data structures  ------------------- G DSA0                                           HSX00               UCB:  81676500 <snip> SDA> SET OUTPUT x.x  SDA> FORMAT 81676500 SDA> SET OUTPUT TT: ! SDA> SPAWN SEAR x.x UCB$L_MAXBCNT < FFFFFFFF.81676618   UCB$L_MAXBCNT                   0000FE007 SDA> eval 0000FE00/200 ! 200 hex being 512 bytes/blocks > Hex = 00000000.0000007F   Decimal = 127         DYN$C_SECURITY  K So this particular disk has a maximum block transfer of 127.  If the QIO(W) 2 request exceeds that, then you'll get a split I/O.  D Somewhere I was told that it's faster if the application is aware ofK UCB$L_MAXBCNT and keeps its I/O smaller than that, rather than to incur the  split I/Os.    Todd* <mckinneyj@cpva.saic.com> wrote in message" news:sVqr6swsNkm5@cpva.saic.com...? > In article <bdc65a53.0406180924.49bc3df0@posting.google.com>, ) >  gspamtackett@yahoo.com (Galen) writes: B > > My application performs write QIOWs to a single multi-gigabyteH > > contiguous file (created contiguous-best-try but actually contiguousA > > since it has only one mapping pointer in its header). MONITOR J > > PROCESS/TOPDIO shows that it is the only process doing any significant > > direct I/O.  > > J > > When the application is running, MONITOR IO shows a high rate of split< > > I/Os, about 200-300 per second. Using some browser-basedB > > instrumentation built into one of my disk controllers (a KZPDCG > > SmartArray 5304A) I can see that the largest I/O being performed at F > > the controller is only 256 blocks, while my application issues its2 > > QIOs for multiple thousands of blocks per I/O. > > F > > This same behavior apparently occurs using an identical file via aH > > vanilla internal builtin SCSI controller and drive. (I can't observeH > > the actual I/O sizes for this hardware, but I can see that the levelH > > of split I/Os and the DIO rate both are similar to what I see on the > > KZPDC.)  > > G > > What is causing the I/Os to be split, when the file is contiguous?? D > > Does the "magic" 256-block maximum actual I/O provide any clues? > > C > > This is an OpenVMS V7.3-2 EV6 500MHz ES40 with 1 GB RAM that is H > > current on ECOs, except that my FIBRE_SCSI V2 may not be the newest.D > > (The release notes for the newest FIBRE_SCSI don't appear to sayJ > > anything that might be related to this issue.) It is not clustered andH > > I am the only user, running the only "user" application (the rest is7 > > HP TCP/IP, DECnet and Web-Based Management Agents.)  > A > I recall that the SCSI disk drivers used to limit the size of a @ > disk transfer - result being split I/Os when you exceeded thisB > limit. Don't recall what the limit is/was - but your descriptionD > suggests its still there with a value of 256 blocks for DKDRIVER -6 > perhaps someone else will follow-up to confirm this.   ------------------------------  # Date: Fri, 18 Jun 2004 20:57:09 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks) - Subject: Re: Split I/Os to contiguous file??? - Message-ID: <saMfZyPRBQ28@cuebid.zko.dec.com>    mckinneyj@cpva.saic.com writes: ) >  gspamtackett@yahoo.com (Galen) writes:   I >> When the application is running, MONITOR IO shows a high rate of split ; >> I/Os, about 200-300 per second. Using some browser-based A >> instrumentation built into one of my disk controllers (a KZPDC F >> SmartArray 5304A) I can see that the largest I/O being performed atE >> the controller is only 256 blocks, while my application issues its 1 >> QIOs for multiple thousands of blocks per I/O.  > A > I recall that the SCSI disk drivers used to limit the size of a @ > disk transfer - result being split I/Os when you exceeded thisB > limit. Don't recall what the limit is/was - but your descriptionD > suggests its still there with a value of 256 blocks for DKDRIVER -6 > perhaps someone else will follow-up to confirm this.  L A gating factor on the size of an I/O is ucb$l_maxbcnt (size in bytes of theB largest I/O transfer this device will support).  Assuming you haveD access to SDA, you can figure out what this value is for any device.  
 $ ANAL/SYS SDA> read iodef  SDA> SHOW DEVICE $1$DGA12  [stuff appears on screen]  SDA> EXAM UCB+UCB$L_MAXBCNT * UCB+00190:  00000000.00020000   "........", SDA> EVAL 20000 / ^D512 			! bytes per block' Hex = 00000000.00000100   Decimal = 256   ( The above is from an HSG80-attached disk   --    M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com    ------------------------------    Date: 18 Jun 2004 16:03:17 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) - Subject: Re: Split I/Os to contiguous file??? 3 Message-ID: <06azPfuDr+OU@eisner.encompasserve.org>   d In article <bdc65a53.0406180924.49bc3df0@posting.google.com>, gspamtackett@yahoo.com (Galen) writes: > H > When the application is running, MONITOR IO shows a high rate of split: > I/Os, about 200-300 per second. Using some browser-based@ > instrumentation built into one of my disk controllers (a KZPDCE > SmartArray 5304A) I can see that the largest I/O being performed at D > the controller is only 256 blocks, while my application issues its0 > QIOs for multiple thousands of blocks per I/O. >   !    256 * 512 = 131072 = 20000 hex   H    Sounds like an 18 bit limitation in the driver.  Hopefully reflectingE    a limitation in the hardware, but perhaps relecting the rareity of E    real users exceeding UNIBUS address space in a single I/O request.    ------------------------------  % Date: Fri, 18 Jun 2004 21:40:38 -0400 " From: Glenn Everhart <gce@gce.com>- Subject: Re: Split I/Os to contiguous file??? 3 Message-ID: <40d399d4$0$3026$61fed72c@news.rcn.com>   ? Most if not all SCSI port drivers limit transfers to 64KB. This @ is less of a real issue on SCSI disks than it might appear sinceA dkdriver will have many I/Os sent to the drive, to be executed at A drive speed, if the disk supports tagged queueing. It does affect A tape max blocksize, but that is not really a performance issue on ? modern tapes much; they have an underlying block structure (and A lots of ECC) which means long records don't really save tape like  they do on 9track.  E The reason for not boosting the port driver I/O size has been mapping D resources in the controllers (and to some extent the processors). IfE most of the mapping resources get tied up for the duration of a large D I/O, other I/O gets starved. In a PC situation the response seems toF be to expect folks not to do large I/O unless they have some reason toC want it. In VMS, the design is to avoid letting any one user tie up D resources enough to materially screw up another. "User" in this caseD can be person or just some process...the same logic applies, and theE application case is perhaps more common. Thus while port drivers with B larger max SCSI I/O have been written (in some it's just setting aD maximum size parameter up and recompiling as I recall), they are not used.   D There is also the unwillingness to have extra long single operationsC where the system might have to wait for them to clear. That too can G take longer for very large operations. That situation occurs more often G than you might think: the XQP often forces read or writecheck (there is E a sysgen parameter to control it) and that check operation forces the S disk driver to single stream mode, i.e., waiting for everything else. Any errors do # that too, but they are less common.    Glenn Everhart   mckinneyj@cpva.saic.com wrote:? > In article <bdc65a53.0406180924.49bc3df0@posting.google.com>, ) >  gspamtackett@yahoo.com (Galen) writes:  > @ >>My application performs write QIOWs to a single multi-gigabyteF >>contiguous file (created contiguous-best-try but actually contiguous? >>since it has only one mapping pointer in its header). MONITORvH >>PROCESS/TOPDIO shows that it is the only process doing any significant
 >>direct I/O.O >>H >>When the application is running, MONITOR IO shows a high rate of split: >>I/Os, about 200-300 per second. Using some browser-based@ >>instrumentation built into one of my disk controllers (a KZPDCE >>SmartArray 5304A) I can see that the largest I/O being performed atbD >>the controller is only 256 blocks, while my application issues its0 >>QIOs for multiple thousands of blocks per I/O. >>D >>This same behavior apparently occurs using an identical file via aF >>vanilla internal builtin SCSI controller and drive. (I can't observeF >>the actual I/O sizes for this hardware, but I can see that the levelF >>of split I/Os and the DIO rate both are similar to what I see on the	 >>KZPDC.)s >>E >>What is causing the I/Os to be split, when the file is contiguous??sB >>Does the "magic" 256-block maximum actual I/O provide any clues? >>A >>This is an OpenVMS V7.3-2 EV6 500MHz ES40 with 1 GB RAM that isAF >>current on ECOs, except that my FIBRE_SCSI V2 may not be the newest.B >>(The release notes for the newest FIBRE_SCSI don't appear to sayH >>anything that might be related to this issue.) It is not clustered andF >>I am the only user, running the only "user" application (the rest is5 >>HP TCP/IP, DECnet and Web-Based Management Agents.)  >  > A > I recall that the SCSI disk drivers used to limit the size of a @ > disk transfer - result being split I/Os when you exceeded thisB > limit. Don't recall what the limit is/was - but your descriptionD > suggests its still there with a value of 256 blocks for DKDRIVER -6 > perhaps someone else will follow-up to confirm this.   ------------------------------  % Date: Fri, 18 Jun 2004 23:07:58 -0500D2 From: David J Dachtera <djesys.nospam@comcast.net>- Subject: Re: Split I/Os to contiguous file??? + Message-ID: <40D3BC1D.A3897996@comcast.net>'   Galen wrote: > @ > My application performs write QIOWs to a single multi-gigabyteF > contiguous file (created contiguous-best-try but actually contiguous? > since it has only one mapping pointer in its header). MONITORsH > PROCESS/TOPDIO shows that it is the only process doing any significant
 > direct I/O.a > H > When the application is running, MONITOR IO shows a high rate of split: > I/Os, about 200-300 per second. Using some browser-based@ > instrumentation built into one of my disk controllers (a KZPDCE > SmartArray 5304A) I can see that the largest I/O being performed atcD > the controller is only 256 blocks, while my application issues its0 > QIOs for multiple thousands of blocks per I/O. > D > This same behavior apparently occurs using an identical file via aF > vanilla internal builtin SCSI controller and drive. (I can't observeF > the actual I/O sizes for this hardware, but I can see that the levelF > of split I/Os and the DIO rate both are similar to what I see on the	 > KZPDC.)e > E > What is causing the I/Os to be split, when the file is contiguous??oB > Does the "magic" 256-block maximum actual I/O provide any clues? > A > This is an OpenVMS V7.3-2 EV6 500MHz ES40 with 1 GB RAM that is F > current on ECOs, except that my FIBRE_SCSI V2 may not be the newest.B > (The release notes for the newest FIBRE_SCSI don't appear to sayH > anything that might be related to this issue.) It is not clustered andF > I am the only user, running the only "user" application (the rest is5 > HP TCP/IP, DECnet and Web-Based Management Agents.)l  B "Split I/O"'s have to do with "window turns", not fragmentation or5 contiguity. The I/O size discussion is a red herring.l  F The item you need to look at is the /WINDOWS value when the volume was@ INITIALIZEd or MOUNTed. See the related doc.'s for a much betterF explanation than I could hope to give. Examine its relationship to theF actual size of the file and see if anything is being exceeded (size of the extent cache, etc.).   D.J.D.     D.J.D.   ------------------------------  % Date: Sat, 19 Jun 2004 01:14:19 -0400f  From: John Santos <JOHN@egh.com>- Subject: Re: Split I/Os to contiguous file??? 4 Message-ID: <1040619010420.404B-100000@Ives.egh.com>  , On Fri, 18 Jun 2004, David J Dachtera wrote:   > Galen wrote: > > B > > My application performs write QIOWs to a single multi-gigabyteH > > contiguous file (created contiguous-best-try but actually contiguous --> ^^^^^^^^^^A > > since it has only one mapping pointer in its header). MONITORn) ---------------> ^^^^ ^^^ ^^^^^^^ ^^^^^^^s  J > > PROCESS/TOPDIO shows that it is the only process doing any significant > > direct I/O.b > > J > > When the application is running, MONITOR IO shows a high rate of split< > > I/Os, about 200-300 per second. Using some browser-basedB > > instrumentation built into one of my disk controllers (a KZPDCG > > SmartArray 5304A) I can see that the largest I/O being performed at>F > > the controller is only 256 blocks, while my application issues its2 > > QIOs for multiple thousands of blocks per I/O. > > F > > This same behavior apparently occurs using an identical file via aH > > vanilla internal builtin SCSI controller and drive. (I can't observeH > > the actual I/O sizes for this hardware, but I can see that the levelH > > of split I/Os and the DIO rate both are similar to what I see on the > > KZPDC.)t > > G > > What is causing the I/Os to be split, when the file is contiguous?? D > > Does the "magic" 256-block maximum actual I/O provide any clues? > > C > > This is an OpenVMS V7.3-2 EV6 500MHz ES40 with 1 GB RAM that is H > > current on ECOs, except that my FIBRE_SCSI V2 may not be the newest.D > > (The release notes for the newest FIBRE_SCSI don't appear to sayJ > > anything that might be related to this issue.) It is not clustered andH > > I am the only user, running the only "user" application (the rest is7 > > HP TCP/IP, DECnet and Web-Based Management Agents.)  > D > "Split I/O"'s have to do with "window turns", not fragmentation or7 > contiguity. The I/O size discussion is a red herring.k > H > The item you need to look at is the /WINDOWS value when the volume wasB > INITIALIZEd or MOUNTed. See the related doc.'s for a much betterH > explanation than I could hope to give. Examine its relationship to theH > actual size of the file and see if anything is being exceeded (size of > the extent cache, etc.).  H As the subject and OP say, the file is contiguous, so window turns don'tD enter into it.  If a file has fewer extents than the /WINDOWS value,D then there is never any need to window turn.  (HELP says the defaultD and minimum value for /windows is 7, but even if it were set somehow@ to 1, no window returns are ever required for contiguous files.)   -- v John Santosi Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Fri, 18 Jun 2004 14:41:15 -0400-, From: "warren sander" <warren.sander@hp.com>' Subject: Re: SWS-2.0 (Apache) & CGI ???i, Message-ID: <40d3392b$1@usenet01.boi.hp.com>  H The reason I migrated the OpenVMS site from Purveyor which it ran for 5+" years to CSWS was because I neededI server side includes (I know purveyor can do them) but I needed them withR" .html and .htm extensions not .htp  H Having server side includes with .html extensions has saved me literally+ thousands of hours in updateing the site too new brand changes.  I Renaming all the current files to .htp and having all the pointers to thec. site and all the bookmarks etc break wasn't an option..  ; Purveyor was great in it's day. But the times have changed.    -warren    ------------------------------    Date: 18 Jun 2004 15:15:23 -0700( From: bob@instantwhip.com (Bob Ceculski)' Subject: Re: SWS-2.0 (Apache) & CGI ??? = Message-ID: <d7791aa1.0406181415.3b6a713a@posting.google.com>   ` "warren sander" <warren.sander@hp.com> wrote in message news:<40d3392b$1@usenet01.boi.hp.com>...J > The reason I migrated the OpenVMS site from Purveyor which it ran for 5+$ > years to CSWS was because I neededK > server side includes (I know purveyor can do them) but I needed them withs$ > .html and .htm extensions not .htp > J > Having server side includes with .html extensions has saved me literally- > thousands of hours in updateing the site toM > new brand changes. > K > Renaming all the current files to .htp and having all the pointers to thee0 > site and all the bookmarks etc break wasn't an	 > option.t > = > Purveyor was great in it's day. But the times have changed.i > 	 > -warrenn  ? purveyor has aliases ... all you had to do is redefine the .htmt? extensions to .htp, just like we do .asp extensions ... simple!l   ------------------------------    Date: 18 Jun 2004 15:17:25 -0700( From: bob@instantwhip.com (Bob Ceculski)' Subject: Re: SWS-2.0 (Apache) & CGI ???r= Message-ID: <d7791aa1.0406181417.233083e5@posting.google.com>   ` "warren sander" <warren.sander@hp.com> wrote in message news:<40d3392b$1@usenet01.boi.hp.com>...J > The reason I migrated the OpenVMS site from Purveyor which it ran for 5+$ > years to CSWS was because I neededK > server side includes (I know purveyor can do them) but I needed them with-$ > .html and .htm extensions not .htp > J > Having server side includes with .html extensions has saved me literally- > thousands of hours in updateing the site to  > new brand changes. > K > Renaming all the current files to .htp and having all the pointers to theN0 > site and all the bookmarks etc break wasn't an	 > option.u > = > Purveyor was great in it's day. But the times have changed.. > 	 > -warrenu  ; if you don't want to do redefinitions, why not just buy thea0 c source code for it from process and change it?8 and purveyor is still great, and soon you will see it on itanium!   ------------------------------  # Date: Fri, 18 Jun 2004 17:48:26 GMTe% From: "John Vottero" <John@mvpsi.com>R/ Subject: Re: VMS Gets Short Shrift at HP World.aA Message-ID: <KVFAc.25556$eH1.11932010@newssvr28.news.prodigy.com>d  . "John Smith" <a@nonymous.com> wrote in message& news:nOudnfgbdeXFkE_dRVn-hw@igs.net... [snip] >.J > If not a .ppt presentation, then perhaps we should each send carly(tm) aG > brick via USPS (just to keep VMS involved) along with a cover letter:a >  > Dear carly(tm),> >uG > The enclosed brick is being sent to you by an irate OpenVMS customer.v >$G > We are disgusted at the lack of marketing and advertising HP does forI
 > OpenVMS. >nI > I'd have come in person to deliver this brick but I fear that I will beuJ > laid-off as our company migrates from OpenVMS to Windows & Linux on DellJ > hardware, so I didn't really think I could afford the trip to California at > this time. >iF > Please throw this brick through the window nearest to your office to attempts' > to understand my frustration with HP.d >v  I This would backfire.  Carly would see "migrates from OpenVMS to Windows &uI Linux" and conclude that HP has to advertise their expertise with Windowse
 and Linux.   ------------------------------  # Date: Fri, 18 Jun 2004 18:38:39 GMTi" From:   VAXman-  @SendSpamHere.ORG/ Subject: Re: VMS Gets Short Shrift at HP World. 0 Message-ID: <00A338CA.68BF4537@SendSpamHere.ORG>  i In article <KVFAc.25556$eH1.11932010@newssvr28.news.prodigy.com>, "John Vottero" <John@mvpsi.com> writes:e/ >"John Smith" <a@nonymous.com> wrote in messages' >news:nOudnfgbdeXFkE_dRVn-hw@igs.net...l >[snip]v >>K >> If not a .ppt presentation, then perhaps we should each send carly(tm) a H >> brick via USPS (just to keep VMS involved) along with a cover letter: >> >> Dear carly(tm), >>H >> The enclosed brick is being sent to you by an irate OpenVMS customer. >>H >> We are disgusted at the lack of marketing and advertising HP does for >> OpenVMS.n >>J >> I'd have come in person to deliver this brick but I fear that I will beK >> laid-off as our company migrates from OpenVMS to Windows & Linux on DelleK >> hardware, so I didn't really think I could afford the trip to California  >at 
 >> this time.s >>G >> Please throw this brick through the window nearest to your office to  >attempt( >> to understand my frustration with HP. >> >,J >This would backfire.  Carly would see "migrates from OpenVMS to Windows &J >Linux" and conclude that HP has to advertise their expertise with Windows >and Linux.S  I You have a valid point there John and no I am not talking about the pointU of Carly's pinhead.i   -- dB http://www.legacy-2000.com  for the *best* OpenVMS system securityC                             solutions that others only claim to be.o -- eK VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COMa            a5   "Well my son, life is like a beanstalk, isn't it?" F   ------------------------------    Date: 18 Jun 2004 11:45:30 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris):/ Subject: Re: VMS Gets Short Shrift at HP World.1= Message-ID: <cf15391e.0406181045.219524a4@posting.google.com>w  e David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<40D0F316.BA7D0FE6@comcast.net>...eF > I feel that pain, also, and I'm in Chicago! Now, there's a doings in8 > Chicago, and I can't afford it! ...without presenting.  D Maybe, since you're in Chicago, you could raise the funds by renting rooms to other attendees. :-)   ? There are very-significant discounts available for HP Certified  Professionals:   Certification Level   Price    Master ASE (MASE)     $395   ASE, CSE              $595   AIS, CSA              $595   APC, ASC              $795   APP, ASP, APS         $795  D Becoming a CSA on OpenVMS takes one exam ($100 at ProMetrics) -- seeE http://www.hp.com/certification/region/americas/csa/bcs/openvms.html e@ That brings the HP World price down from $1,790 (with early-bird discount) to only $595.a  1 Certification Levels and acronyms, for reference:d Master ASE (MASE)e! Accredited Systems Engineer (ASE)$% Certified Systems Administrator (CSA)   Certified Systems Engineer (CSE)$ Accredited Presales Consultant (APC)! Accredited Sales Consultant (ASC) ' Accredited Integration Specialist (AIS)$# Accredited Sales Professional (ASP)t& Accredited Presales Professional (APP)$ Accredited Platform Specialist (APS)   ------------------------------    Date: 18 Jun 2004 15:19:39 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)/ Subject: Re: VMS Gets Short Shrift at HP World. 3 Message-ID: <qluKH6O267n$@eisner.encompasserve.org>2  R In article <h6CdnSYVvefZC0zdRVn-ug@igs.net>, "John Smith" <a@nonymous.com> writes:N > Perhaps then it it time for a .ppt slideshow that can be conceived, producedL > and sent to HP executives worldwide and members of the BoD, entitled "Lies > My Vendor Told Me".-  I If that's not enough, I've got a tape of the old DigiBits commercials. Ink BETA format!  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"u& 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfbL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O  Save Model Rocketry from the HSA!   http://www.space-rockets.com/congress.htmlm   ------------------------------  # Date: Fri, 18 Jun 2004 20:15:31 GMTt" From:   VAXman-  @SendSpamHere.ORG/ Subject: Re: VMS Gets Short Shrift at HP World. 0 Message-ID: <00A338D7.F0D8E589@SendSpamHere.ORG>  q In article <cf15391e.0406181045.219524a4@posting.google.com>, keithparris_NOSPAM@yahoo.com (Keith Parris) writes: f >David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<40D0F316.BA7D0FE6@comcast.net>...G >> I feel that pain, also, and I'm in Chicago! Now, there's a doings ine9 >> Chicago, and I can't afford it! ...without presenting.. >.E >Maybe, since you're in Chicago, you could raise the funds by rentingI >rooms to other attendees. :-) >r@ >There are very-significant discounts available for HP Certified >Professionals:s >  Certification Level   Price >  Master ASE (MASE)     $395o >  ASE, CSE              $595a >  AIS, CSA              $595  >  APC, ASC              $795b >  APP, ASP, APS         $795i >dE >Becoming a CSA on OpenVMS takes one exam ($100 at ProMetrics) -- see F >http://www.hp.com/certification/region/americas/csa/bcs/openvms.html A >That brings the HP World price down from $1,790 (with early-birdc >discount) to only $595. >e2 >Certification Levels and acronyms, for reference: >Master ASE (MASE)" >Accredited Systems Engineer (ASE)& >Certified Systems Administrator (CSA)! >Certified Systems Engineer (CSE)a% >Accredited Presales Consultant (APC)O" >Accredited Sales Consultant (ASC)( >Accredited Integration Specialist (AIS)$ >Accredited Sales Professional (ASP)' >Accredited Presales Professional (APP)a% >Accredited Platform Specialist (APS)y  F Where are the operating system software accreditations?  How about an E acronym like Accredited Engineering Specialst of Operating Systems or,> AESOP.  (Or would such a fabled accreditation be too much?) ;0  C How about a tier of accreditations for hobbyists?  There could be a E Beginner Accredited Technology Enthusiast.  Of course, there might be E a few negative connotations if one goes for the master accreditation.   s -- tB http://www.legacy-2000.com  for the *best* OpenVMS system securityC                             solutions that others only claim to be.h -- mK VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COMn            z5   "Well my son, life is like a beanstalk, isn't it?" n   ------------------------------    Date: 18 Jun 2004 15:20:57 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)/ Subject: Re: VMS Gets Short Shrift at HP World.n3 Message-ID: <44XuDXb1ugzR@eisner.encompasserve.org>u  [ In article <i%7Ac.34820$Hg2.6969@attbi_s04>, "Dave Gudewicz" <k9jdk@NOSPAMarrl.net> writes: L > How did "BoD" get changed to "Body" in the note I posted?  I'm glad it did2 > now, the Landru thing took some of the edge off.  < Brobably someone's spell checker that mucked with the TLA...   > He's dead Jim.  Or is he?.  0 I believe his last words were "I'm dead Jim" :-)  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfrL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O  Save Model Rocketry from the HSA!   http://www.space-rockets.com/congress.htmlr   ------------------------------    Date: 18 Jun 2004 15:27:03 -06004 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)/ Subject: Re: VMS Gets Short Shrift at HP World.23 Message-ID: <ZmfHrN3vf9K6@eisner.encompasserve.org>s  U In article <00A337F7.DF760209@SendSpamHere.ORG>,   VAXman-  @SendSpamHere.ORG writes:nM > According to a recent article I read from a URL posted here or in a similareM > group, it is PPT which is dumbing down management.  You expect them to makecO > any sage and or educated decisions when their already feeble minds have been e< > further crippled by a Micro$hit PPT brick upside the head?  L At the recent CARTS meeting we got handouts of a couple "standard HP format"L powerpoint presentations. What I want to know is what is the part number forH the special color inkjet cartridge that HP uses internally. The one that looks like:u  : 	+-------+-------+---------------------------------------+ 	|	|	|					| 	|	|	|					| 	|	|	|					| 	|	|	|					| 	|	|	|					| 	| Red	|Yellow	|		Blue			| 	|	|	|					| 	|	|	|					| 	|	|	|					| 	|	|	|					| 	|	|	|					| 	|	|	|					|: 	+-------+-------+---------------------------------------+  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdfsL     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  O  Save Model Rocketry from the HSA!   http://www.space-rockets.com/congress.html    ------------------------------  % Date: Fri, 18 Jun 2004 16:45:21 -0400t# From: "John Smith" <a@nonymous.com>h/ Subject: Re: VMS Gets Short Shrift at HP World.o, Message-ID: <MKGdnf9TgrT4yU7d4p2dnA@igs.net>  > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message7 news:cf15391e.0406181045.219524a4@posting.google.com...t? > David J Dachtera <djesys.nospam@comcast.net> wrote in messagea' news:<40D0F316.BA7D0FE6@comcast.net>...tH > > I feel that pain, also, and I'm in Chicago! Now, there's a doings in: > > Chicago, and I can't afford it! ...without presenting. >pF > Maybe, since you're in Chicago, you could raise the funds by renting > rooms to other attendees. :-)i >CA > There are very-significant discounts available for HP Certified  > Professionals: >   Certification Level   Price  >   Master ASE (MASE)     $395 >   ASE, CSE              $595 >   AIS, CSA              $595 >   APC, ASC              $795 >   APP, ASP, APS         $795 >wF > Becoming a CSA on OpenVMS takes one exam ($100 at ProMetrics) -- seeF > http://www.hp.com/certification/region/americas/csa/bcs/openvms.htmlB > That brings the HP World price down from $1,790 (with early-bird > discount) to only $595.i >p3 > Certification Levels and acronyms, for reference:w > Master ASE (MASE)n# > Accredited Systems Engineer (ASE)o' > Certified Systems Administrator (CSA)P" > Certified Systems Engineer (CSE)& > Accredited Presales Consultant (APC)# > Accredited Sales Consultant (ASC)e) > Accredited Integration Specialist (AIS).% > Accredited Sales Professional (ASP)e( > Accredited Presales Professional (APP)& > Accredited Platform Specialist (APS)    I I'd be hugely surprised if any of the Sales accreditations include havingeK any knowledge of OpenVMS and why is should be sold in preference to windowso	 or linux.e   ------------------------------  % Date: Fri, 18 Jun 2004 22:52:31 -0500P2 From: David J Dachtera <djesys.nospam@comcast.net>/ Subject: Re: VMS Gets Short Shrift at HP World. + Message-ID: <40D3B87F.4E319168@comcast.net>e   Keith Parris wrote:o > g > David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<40D0F316.BA7D0FE6@comcast.net>...-H > > I feel that pain, also, and I'm in Chicago! Now, there's a doings in: > > Chicago, and I can't afford it! ...without presenting. > F > Maybe, since you're in Chicago, you could raise the funds by renting > rooms to other attendees. :-)) > A > There are very-significant discounts available for HP Certified- > Professionals: >   Certification Level   Pricec >   Master ASE (MASE)     $395 >   ASE, CSE              $595 >   AIS, CSA              $595 >   APC, ASC              $795 >   APP, ASP, APS         $795 > F > Becoming a CSA on OpenVMS takes one exam ($100 at ProMetrics) -- seeF > http://www.hp.com/certification/region/americas/csa/bcs/openvms.htmlB > That brings the HP World price down from $1,790 (with early-bird > discount) to only $595.i  F Now, balance that against the cost of acquiring the accreditation* and compute the ROI.  C *: ...unless you can ace the test drawing from life/work experienceh (possible, but not likely)./  H Sometimes I think these folks out east just don't grok the working class or our lot....   D.J.D.   ------------------------------  % Date: Fri, 18 Jun 2004 22:48:15 -0500g2 From: David J Dachtera <djesys.nospam@comcast.net>/ Subject: Re: VMS Gets Short Shrift at HP World.i+ Message-ID: <40D3B77F.817167E6@comcast.net>s   Bob Kaplow wrote:l > T > In article <h6CdnSYVvefZC0zdRVn-ug@igs.net>, "John Smith" <a@nonymous.com> writes:P > > Perhaps then it it time for a .ppt slideshow that can be conceived, producedN > > and sent to HP executives worldwide and members of the BoD, entitled "Lies > > My Vendor Told Me".t > K > If that's not enough, I've got a tape of the old DigiBits commercials. Ini > BETA format!  F I have a Beta machine that could be coaxed back to life, with a little
 3-in-1 oil...1   D.J.D.   ------------------------------  + Date: Fri, 18 Jun 2004 19:55:18 +0000 (UTC) * From: bleau@UMTOF.UMD.EDU (Lawrence Bleau); Subject: Re: Why does idle telnet session get disconnected?s0 Message-ID: <cavhb6$k3g$1@grapevine.wam.umd.edu>  K Hello, c.o.v.  I think we finally found a reason for the disconnect problem A we've observed here, and thought I'd share it with the newsgroup.r  F The problem was that an idle terminal session would be disconnected onJ later versions of VMS and TCPIP, but not earlier ones.  Additionally, theyH would get disconnected for certain off-campus sites (my boss's home) but3 not for telnet sessions originating from on campus.t  D It turns out two things were needed for this problem.  First was theJ upgrade to some later version of TCPIP that increased the default keepidleD period; earlier versions had it set at 150 (75 secs), later versions defaulted it to 14400 (2 hrs).  K The second thing that happened here was that, unknown to me at the time, myBH boss purchased an Apple Airport, sort of a router firewall combo for hisK home.  This, I imagine, intercepted incoming packets from tcp sessions, and I forwarded the packets over a separate circuit to the target system at hisOE home, reversing the process for outgoing packets.  This device has an K internal timer; by experimentation we determined it to be about 60 minutes.E  J It didn't see any packets from the idle telnet session for >60 minutes, soK it disconnected it.  The keepalive packets never reached the telnet client, E so the VMS telnet server thought the session was dead and deleted the H process, which was almost exactly at 2 hours.  The telnet client noticed the dead connection sooner.C  J When we set keepidle at exactly 7200 (1 hr) his session was kept alive forI hours on end, even though it was idle.  We increased keepidle in 1-minutew: increments, and at 1:05 the idle session was disconnected.  H When telnet sessions are started from on campus, there's no intermediateF firewall, so the telnet client received and correctly responded to the8 keepalive packets; hence why there's no problem at work.  J So, somewhat to my relief, we concluded that this problem was not VMS, notE TCP/IP, not the campus network, not the telnet server, not the telnetaG client, but the boss's own fault in making a network config change(!). I/ Hey, do solutions get any better than this? :-)-  K Anyway, we set keepidle down to 5 minutes (his decision, not mine) and it's K now working like a charm; no one gets disconnected.  Case closed; enjoy thet weekend!   Lawrence Bleau University of Maryland" Physics Dept., Space Physics Group 301-405-6223 bleau@umtof.umd.eduu   ------------------------------   Date: 18 Jun 2004 21:23:22 GMT, From: bill@gw5.cs.uofs.edu (Bill Gunshannon); Subject: Re: Why does idle telnet session get disconnected?e+ Message-ID: <2jh4q9F12406qU1@uni-berlin.de>   0 In article <cavhb6$k3g$1@grapevine.wam.umd.edu>,- 	bleau@UMTOF.UMD.EDU (Lawrence Bleau) writes:tM > Hello, c.o.v.  I think we finally found a reason for the disconnect problem C > we've observed here, and thought I'd share it with the newsgroup.r > H > The problem was that an idle terminal session would be disconnected onL > later versions of VMS and TCPIP, but not earlier ones.  Additionally, theyJ > would get disconnected for certain off-campus sites (my boss's home) but5 > not for telnet sessions originating from on campus.) > F > It turns out two things were needed for this problem.  First was theL > upgrade to some later version of TCPIP that increased the default keepidleF > period; earlier versions had it set at 150 (75 secs), later versions  > defaulted it to 14400 (2 hrs). > M > The second thing that happened here was that, unknown to me at the time, myuJ > boss purchased an Apple Airport, sort of a router firewall combo for hisM > home.  This, I imagine, intercepted incoming packets from tcp sessions, and-K > forwarded the packets over a separate circuit to the target system at hiscG > home, reversing the process for outgoing packets.  This device has an M > internal timer; by experimentation we determined it to be about 60 minutes.. > L > It didn't see any packets from the idle telnet session for >60 minutes, soM > it disconnected it.  The keepalive packets never reached the telnet client,rG > so the VMS telnet server thought the session was dead and deleted themJ > process, which was almost exactly at 2 hours.  The telnet client noticed > the dead connection sooner.e > L > When we set keepidle at exactly 7200 (1 hr) his session was kept alive forK > hours on end, even though it was idle.  We increased keepidle in 1-minute < > increments, and at 1:05 the idle session was disconnected. > J > When telnet sessions are started from on campus, there's no intermediateH > firewall, so the telnet client received and correctly responded to the: > keepalive packets; hence why there's no problem at work. > L > So, somewhat to my relief, we concluded that this problem was not VMS, notG > TCP/IP, not the campus network, not the telnet server, not the telnet I > client, but the boss's own fault in making a network config change(!).  1 > Hey, do solutions get any better than this? :-)$ > M > Anyway, we set keepidle down to 5 minutes (his decision, not mine) and it's0M > now working like a charm; no one gets disconnected.  Case closed; enjoy the0
 > weekend! >   C This comes as no surprise to me although I opted to keep out of the D original discusion.  Our campus did the same thing, probably for theB same reason.  It amazes me how sales droids can convince otherwiseD competent network people to break the TCPIP protocol with such ease.  D The supposed reason for doing this was to prevent "an attacker" fromB opening a lot of sessions and leaving them behind.  Sadly, that isC exactly what this "solution" does. After exceeding the idle time if H the user on the remote end types something his packets don't get throughC and his client drops the session.  But, unless something causes thetC host to send data the idle session will stay there until explicitlyr! killed or the system is rebooted.t   Idiots!!   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   l   ------------------------------   End of INFO-VAX 2004.338 ************************UNIBUS address space in a single I/O request.    ------------------------------  % Date: Fri, 18 Jun 2004 21:40:38 -0400 " From: Glenn Everhart <gce@gce.com>- Subject: Re: Split I/Os to contiguous file??? 3 Me----------*/

#if (!defined(SFX) || defined(SFX_EXDIR))
    if (FUNCTION == ROOT) {
        Trace((stderr, "initializing root path to [%s]\n",
          FnFilter1(pathcomp)));
        if (pathcomp == (char *)NULL) {
            rootlen = 0;
            return MPN_OK;
        }
        if (rootlen > 0)        /* rootpath was already set, nothing to do */
            return MPN_OK;
        if ((rootlen = strlen(pathcomp)) > 0) {
            int had_trailing_pathsep=FALSE, has_drive=FALSE, add_dot=FALSE;
            char *tmproot;

            if ((tmproot = (char *)malloc(rootlen+3)) == (char *)NULL) {
                rootlen = 0;
                return MPN_NOMEM;
            }
            strcpy(tmproot, pathcomp);
            if (isalpha((uch)tmproot[0]) && tmproot[1] == ':')
                has_drive = TRUE;   /* drive designator */
            if (tmproot[rootlen-1] == '/' || tmproot[rootlen-1] == '\\') {
                tmproot[--rootlen] = '\0';
                had_trailing_pathsep = TRUE;
            }
            if (has_drive && (rootlen == 2)) {
                if (!had_trailing_pathsep)   /* i.e., original wasn't "x:/" */
                    add_dot = TRUE;    /* relative path: add '.' before '/' */
            } else if (rootlen > 0 && (SSTAT(tmproot, &G.statbuf) ||
                       !S_ISDIR(G.statbuf.st_mode))) /* path does not exist */
            {
                if (!G.create_dirs /* || iswild(tmproot) */ ) {
                    free(tmproot);
                    rootlen = 0;
                    /* skip (or treat as stored file) */
                    return MPN_INF_SKIP;
                }
                /* create the directory (could add loop here scanning tmproot
                 * to create more than one level, but why really necessary?) */
                                                                                                                                                                                                                                                           