1 INFO-VAX	Mon, 11 Dec 2006	Volume 2006 : Issue 681       Contents: Re: Bad Shadow set member  Re: Bad Shadow set member ! Re: IEEE Decimal Float on Itanium ! Re: IEEE Decimal Float on Itanium * Re: Managed vs unmanaged switch in cluster* Re: Managed vs unmanaged switch in cluster* Re: Managed vs unmanaged switch in cluster Re: mysql client mysql client  Re: NAT router software on VMS ?  Re: NAT router software on VMS ?) Re: OpenVMS Weekly Audio Update - Show #1 % OpenVMS Weekly Audio Update - Show #1 ) Re: OpenVMS Weekly Audio Update - Show #1 ) Re: sys v0600 patch for openvms 8.2 alpha  Re: The Hole in Cerner's Logic Re: The Hole in Cerner's Logic Re: The Hole in Cerner's Logic Re: Who called me?* Re: [DEFRAG] Umpteen OPEN files in logfile  F ----------------------------------------------------------------------  % Date: Sun, 10 Dec 2006 13:36:57 +0100 4 From: Jur van der Burg <"vdburg at hotmail dot com">" Subject: Re: Bad Shadow set member4 Message-ID: <457bff6a$0$323$e4fe514c@news.xs4all.nl>  ' That depens on how broken the disk was.   J  > But the other disk was broken, so no copy is even possible. Wouldn't itD  > make sense for the shadowing software to mount the only available
  > member?  & Sure. Don't use /INCLUDE in that case.    >And even with a merge, theC  > software is going to have to randomly pick one or the other disk   > anyway, right?   F If you mean for copying the data, no. There's still a 'master' member,F but the handling in case of errors is different than with a full copy.C In the merge case any errors will be propagated from the bad to the B good member. The only thing that counts in such a case is that the disk are equal in the end.   Jur.  
 AEF wrote: > Jur van der Burg wrote: G >> That's not a bug. You asked specifically for NO copy, and due to the , >> power outage a merge copy is required.... > I > But the other disk was broken, so no copy is even possible. Wouldn't it C > make sense for the shadowing software to mount the only available D > member? If you had a one-member shadow set, you're saying it stillH > wouldn't mount because of the power outage? And even with a merge, theB > software is going to have to randomly pick one or the other disk > anyway, right? > G >> Perhaps you would like the /POLICY=REQUIRE_MEMBERS for mount so that , >> it insists on both members to be present. > 4 > I think that's exactly what the OP *doesn't* want. >  > AEF  >  >> Jur.  >> >> Tom Linden wrote:K >>> On Sat, 09 Dec 2006 19:12:40 -0800, AEF <spamsink2001@yahoo.com> wrote:  >>>  >>>> Tom Linden wrote:I >>>>> On Sat, 09 Dec 2006 12:19:35 -0800, <norm.raphael@metso.com> wrote:  >>>>> L >>>>>> "Tom Linden" <tom@kednos-remove.com> wrote on 12/09/2006 09:31:17 AM:
 >>>> [...]M >>>>> Sorry, should have been more clear.  The mount caommand issued manually  >>>>> worked fine  as below 4 >>>>> MOUNT DSA0:/CLUSTER/NOASSIST/INCLUDE/NOCOPY/ -$ >>>>>   SHADOW=($42$DKA1300:) COMMON >>>>> J >>>>> What I was puzzling over was the mount command from the startup file
 >>>>> that	 >>>>> had N >>>>> both members listed failed.  Why wouldn't mount, knowing that one member	 >>>>> had N >>>>> failed, take corrective action and mount the working member.  This seems) >>>>> fundamental to me for a shadow set. " >>>> What was the error message???/ >>> Looking through my logs this is all I found  >>>  >>> $     MOUNT L >>> DSA0:/CLUSTER/NOASSIST/INCLUDE/NOCOPY/SHADOW=($42$DKA1200:,$42$DKA1300:)
 >>> COMMON. >>> %MOUNT-F-SHDWCOPYREQ, shadow copy required >>> K >>> Unless the above command is incorrect, this is a bug.  If a member of a  >>> shadow set fails, A >>> the mount command should recover,  otherwise have shadow sets  >>> 
 >>>> [...] >>>> >>>> AEF >>>> >>>  >>> K >>> --Using Opera's revolutionary e-mail client: http://www.opera.com/mail/  >    ------------------------------  % Date: Sun, 10 Dec 2006 18:50:48 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> " Subject: Re: Bad Shadow set member7 Message-ID: <71018$457c9d6a$cef8887a$9182@TEKSAVVY.COM>    Tom Linden wrote: M > In my case bad meant dead, and that is the point.  When the system rebooted K > the the good member would not mount, but as i explained I had to manually I > mount it, which seems to me to be a flaw in the concept since the whole K > purpose of shadowing is to provide a higher degree of disaster tolerance.     I The VMS sofware makes a lot of effort to prevent bad/old/stale data from  9 being used as the source which overwrites your good data.     J On a running system, VMS has a good situational awareness of what is good K and bad in a shawdowset. But when you reboot, it can only count on reading  I bits on a drive , as well as the order in which drives are listed in the  < MOUNT command in order to reconstruct situational awareness.  L The first drive listed in a mount command has significance since this is an J implicit source of a copy operation if other drives are not valid members  of the shadowset.   C Now , /INCLUDE tells VMS to mount all disks that were part of that  K shadowset at the time of a crash/shutdown.  The HELP text mentions that it  F needs to be given only one physical drive specification (where it can I lookup the names of other participating drives at the time of the crash).   I However, in your case, since one of the members was no longer available,  K VMS was incapable of rebuilding the shadowset to the same config as it was  > at the time of shutdown since one drive could not be included.  I Personally, I prefer to list the drives in the mount command in an order  I where the first drive is the one I know is to be used as a source should  L any copy/merge operation been needed.  In a multi node situation, there are L cases where the node with the "older" drives must be used as a source since I it contains the valid production data.  So you create the shadowset with  J that drive only and then bring the others into the shadowset. And that is F where the significance of the first drive on the list comes into play.   ------------------------------    Date: 11 Dec 2006 08:37:50 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) * Subject: Re: IEEE Decimal Float on Itanium3 Message-ID: <TTkT9TtUbMdF@eisner.encompasserve.org>   q In article <457a2f58$0$49195$14726298@news.sunsite.dk>, =?ISO-8859-15?Q?Arne_Vajh=F8j?= <arne@vajhoej.dk> writes:  > F > decimal is actually a builtin type in the newest "big" language: C#.  C    New?  I'm quite sure that thing has been around, not gaining any #    foothold, for over a decade now.    ------------------------------    Date: 11 Dec 2006 08:44:11 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) * Subject: Re: IEEE Decimal Float on Itanium3 Message-ID: <po$jeCOUt6+d@eisner.encompasserve.org>   n In article <1165730356.936089.246350@80g2000cwy.googlegroups.com>, "Bob Gezelter" <gezelter@rlgsc.com> writes: > G > I too have used binary integers for years in such situations, because A > in most cases that I have encountered, the scaling was known in G > advance. This is, in effect, using integer with an understood decimal  > point.  G    We've done that for years on processors which had no floating point. C    You're just implementing fixed point by hand, which is fine.  Of B    course the folks who love PL/I will wonder why you bother to do    it by hand.   ------------------------------    Date: 11 Dec 2006 05:30:56 -0800% From: briggs@eisner.encompasserve.org 3 Subject: Re: Managed vs unmanaged switch in cluster B Message-ID: <1165843855.956333.205630@79g2000cws.googlegroups.com>   Schroeder, AJ wrote: > JF Mezei wrote: E > > Also, in terms of speeds, since a switch is store-and-forward and L > > introduces a latency in packet delivery, is that significant in terms of; > > clustering and most notably MSCP access over ethernet ?  > H > When talking about switches vs. hubs, switches will almost *always* beC > faster because the switch stores MAC addresses in it's forwarding J > database. In turn that cuts down on the amount of ARP traffic that takes > place.  1 This bit about ARP traffic is arguably incorrect.   D MAC address caching helps to optimize unicast traffic, not broadcastD traffic.  ARP requests are broadcast frames and are still flooded toE all ports.  ARP replies are unicast traffic and are treated just like  any other unicast frame.  @ There is nothing about being on a switch that causes stations toA send any more or any fewer ARPs than they would if were on a hub.   ? Switches break up the collision domain.  They do not reduce the  broadcast domain.   ? However...  If you were to break your network up into VLANs and @ route between them, that would reduce the size of each broadcast< domain and _that_ could do something about the volume of ARP; broadcast requests received at each station on the network.   0 A switch can do [at least] three things for you:  > 1.  Since it creates and maintains a MAC address cache, it canD     optimize unicast traffic by forwarding frames only to the single6     port through which the target device is reachable.  B     This cuts down on unwanted traffic received at each station on     the network.  D 2.  Since this also means that not all switch ports will be involved=     in forwarding each frame, the switch can process multiple      frames concurrently.  A     This means that the switch can handle aggregate traffic loads B     greater than any single switch port could process.  (El cheapo@     switches may not support the high backplane forwarding rates+     needed to take full advantage of this).   D     A hub is limited to what a single port can handle.  Half-duplex.  D 3.  If the switch supports VLANs, it can partition your network into>     smaller virtual networks, allowing you to reduce broadcast-     traffic and potentially improve security.   G > Cisco makes a great switch, and you can pick up older switches off of H > ebay that will have IOS, and the managability you are looking for. TheD > only thing that I am not sure about is the port mirroring. I wouldI > assume that since it is a managed switch that you can do a port mirror, & > since that is a management function.  A Cisco switches using CatOS use the "set span" command.  Only core E switches and very old closet switches are likely to be running CatOS.   @ Cisco switches using IOS use the "monitor" command.  If I recallG correctly, older releases use "port monitor" in interface configuration A mode.  Newer releases use "monitor" in global configuration mode.    ------------------------------  + Date: Mon, 11 Dec 2006 18:20:31 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk3 Subject: Re: Managed vs unmanaged switch in cluster , Message-ID: <elk7hf$alh$1@south.jnrs.ja.net>  ^ In article <ql4rn2hft4974vc2h9vvisbav8irr8sppf@4ax.com>, Joe Bloggs <JBloggs@acme.com> writes:- >On Fri, 08 Dec 2006 15:50:50 -0500, JF Mezei & ><jfmezei.spamnot@teksavvy.com> wrote: > M >>Right now, my cluster is on a 10mbps hub. Time to upgrade. (and I ned more  	 >>ports).  >>G >>However, the affordable switches are all unmanaged and offer no port  N >>mirroring capabilities. (aka: ethermon becomes useless when trying to debug  >>network problems). > F >One possibility:  the Linksys wap/switches (WRT54xx)  running linux  ' >(via open-source firmware, eg DD-WRT)   >  >Price is right, ($60-120 USD).  > 7 >but only 4 switch ports.  and No port mirroring, afaik = >(afaik, if port mirroring can be done, I'd love to know how)  > $ >http://en.wikipedia.org/wiki/WRT54G  M But as it says in that article be careful about which version of the hardware M you purchase the later versions of the WRT54G and WRT54GS were first neutered I by reducing flash and Ram memory and then had linux replaced by Vx-Works. J DD-WRT is one of the few third-party firmware to support some of the later
 versions (see I http://www.dd-wrt.com/wiki/index.php/Version_5_and_6_Router_Information)  D but even it doesn't support the newest hardware version (version 7).M If you are going to buy a WRT54G or WRT54GS you really need to go to the shop O and see the serial number to determine which version it is (and most shops will ' probably only have the later versions).   J However linksys did release the WRT54GL which is basically the same as theN WRT54G version 4.0 and runs Linux based firmware (and is supported by DD-WRT)      David Webb   Security team leader CCSS Middlesex University   ------------------------------    Date: 11 Dec 2006 10:49:07 -0800( From: "Rich Jordan" <jordan@ccs4vms.com>3 Subject: Re: Managed vs unmanaged switch in cluster C Message-ID: <1165862947.793493.127750@n67g2000cwd.googlegroups.com>    Rich Jordan wrote: > JF Mezei wrote: N > > Right now, my cluster is on a 10mbps hub. Time to upgrade. (and I ned more > > ports).  > > H > > However, the affordable switches are all unmanaged and offer no portO > > mirroring capabilities. (aka: ethermon becomes useless when trying to debug  > > network problems). > > N > > Also, my netgear box with a built-in unmanaged switch has shown it doesn'tM > > handle cluster traffic 100% properly. Since low end switches are designed O > > for the PC market, I am a little weary of buying those consumer brands like  > > d-link or linksys. > > O > > In practice, is the loss of ethermon capabilities with use of a dumb switch  > > an issue ? > > K > > Would I be better off buying an older used Cisco switch with management L > > capbilities (also gives me experience on Cisco's IOS), or is buying just- > > any cheap switch brand new still better ?  > > G > > (I don't need gigabit speeds, since as far as I know, HP hasn't yet O > > produced a q-bus gigabit ethernet adaptor for my VAXes...) ajnd even on the @ > > alphas, it is 100mbps with the built-in ethernet adaptors.). > > E > > Also, in terms of speeds, since a switch is store-and-forward and L > > introduces a latency in packet delivery, is that significant in terms of; > > clustering and most notably MSCP access over ethernet ?  > A > We've been running work clusters for years on unmanaged Netgear F > switches.  At home I run a Gigafast consumer 16 port, no problems at > all. > H > When I tested transitioning my home cluster from a 10Base-T hub to theC > first Gigafast 8 port switch saw a noticeable improvement in MSCP F > access speeds between 100Mbit nodes, but I don't have my timings anyD > more.  ZIPping files across the network (ZIP on a PWS600au, sourceH > files and destination ZIP file on a VAXstation 3100m76) was measurably6 > faster (3-5%) than with the hub, which surprised me. > F > My Gigafast has a pseudo VLAN capability set with DIP switches.  YouI > can partition it into 8 separate 2-port VLANs, 4 x 4-port VLANs, or 2 x H > 8-port VLANs.  In the last config (and maybe others, I don't have docsI > handy) one of the ports was a 'mirror' port that could receive all data ) > being sent across that particular VLAN.  > E > I never used the capability so I can't speak for how well it works. H > Gigafast still lists a 24 port unit on their site with the capability. > No info on cost though.  >  > Rich  A Just in passing.  I found my gigafast user guide and it shows the  following info:   3 EZ series switches (EZ1600) are unmanaged and basic F EE series switches (EE1600) are managed via serial console and have an accessory slot. F EF series switches (EF1600-SI) are unmanaged but can set up VLANs with  mirror ports using DIP switches.  D I've seen the occasional EF and EE switch go fairly inexpensively on ebay,   F I have the EF1600-SI and it was quite inexpensive, and has worked justG fine, BUT its only use has been in a home environment; I'm not vouching < for their overall quality or manageability beyond that basicD experience.  I did play with the trunking once long ago and ethermonG was able to see all the traffic I expected to see on the various ports.   A BTW, you need to be careful about routers and modems that contain @ 'switches'.  If you look at the specs you might notice that onlyF support for TCPIP is listed.  The firmware in the units that do NAT orE connection ot broadband may very well not know how to deal with other F protocols (even though the switch should be protocol transparent).  WeF had a heck of a time with some Motorola powerline network and wirelessE network units that turned out to be systemic; they just didn't handle ; anything other than TCPIP despite having a 'bridging' mode.    ------------------------------    Date: 11 Dec 2006 01:14:22 -0800! From: "Ian Miller" <gxys@uk2.net>  Subject: Re: mysql client A Message-ID: <1165828462.329034.95040@73g2000cwn.googlegroups.com>    Have you tried the kit from 2 http://www.pi-net.dyndns.org/anonymous/kits/#mysql   I think it includes the client.    ------------------------------   Date: 11 Dec 2006 02:57:09 GMT# From: malakasoy <pana@gamithis.net>  Subject: mysql client : Message-ID: <Xns98968DE963399panagamithisnet@203.50.5.233>   Hi  G Does anyone have a precompiled mysql client for openvms 7.2 ? If so, i  % would really appreciate a link to it.   F i've downloaded the source and tried to compile the client, but i get . numerous errors and can never get it to work.    TIA  Peter    ------------------------------    Date: 11 Dec 2006 07:09:48 -0800( From: "Rich Jordan" <jordan@ccs4vms.com>) Subject: Re: NAT router software on VMS ? A Message-ID: <1165849788.487009.99620@f1g2000cwa.googlegroups.com>    JF Mezei wrote: M > The telco here doesn't regard the DSL service as requiring the same uptimes D > as the POTS telephone service even though both usethe same copper. > L > Twice in the past month, I have had problems with DSL connectivity becauseC > of a Bell Canada infrastructure problem (their ATM/BAS machines).  >  > Normally, one would have:  >  > < > [DSL-MODEM]-------[NAT ROUTER]----+-----[NODE1  10.0.0.11]= >                                    +-----[NODE2  10.0.0.12] = >                                    +-----[NODE3a 10.0.0.13]  > $ > How difficult would it be to have: >  > < > [DIALUP-MODEM]--[NODE3b]----------+-----[NODE1  10.0.0.11]= >                                    +-----[NODE2  10.0.0.12] = >                                    +-----[NODE3a 10.0.0.13]  > P > NODE3a is NODE3's first ethernet port, NODE3b is NODE3's second ethernet port. >  > H > When the DSL line fails, NODE3 would establish a PPP connection from aH > serial port to an alternate ISP and then act as a NAT router using its/ > second ethernet interface to talk to the lan.  >  > B > I know that there are NAT router software on Linux machines. How5 > complex/difficult would it be to port that to VMS ?   F I would think its not worth the effort, personally, but I really don'tE know how difficult it would be.  Why would you want to spend the time G on it (other than for the challenge or making something else work under G VMS) when you can easily purchase a dialup capable NAT device (Kingston C EtheRX, 3Com Officeconnext, Sonicwall -M models)?  The sonicwall in F particular can replace your NAT router and separate modem, and provide automatic failover.    Rich   ------------------------------  % Date: Mon, 11 Dec 2006 23:01:48 +0800  From: prep@prep.synonet.com ) Subject: Re: NAT router software on VMS ? 0 Message-ID: <873b7mxyqr.fsf@k9.prep.synonet.com>  / JF Mezei <jfmezei.spamnot@teksavvy.com> writes:   E > The telco here doesn't regard the DSL service as requiring the same D > uptimes as the POTS telephone service even though both usethe same	 > copper.  > D > Twice in the past month, I have had problems with DSL connectivity@ > because of a Bell Canada infrastructure problem (their ATM/BAS > machines). >  > Normally, one would have:  >  > < > [DSL-MODEM]-------[NAT ROUTER]----+-----[NODE1  10.0.0.11]= >                                    +-----[NODE2  10.0.0.12] = >                                    +-----[NODE3a 10.0.0.13]  > $ > How difficult would it be to have: >  > < > [DIALUP-MODEM]--[NODE3b]----------+-----[NODE1  10.0.0.11]= >                                    +-----[NODE2  10.0.0.12] = >                                    +-----[NODE3a 10.0.0.13]  > P > NODE3a is NODE3's first ethernet port, NODE3b is NODE3's second ethernet port. >  > H > When the DSL line fails, NODE3 would establish a PPP connection from aH > serial port to an alternate ISP and then act as a NAT router using its/ > second ethernet interface to talk to the lan.  >  > B > I know that there are NAT router software on Linux machines. How5 > complex/difficult would it be to port that to VMS ?   A With wich IP stack? The old UCX stuff did have a (semi)documented + hook point you could use, but that is gone.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 11 Dec 2006 01:02:31 -0800! From: "Ian Miller" <gxys@uk2.net> 2 Subject: Re: OpenVMS Weekly Audio Update - Show #1A Message-ID: <1165827751.910010.47090@80g2000cwy.googlegroups.com>   : Enough about the format, what do you think of the content?   ------------------------------  % Date: Sun, 10 Dec 2006 12:11:12 +0100 / From: Paul Sture <paul.sture.nospam@hispeed.ch> . Subject: OpenVMS Weekly Audio Update - Show #1J Message-ID: <paul.sture.nospam-8DF216.12111210122006@mac.sture.homeip.net>  : OpenVMS Weekly Audio Update is up on the OpenVMS.org site:  9 http://www.openvms.org/stories.php?story=06/12/09/2471408    It is in MP3 format.   --  
 Paul Sture   ------------------------------  # Date: Sun, 10 Dec 2006 14:31:52 GMT " From:   VAXman-  @SendSpamHere.ORG2 Subject: Re: OpenVMS Weekly Audio Update - Show #10 Message-ID: <00A5FFC6.7D281708@SendSpamHere.ORG>  | In article <paul.sture.nospam-8DF216.12111210122006@mac.sture.homeip.net>, Paul Sture <paul.sture.nospam@hispeed.ch> writes: >  > ; >OpenVMS Weekly Audio Update is up on the OpenVMS.org site:  > : >http://www.openvms.org/stories.php?story=06/12/09/2471408 >  >It is in MP3 format.     K Yes Paul, it is MP3 format.  It's a podcast Paul; albeit, the audio says it L is NOT a podcast.  Just keep in mind the famous quotation from Shakespeare's& "Romeo and Juliet" Act II, Scene 2. ;)  L Happily, it's not in a proprietary format ill-conceived by that schlock out-: fit in Redmond, Wash. thus, everyone can listen to it!  :D --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 11 Dec 2006 10:08:27 -0800! From: "Ian Miller" <gxys@uk2.net> 2 Subject: Re: sys v0600 patch for openvms 8.2 alphaC Message-ID: <1165860506.974725.109660@j72g2000cwa.googlegroups.com>    Its not there yet   1 ftp://ftp.itrc.hp.com/openvms_patches/alpha/V8.2/    ------------------------------    Date: 11 Dec 2006 04:30:48 -0800< From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>' Subject: Re: The Hole in Cerner's Logic B Message-ID: <1165840248.313557.303540@79g2000cws.googlegroups.com>   David J Dachtera wrote:   O > O.k. So, Cerner is dropping VMS in favor of HP-UX because of the whole Oracle  > thing. Fair enough.   E I suspect it is more HP not knowing how to cherish 'a good thing' and D not knowing how to leave that good thing alone. I have my reasons toG believe that HP strongly suggested to Cerner to move to HPUX. Cerner is B likely to say 'ok, we'll go to Unix... in fact we have a fine UnixC implementation already on Aix where we are not tossed around by the : self appointed fashion police from HP. It's sad. Very sad.   Cheers,  Hein.    ------------------------------    Date: 11 Dec 2006 07:25:21 -0800! From: "Ian Miller" <gxys@uk2.net> ' Subject: Re: The Hole in Cerner's Logic @ Message-ID: <1165850721.554042.9610@79g2000cws.googlegroups.com>   David J Dachtera wrote:  ....O > O.k. So, Cerner is dropping VMS in favor of HP-UX because of the whole Oracle Q > thing. Fair enough. No one's got the balls to turn up the heat under Oracle, so  > that's going nowhere.     5 Can you point at a reference for Cerner dropping VMS?    ------------------------------  % Date: Sun, 10 Dec 2006 23:51:58 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ' Subject: Re: The Hole in Cerner's Logic 8 Message-ID: <3a440$457ce401$cef8887a$32119@TEKSAVVY.COM>   David J Dachtera wrote: O > O.k. So, Cerner is dropping VMS in favor of HP-UX because of the whole Oracle Q > thing. Fair enough. No one's got the balls to turn up the heat under Oracle, so  > that's going nowhere.     I Has Cerner formally announced it would stop developping a VMS version of  H its product ? If that is the case, it would appear to me that this is a  HUGE blow to VMS.   K One of the few allowed remaining niches for was the medical field. If that  ! is gone, this is really bad news.   G BTW, NASA has DEC Alphas as workstations to support space missions for  L shuttle. They are considering going to 8086s running Linux. And that leaves L the door open to any industry standard computer maker. HP is this likely to F lose a big customer and Tru64 revenus that won't be replaced by HP-UX.    J It is interesting that HP use that 10 billion porting fund to help Oracle K port to HP-UX on that IA64 thing, but didn't get enough commitment to have  + the same happen for VMS on that IA64 thing.   K IA64 is killing VMS.  Pressure has to rise to get HP to port VMS to the 64  L bit 8086 ASAP to allow VMS to grow again and no longer have any clouds over  its future.    ------------------------------    Date: 10 Dec 2006 19:57:00 -0800< From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com> Subject: Re: Who called me? C Message-ID: <1165809420.916921.158460@j44g2000cwa.googlegroups.com>    roger wrote:A > OK.  I finally got it working...  Here is an example of calling / > TBK$SHOW_TRACEBACK() for everyone to enjoy...    Thanks for the example Roger!   D At times I work with Bryan and his friends. I adapted your example aF little for Bryan to show how one could return data back into the Cobol caller.  Example of that below.   Hein.    $ cc get_callers_module_name' $ link main,sub,get_callers_module_name 
 $ run main My Test: Calling my-sub  My Sub: Calling stack_dump_here  My Sub: Name = MY_SUB:MY_SUB My Test: Back from  my-sub   $ type main.cob  IDENTIFICATION DIVISION. PROGRAM-ID. my-main-program. PROCEDURE DIVISION.  MY_MAIN SECTION. MAIN. &     DISPLAY "My Test: Calling my-sub".     CALL "my-sub".)     DISPLAY "My Test: Back from  my-sub". 
     STOP RUN.    $ type  sub.cob  IDENTIFICATION DIVISION. PROGRAM-ID. my-sub.    DATA DIVISION. WORKING-STORAGE SECTION. 01 my-name PIC X(32).  PROCEDURE DIVISION.  MY_MAIN SECTION. MAIN. .     DISPLAY "My Sub: Calling stack_dump_here".>     CALL "get_callers_module_name" using by reference my-name.'     DISPLAY "My Sub: Name = ", my-name.      EXIT PROGRAM.     $ type GET_CALLERS_MODULE_NAME.c /*> ** This example program shows how to call TBK$SHOW_TRACEBACK()B ** to obtain program, module, routine, line info about the caller. **E **  Hein van den Heuvel, Based on an example written by: Roger Tucker  */   #define __NEW_STARLET      #include <starlet.h> #include <stsdef.h>  #include <lib$routines.h>  #include <descrip.h> #include <libicb.h>  #include <ssdef.h>   #include <stdio.h> #include <string.h>    static int line_counter; static char *name_pointer;C static unsigned long (*tbk$show_traceback)() = 0;   /* fill in just  once */   9 static unsigned long find_image_symbol(const char *image, '         const char *name, void **addr); A static void stack_dump(unsigned __int64 fp, unsigned __int64 pc);   7 void get_callers_module_name(char *callers_name_buffer)      { F     name_pointer = callers_name_buffer; /* stick in global variable */     line_counter = 0;       callers_name_buffer[0] =  0;        /* Get my callers FP & PC */%     struct _invo_context_blk context; (     lib$get_curr_invo_context(&context);?     unsigned long status = lib$get_prev_invo_context(&context); 9     if (!$VMS_STATUS_SUCCESS(status)) lib$signal(status); 4     unsigned __int64 fp = context.libicb$q_ireg[29];;     unsigned __int64 pc = context.libicb$q_program_counter;      stack_dump(fp,pc);       }    /*D ** This function is called by tbk$show_traceback for each line to be outputE ** Here we are just interested in the first line, and stash aways its  data.  */; static void my_action_routine (struct dsc$descriptor *line)      {      if (1 == line_counter++)	         {          char *p;+         strtok(line->dsc$a_pointer, " \t"); ;         strncpy  ( name_pointer,  strtok(NULL, " \t"), 30); '         strcat   ( name_pointer, ":" ); :         strncat ( name_pointer,  strtok(NULL, " \t"), 31 - strlen(name_pointer));	         }      }        /*0 ** stack_dump() - this prints out the stack-dump0 **      If first dynamically activates TRACE.EXE4 **      and finds the address of TBK$SHOW_TRACEBACK.< **      Note: TBK$SHOW_TRACEBACK says it needs the SP but by9 **      looking at the 7.1 source listings it's not used.  */  @ static void stack_dump(unsigned __int64 fp, unsigned __int64 pc)     {      unsigned long status;      if (!tbk$show_traceback)	         { @         status = find_image_symbol("TRACE","TBK$SHOW_TRACEBACK",.                 (void **)&tbk$show_traceback);=         if (!$VMS_STATUS_SUCCESS(status)) lib$signal(status); 	         }      if (tbk$show_traceback) 	         { 6         status = tbk$show_traceback(fp, /* FAULT FP */E                 0,                      /* FAULT SP - NOT USED!!!! */ 0                 pc,                     /* PC */9                 0,                      /* detail flag */ A                 &my_action_routine,     /* User action routine */                  0); =         if (!$VMS_STATUS_SUCCESS(status)) lib$signal(status); 	         }      }     9 static unsigned long find_image_symbol(const char *image, &         const char *name, void **addr)     { G     struct dsc$descriptor image_dsc = { 0, DSC$K_DTYPE_T, DSC$K_CLASS_S  };F     struct dsc$descriptor name_dsc = { 0, DSC$K_DTYPE_T, DSC$K_CLASS_S };  +     image_dsc.dsc$w_length = strlen(image); ,     image_dsc.dsc$a_pointer = (char *)image;  )     name_dsc.dsc$w_length = strlen(name); *     name_dsc.dsc$a_pointer = (char *)name;  F     return lib$find_image_symbol(&image_dsc, &name_dsc, (void *)addr);     }    ------------------------------  % Date: Sun, 10 Dec 2006 19:59:59 -0500 3 From: "Richard B. Gilbert" <rgilbert88@comcast.net> 3 Subject: Re: [DEFRAG] Umpteen OPEN files in logfile : Message-ID: <D7OdnePRN-YWMOHYnZ2dnUVZ_ternZ2d@comcast.com>    Peter 'EPLAN' LANGSTOEGER wrote:H > I run the DEFRAG (currently V2.9) on a daily (L2) and on a weekly (L1)G > basis for years now and I always wondered why in the logfile I always G > find myriads of "OPEN" files ("file was open; no action taken") which I > I still don't understand (because there is nothing running - except the : > Defragger of course - which would hold said files open).  I Which files are reported as open?  The operating system always has files  E open, typically on the system disk but some may be located elsewhere.    $ SHOW DEVICE/FILES 8 should show which files are open and what has them open.  H If you want to fully defragment the system disk you may have to build a H minimal system on another disk and boot from it in order to get the job  done.    ------------------------------   End of INFO-VAX 2006.681 ************************                                                                                                                                                                                                                                                                                                                                                          > 500 I never heard of the REST command.  Try HELP.> <<< REST 07 >>> 500 I never heard of the REST command.  Try HELP. 	 <<< PWDm2 >>> 257 "/disk$misc/decus" is current directory.	 <<< PWDe2 >>> 257 "/disk$misc/decus" is current directory. <<< REST 07 >>> 500 I never heard of the REST command.  Try HELP.E <<< TYPE I >>> 200 Type I ok.	 <<< PWD02 >>> 257 "/disk$misc/decus" is current directory. <<< TYPE I >>> 200 Type I ok. <<< REST 07 >>> 500 I never heard of the REST 