1 INFO-VAX	Sat, 29 Oct 2005	Volume 2005 : Issue 604       Contents:J Alphaserver 4100 and Mylex DAC960 RAID controller documentation/utilities?N Re: Alphaserver 4100 and Mylex DAC960 RAID controller documentation/utilities? Compile on VAX/VMS Disk Optimizer 2.4* Re: Lack of response from HP field service Re: LASER symbiont4 OT: Digital PWS 500 + OpenVMS CD's  for sale on EBAYA Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....  VMS? Re: VMS? Re: Writing an OpenVMS ACP Re: Writing an OpenVMS ACP Re: Writing an OpenVMS ACP Re: Writing an OpenVMS ACP Re: Writing an OpenVMS ACP  F ----------------------------------------------------------------------  % Date: Sat, 29 Oct 2005 12:39:23 +0100 $ From: Sampsa Laine <news@sampsa.com>S Subject: Alphaserver 4100 and Mylex DAC960 RAID controller documentation/utilities? 0 Message-ID: <2005102912392316807%news@sampsacom>  G I just picked up one of these on ebay to run VMS on for fun. I have NO  G idea how to configure the Mylex RAID controller that came with it, can  @ someone point me towards a place I can get documentation and/or  drivers/utilities.   Sampsa   ------------------------------  % Date: Sat, 29 Oct 2005 12:01:34 -0400 , From: "Richard Tomkins" <tomkinsr@istop.com>W Subject: Re: Alphaserver 4100 and Mylex DAC960 RAID controller documentation/utilities? 6 Message-ID: <1130601854_862@spool6-east.superfeed.net>  G The Mylex requires a floppy disk to be run on the system and within the L application on the floppy you can configure and control the RAID. There is a< second floppy with the driver and management software on it.   A word of caution.  K There are Mylex controllers and there are Digital controllers made by Mylex L and each has their own firmware. Mylex floppies do not work with the DigitalJ and vice versa. If you ignore this warning you will just have so much dead1 hardware on your hands and it is not recoverable.   I Next item, the firmware cannot be changed between the two, we had special J SCSI chips put on our Mylex that worked properly and thus differently withK SCSI reconnect and other minor problems. Again, ignore this warning and you @ will just have so much dead hardware on your hands and it is not recoverable.  K The requisite firmware and driver may be found on the Alpha firmware update A CD, but I have not looked in years so I cannot remember for sure.    rtt     1 "Sampsa Laine" <news@sampsa.com> wrote in message * news:2005102912392316807%news@sampsacom...H > I just picked up one of these on ebay to run VMS on for fun. I have NOH > idea how to configure the Mylex RAID controller that came with it, canA > someone point me towards a place I can get documentation and/or  > drivers/utilities. >  > Sampsa >       O ----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==---- S http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups K ----= East and West-Coast Server Farms - Total Privacy via Encryption =----    ------------------------------    Date: 29 Oct 2005 10:11:14 -0700 From: usenet@tenox.tc  Subject: Compile on VAX/VMS B Message-ID: <1130605874.881452.73540@z14g2000cwz.googlegroups.com>   Hi,   D Could somebody with VAX/VMS 6.1 compile for me a small programm in CD and send back the binary? Or perhaps give me access to the system? I? only have access to Alpha OpenVMS at the moment. :-( This thing  requires Curses library.  $ If so please contact tenox@tenox.tc.   Thanks!  -a   ------------------------------  % Date: Sat, 29 Oct 2005 12:50:05 +0000 1 From: andyg <andyg.1xo1wz@nomx.sysadminforum.com>  Subject: Disk Optimizer 2.4 1 Message-ID: <andyg.1xo1wz@nomx.sysadminforum.com>   H After trying to defrag a volume (inex of 90!) I can start the script andB then try and monitor this script, after about 10 seconds a warning@ window appears "DFG-W-MONNOVOL ***** disk is not being currently
 defragmented"   ? Ive tried different settings but this is just playing ball why? F The disk is currently in a mirrored Raid 5 and im running VMS 7.2-1 on a 3 node cluster. 
 Any ideas? Cheers Andy      --   andyg " posted via http://sysdminforum.com   ------------------------------  % Date: Sat, 22 Oct 2005 19:35:24 -0700  From: Z <Z@no.spam> 3 Subject: Re: Lack of response from HP field service & Message-ID: <IFC6f.37$xk2.21@fe06.lga>   Alan Greig wrote: E > Is this typical now or have I just been unlucky? Two weeks with no  , > response let alone a repair is ridiculous.  , In my experience, that's not typical at all.   ------------------------------  % Date: Sat, 22 Oct 2005 16:00:57 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net> Subject: Re: LASER symbiont + Message-ID: <435AA889.4E9DF454@comcast.net>    Albrecht Schlosser wrote:  >  > GeorgeC wrote: > > Hi all,  > > 1 > > I am running OpenVMS V7.2-2 on an Alpha 1200.  > > A > > I have a few LAT print queues using the LASER print symbiont. C > > These printers are connected to the network via Emulex (NETQUE) D > > print servers.  The printers are fitted with ethernet interfacesC > > so they are capable of being connected to the network directly.  > > H > > I want to connect the printers directly on the network and eliminateD > > the print servers. I can do that by converting the LAT queues toJ > > TCPIP queues and use the default TCPIP print symbiont TCPIP$TELNETSYM.6 > > In fact, I do this already with some of my queues. > > However there is a snag... > > K > > User written applications rely on queue form definitions (LANDSCAPE,..) 6 > > that do not work with the default TELNET symbiont. > I > I never heard (read) that TCPIP$TELNETSYM can't handle form definitions J > (correctly). AFAIK TELNETSYM doesn't differ from LATSYM in this respect.  A Depends on the UCX version. It was known to be broken in V4.2 and D earlier, and I think may have been "fixed" in some ECO for V4.3. Not 100% certain on that...    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  % Date: Sat, 29 Oct 2005 16:51:46 +0200 Q From: "Jimmi Aakjr (8310)" <bitnissen#spammer_format_your_harddisk#@hotmail.com> = Subject: OT: Digital PWS 500 + OpenVMS CD's  for sale on EBAY 8 Message-ID: <6337m1lb10adbq1f2hoaak9t285aign3cv@4ax.com>   Hi Guys   0 I have my AlphaStation PWS 500 for sale on Ebay    starting bid is EUR 30  
 Best regards     Jimmi, Denmark  < x-posted to : comp.sys.dec.micro, comp.sys.doc, comp.os.vms,, comp.os.linux.alpha,linux.debian.ports.alpha  Y http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&ssPageName=ADME:L:LCA:US:11&item=5824543901     8 DEC (Digital Equipment) Alpha Personal Workstation 500a. Configuration: CPU 500Mhz, 2MB B-Cache 
 320 MB Memory  S3 (VGA-connector) Video card. 1x 4.3 GB Disk (Wide-SCSI)
 CDROM, Floppy ( Ethernet 10/100 Mbit Twisted-Pair (RJ45) OpenVMS 7.3-1 is installed  * A complete set of OpenVMS cd's is included< (12/2002) OpenVms Alpha Software Products Library, Disc 1-9 = (12/2002) OpenVms Alpha Software Online Documentation Library  (Windows) Disc 1-3= (12/2002) OpenVms Alpha Software Online Documentation Library  (OpenVMS) Disc 1-30 (08/2002) Compaq TCP/IP services Version 5.3 MUP# (09/2001) Compaq System Tools  1.4  2 (06/2002) hp OpenVms Documentation v7.3-1 Disc 1-29 (06/2002) hp OpenVms Alpha Operation System Version 7.3-1 & (04/2001) OpenVms Freeware v5 Disc 1-2& (01/1999) OpenVms Freeware v4 Disc 1-2 (01/1996) OpenVms Freeware v3 8 (06/2002) HP OpenVms Alpha Layered Products 7.3-1 disc 1) (06/2002) Hp OpenVMS Open Source Tools v1 , (05/2002) Alpha Systems Firmware Update V6.2? OpenVms 7.3-1 is installed (No licenses). It is possible to run @ TRU64(Unix), Windows NT 4, Windows 2000 (beta) and linux on this machine.+ The machine is in good technical condition. ?  The PWS is located in Denmark (Aarhus), pick-up is possible or B delivery by GLS or DHL (within the EU). Buyer pays shipping costs.! (EUR 50-80 depending on country).    ------------------------------   Date: 29 Oct 2005 05:44:11 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>J Subject: Re: UCX performance on VMS 6.2 - Unexpected rise in CPU usage....? Message-ID: <DTiotGxQ0bj6-pn2-0XKbjHdDRutd@dave2_os2.home.ours>   E On Thu, 27 Oct 2005 09:48:23 UTC, "Lawrie" <stroker_ace@hotmail.com>   wrote:    > I don't poll every 1ms though. > G > I only poll the sockets once each time a message arrives on the input D > queue (max rate 30 per second) to see if any sockets are ready forI > writing. I then try and send write all unsent data to each socket until C > it blocks. I then return  to by blocking read on the input queue.  >  > Cheers >  > Lawrie  F Aha, that makes sense. I obviously misunderstood '(wiith a timeout of 5 1/1000 of a second)'. BTW how big are these messages?    --   Cheers - Dave W.   ------------------------------    Date: 29 Oct 2005 01:45:38 -0700 From: "Zach" <netrek@gmail.com> 
 Subject: VMS? B Message-ID: <1130575538.015015.28440@g43g2000cwa.googlegroups.com>  4 Is there a free version of VMS that will run on x86?  F Can someone recommend a good free book/guide/tutorial on VMS as both a user and sys admin perspective?    Zach   ------------------------------  % Date: Sat, 29 Oct 2005 12:30:13 +0100 6 From: "Alex Daniels" <AlexNoSpamDaniels@themail.co.uk> Subject: Re: VMS? 4 Message-ID: <43635d46$0$362$da0feed9@news.zen.co.uk>  + "Zach" <netrek@gmail.com> wrote in message  < news:1130575538.015015.28440@g43g2000cwa.googlegroups.com...6 > Is there a free version of VMS that will run on x86?  K Not a native one, but you get an old VAX from Ebay for next to nothing, or  $ an old Alpha for only a little more.  M Alternativly there are a couple of VAX emulators that are freeware, SIMH and  F TS10. Then there is also Charon-VAX which is available in hobbyist or  commerial versions.   M If it's for personal use, then check out the hobbyist license program, for a   free license. F http://h71000.www7.hp.com/openvms/license_software_programs/index.html  H > Can someone recommend a good free book/guide/tutorial on VMS as both a! > user and sys admin perspective?  >  > Zach  D The documentation set is outstanding. http://h71000.www7.hp.com/doc/   Alex     ------------------------------    Date: 29 Oct 2005 03:45:04 -07002 From: "Max" <max.jarvis@gb.schneider-electric.com># Subject: Re: Writing an OpenVMS ACP C Message-ID: <1130582704.670833.217750@g44g2000cwa.googlegroups.com>   G Hi Larry, thanks for your kind reply. Actually I'm new to VMS, and what F I'm doing is porting a real-time system to VMS from RSX-11M (father ofF VMS). Don't know whether you remember RSX... In RSX, writing an ACP isE an efficient way of implementing a client-server arrangement. What we D do is this: invent an imaginary device (we call them pseudo-devices)F and write a driver and an ACP for it. The driver does nothing but wakeF up the ACP process, which is the server. The client processes ('tasks'G in RSX) communicate with the server by doing QIOs to the pseudo-device. E Data transfer can be in either direction, depending on whether it's a F read QIO or a write QIO. This is the functionality I need to implementG in VMS. From my reading of the manuals, the alternative to using an ACP D seems to be to use mailboxes, but that doesn't sound as efficient: IF reckon you'd need a mailbox for the server and one for each client (ofG which there might be dozens), and each transaction would involve 4 QIOs C (client sends request, server receives request, server sends reply, ? client receives reply) whereas with an ACP there's only one QIO + involved. Any advice gratefully received...    Max.     Larry Kilgallen wrote:z > In article <1130517970.075291.268550@o13g2000cwo.googlegroups.com>, "Max" <max.jarvis@gb.schneider-electric.com> writes: > G > > I need to write an ACP (Ancillary Control Processor) for OpenVMS. I F > > can't seem to find anything in the OpenVMS manuals about how to doK > > this, e.g. how does the ACP get launched, how does it get woken up when J > > a QIO is sent to it, how does it pick up the QIO, how does it completeJ > > the QIO, etc. Is anyone willing to let me see source code for a VMS orI > > OpenVMS ACP? Or can anyone tell me where to find the information in a 	 > > book?  > E > The best source of information is the VMS Source Listings Kit which C > contains various examples.  Note that ACPs only work with certain B > device drivers designed to work with them, so typically you also2 > need to understand how to write a device driver. > 9 > But you knew that if you _really_ have to write an ACP.  > D > You could also try describing your problem to look for suggestionsB > regarding other ways to solve it.  But if your employer requires> > you to write an ACP then they should spend the money for the > source listings kit.   ------------------------------    Date: 29 Oct 2005 03:49:18 -07002 From: "Max" <max.jarvis@gb.schneider-electric.com># Subject: Re: Writing an OpenVMS ACP C Message-ID: <1130582958.031245.196570@f14g2000cwb.googlegroups.com>   G Hi Bob, thanks for your reply. Have you any ideas about the name of the B advanced device driver book you mention? Is it one of the standardE OpenVMS manuals provided by HP, or something published by a different B publisher? I think we may already have an old set of VMS listings,D possibly in microfiche, so will have a look there for the sources of> the mag tape ACP or something like that. Thanks again.... Max.   ------------------------------  % Date: Sat, 29 Oct 2005 07:15:25 -0400 . From: JF Mezei <jfmezei.sopamnot@teksavvy.com># Subject: Re: Writing an OpenVMS ACP , Message-ID: <436359A5.FB6A93E6@teksavvy.com>  
 Max wrote:I > in VMS. From my reading of the manuals, the alternative to using an ACP D > seems to be to use mailboxes, but that doesn't sound as efficient:  H You cab use DECNET where a single process can create a network object toD which any client can connect and send/receive transactions each with their own channel.  D (the server has one mailbox channel to DECNET which receives controlH messages to tell it when someo is connecting, disconnecting etc, and canM then use that message to conform a connection request, exchange of data etc).   B This also allows you to have client processes on different nodes.     F Anbother facility is ICC, the intra-cluster communication system. ThisA is not as clean as DECNET, but much more efficient. Any node in a  cluster can participate.   ------------------------------  # Date: Sat, 29 Oct 2005 14:05:15 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> # Subject: Re: Writing an OpenVMS ACP > Message-ID: <vkL8f.140041$G8.63060@text.news.blueyonder.co.uk>  
 Hello Max,  C It all depends on the latency target you're attempting to meet. How B "real-time" is your application in terms of the response times and) throughputs you need to reliably achieve?   H In my experience an Alpha (or Integrity) system is so much faster than aJ PDP11 that the overhead of mailboxes is not that great in terms of elapsed time.   I The big changes from a PDP11 are of course the virtual address space of a H process and the ability to have many things happening in parallel with aI multi-CPU box. Both of those should cause you to think how to lay out the L application as part of your porting process. You may find that you no longerK need multiple processes (rough equivalent of RSX tasks). ASTs and so on you L should already be thoroughly familiar with. You may be able to implement theJ whole thing as a series of AST routines with a noddy top level. Threads (aJ lightweight scheduling mechanism - look up DECthreads) can also be useful.  I My preferred mechanism for minimal latency message passing on a real-time I VMS system is to use the lock manager to deliver message notifications in L the lock block and to have all the processes mapping a global section, usingG global sections for message passing and for the common data structures. L Again the lock manager has an important part to play in synchronising accessG to common data structures and flags. The key here is to think multi-CPU B (thus parallel activities) and have a fine granularity on the dataF structures so that you don't hold off processing on everything becauseJ you're updating something that only a few parts of the code are interestedL in. It's not too difficult to implement and it's reasonably easy to test and debug.  G I wouldn't recommend the device driver / ACP route unless you have some H really good reasons to go there. VMS has plenty of alternatives for whatH you've described that you want to do. A noddy device driver as a messageK passing agent may be useful to preserve device characteristics if you don't K want to make changes to the source code. I suspect you'll be making changes J anyway, so assuming it to be well-structured code using subroutines for IOA then changes should (in principle) be limited to a few key areas.   H One big question (based on experience) - what kind of IO devices are youL going to need and can you find their equivalent on the Alpha / Integrity boxL that you're planning to move to? Interrupt latency on Alpha / Itanium can beK worse than on a PDP11 with Q-bus (or microVAX II with Q-bus) because of all I the buffering in the PCI bus. The Q-bus is a pretty basic bus (by current H standards) and thus the CPU is closely involved with it - which leads toL very good low interrupt latency characteristics. That really matters in some real-time applications.  --     Hope this helps, Colin. ) colin DOT butcher AT xdelta DOT co DOT uk E It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------  % Date: Sat, 29 Oct 2005 11:08:10 -0400 2 From: "Stanley F. Quayle" <squayle@insight.rr.com># Subject: Re: Writing an OpenVMS ACP / Message-ID: <4363581A.28909.26E19EAB@localhost>   " On 29 Oct 2005 at 3:45, Max wrote:D > Hi Larry, thanks for your kind reply. Actually I'm new to VMS, andB > what I'm doing is porting a real-time system to VMS from RSX-11M > (father of VMS).  E Ah, the problem's one of different definitions.  I have done lots of  F real-time programming on VAX, and have never written an ACP.  Or even  a device driver.  G > From my reading of the manuals, the alternative to using an ACP seems ? > to be to use mailboxes, but that doesn't sound as efficient:    = Mailboxes can be pretty quick -- since they're just a memory  F transfer.  You also get record-oriented operation, since you can find - out the size of the data from the other side.   B You could also use the Pseudo Terminal facility, which looks more  like a serial line.   < And there's other options for DECnet/IP/cluster things, too.  A In the server you'd want to have an AST delivered when a message  F arrives.  This allows you to make everything "interrupt-driven", with ? a main-line code section that wakes up from time to time to do   bookkeeping stuff.  F > I reckon you'd need a mailbox for the server and one for each client# > (of which there might be dozens)    D No big deal.  You can get gigs and gigs of RAM on Alpha and Itanium 
 systems.  :-)   A Your server's mailbox can have a group or system logical name so   everyone can find it...   F Have each client include its MBA device name in each message, and you F won't to have to manage this on the server side.  If you can't send a 2 response to the client's mailbox, he's gone away..  B > and each transaction would involve 4 QIOs (client sends request,E > server receives request, server sends reply, client receives reply) 4 > whereas with an ACP there's only one QIO involved.  @ Perhaps, but the equivalent in VMS would be writing to a shared ? memory area (global section in VMS), and have everyone wait on  # something like a shared event flag.   > This makes your entire set of applications WAAAAY too closely-C coupled.  I recommend you let the OS do most of the heavy lifting,  A unless latency is a big issue.  Which you can't quantify without   testing it, anyway.   
 --Stan Quayle  Quayle Consulting Inc.  
 ----------- Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363 3 8572 North Spring Ct., Pickerington, OH  43147  USA 0 stan-at-stanq-dot-com       http://www.stanq.com) "OpenVMS, when downtime is not an option"    ------------------------------   End of INFO-VAX 2005.604 ************************