1 INFO-VAX	Tue, 28 Nov 2006	Volume 2006 : Issue 655       Contents:- Re: Batch Queue Jobs Stuck In Starting Status - Re: Batch Queue Jobs Stuck In Starting Status 0 Re: Cluster connection lost when one link fails?0 Re: Cluster connection lost when one link fails?0 Re: Cluster connection lost when one link fails?0 Re: Cluster connection lost when one link fails?0 Re: Cluster connection lost when one link fails?0 Re: Cluster connection lost when one link fails? Default sequential file size  Re: Default sequential file size  Re: Default sequential file size Future of Mozilla on VMS Re: Future of Mozilla on VMS Re: Future of Mozilla on VMS Re: interested in wombats 4 Re: libXrender 0.9.0 for OpenVMS ALPHA 8.2 or later. MMOV under OpenVMS 8.3 Re: MMOV under OpenVMS 8.3 Re: MMOV under OpenVMS 8.3 Re: Mouse: thumbwheel support ! Re: public-key ssh into VMS 7.3-1 ! Re: public-key ssh into VMS 7.3-1 / Re: renderext 0.9 for OpenVMS (X11 definitions) 4 Re: Thoughts on the book: DEC is dead, long live DEC4 Re: Thoughts on the book: DEC is dead, long live DEC Re: [sFTP] Current situation?  Re: [sFTP] Current situation?   F ----------------------------------------------------------------------    Date: 28 Nov 2006 07:53:58 -08000 From: "Andrew Butchart" <andrew@floatingbear.ca>6 Subject: Re: Batch Queue Jobs Stuck In Starting StatusB Message-ID: <1164729238.699910.262140@80g2000cwy.googlegroups.com>  F Well - I've finally managed to persuade our operations people to do anD ana/disk/repair (took several meetings).  Although there weren't anyG bad "errors" found, a lot of mis-reported files did relate to old jobs.   B Last night everything ran fine - my fingers are crossed for future runs.   F They are also looking into Peter's suggestion that he sent me off listA to check the patch level of the queuemanager but since we haven't E applied any patches to the OS since 1996 they're taking it slowly and 
 carefully.   Thanks everyone for the help.    Andrew B   ------------------------------    Date: 28 Nov 2006 08:33:37 -0800+ From: "Dave Gullen" <dave.gullen@gap.co.uk> 6 Subject: Re: Batch Queue Jobs Stuck In Starting StatusA Message-ID: <1164731617.196365.94390@80g2000cwy.googlegroups.com>   : Might be worth checking the size of the QMAN Journal file, SYS$SYSTEM:QMAN$JOURNAL.DAT.  F If it's very big, try this (undocumented in 7.1) command to shrink it.   $ MC JBC$COMMAND DIAG 7    Dave   Andrew Butchart wrote:  H > Well - I've finally managed to persuade our operations people to do anF > ana/disk/repair (took several meetings).  Although there weren't anyI > bad "errors" found, a lot of mis-reported files did relate to old jobs.  > D > Last night everything ran fine - my fingers are crossed for future > runs.  > H > They are also looking into Peter's suggestion that he sent me off listC > to check the patch level of the queuemanager but since we haven't G > applied any patches to the OS since 1996 they're taking it slowly and  > carefully. >  > Thanks everyone for the help.  > 
 > Andrew B   ------------------------------    Date: 27 Nov 2006 23:09:12 -0800/ From: "Volker Halle" <volker_halle@hotmail.com> 9 Subject: Re: Cluster connection lost when one link fails? A Message-ID: <1164697752.857153.96110@45g2000cws.googlegroups.com>    Malcolm,  A yes, I have seen a CLUEXIT crash last week, where - due to a link E failure - a node crashed with CLUEXIT, although there was a correctly F working second SCA channel available all the time (OpenVMS V7.3-2 with* VMS732_UPDATE-V0400 + some other patches).  C You should have a look at the CLUEXIT crashdump with commands like:    SDA> SHOW LAN/TIMESTAMP 
 SDA> PE VC SDA> PE CHAN SDA> PE TRACE/SORT% SDA> LAN INTERNAL_COUNTERS/DEVICE=EWB   G The most 'suspect' symptom in my case was, that one of the VCs remained % in status: 'Path Notif_Pend Closing'.   B If you are running a console manager application, also look at theD console messages (including the timestamps) of the systems involved.  @ To decode the EWB errlog entries on an ES40, you need to run SEAG (System Event Analyzer included in WEBES). These errors are most likely " 'possible duplex mismatch' errors.   --- , Volker Halle, Invenate GmbH, OpenVMS Support  # An OpenVMS crashdump analysis a day $ makes the Windows headaches go away.   ------------------------------  % Date: Tue, 28 Nov 2006 08:21:53 -0800 , From: "Malcolm Dunnett" <dunnett@mala.bc.ca>9 Subject: Re: Cluster connection lost when one link fails?  Message-ID: <456c61b4$1@flight>   ; "Volker Halle" <volker_halle@hotmail.com> wrote in message  ; news:1164697752.857153.96110@45g2000cws.googlegroups.com...   C > yes, I have seen a CLUEXIT crash last week, where - due to a link G > failure - a node crashed with CLUEXIT, although there was a correctly H > working second SCA channel available all the time (OpenVMS V7.3-2 with, > VMS732_UPDATE-V0400 + some other patches).  /   I'm at VMS732_UPDATE-V600 + SYS V10 + F11X V5    > E > You should have a look at the CLUEXIT crashdump with commands like:  >  > SDA> SHOW LAN/TIMESTAMP  > SDA> PE VC > SDA> PE CHAN > SDA> PE TRACE/SORT' > SDA> LAN INTERNAL_COUNTERS/DEVICE=EWB  > I > The most 'suspect' symptom in my case was, that one of the VCs remained ' > in status: 'Path Notif_Pend Closing'.   D   That appears to be the same thing I am seeing. The timestamp showsB SCA messages being sent and received on EWA just before the crash,A but nothing on EWB for  the previous 30 seconds. The lan counters @ show EWB as down and EWA as UP. I see the same status on the VC:  * MALVM9 VC Summary 27-NOV-2006 07:53:38.94:I            - Channels -                                            -Pipe   Quota-     Total    Total L   Node    ECS Open Dead     VC Open Time         VC Close Time    Mgmt  Max  Cur    Bytes   Errors  Status M   ----    --- ---- ----     ------------         -------------    ----  ---     ---    -----   ------  ------K   MALVM8    2    2     0  26-NOV 22:48:25.14  (No time)             0   31    16   20600508      11  Open PathK   MALVM3    1    1     1  26-NOV 22:51:06.59  27-NOV 07:53:08.11    0   16  - 1  387807903     187  Path Notif_Pend Closing K   MALVM9    1    1     0  26-NOV 22:48:14.91  (No time)             0    8   1        318       0  Open Path   < The console log on MALVM3 shows it lost connection to MALVM9@ at the time of the crash. MALVM3 and MALVM9 each have two votes,( MALVM8 has one - so it MALVM3 and MALVM99 lose communication (but both could still speak to MALVM8) 8 I guess it's the luck of the draw which one will CLUEXIT    B > To decode the EWB errlog entries on an ES40, you need to run SEAI > (System Event Analyzer included in WEBES). These errors are most likely $ > 'possible duplex mismatch' errors. > J   That's likely the case ( I don't have WEBES installed on these systems).L I'm sure the actual error was triggered by the switch going through its POST8 and doing various weird things to the port as it did so.  ' Have you logged a call on this problem?    ------------------------------  % Date: Tue, 28 Nov 2006 08:23:23 -0800 , From: "Malcolm Dunnett" <dunnett@mala.bc.ca>9 Subject: Re: Cluster connection lost when one link fails?  Message-ID: <456c620e$1@flight>   - "R.A.Omond" <Roy@Omond.net> wrote in message  . news:ekh88t$hiq$1$830fa795@news.demon.co.uk...  ? > Can you check what the values of the SYSGEN parameters are in 
 > both ES40s:  >  > TIMVCFAIL  > LAN_FLAGS  >   K Those parameters are at their default values (1600 and 0) on both systems.     ------------------------------    Date: 28 Nov 2006 09:21:32 -0800/ From: "Volker Halle" <volker_halle@hotmail.com> 9 Subject: Re: Cluster connection lost when one link fails? B Message-ID: <1164734492.295427.215300@45g2000cws.googlegroups.com>   Malcolm,  % I have done some additional research:   ? The code to close the VC will wait, until all packets have been F transmitted. In our case, I still find packets in the Transmit PendingE queue of the failing LAN interface. As long as they won't be finished D or aborted, the VC will not leave the 'Notif_Pend Closing' state and2 therefore can't re-form using the working channel.   SDA> SHOW LAN/FULL/DEV=EWB ... < Transmit pending queue  88AFA3D8  Status:  Valid, 4 elements ...   + SDA> VALI QUE/LIST <use address from above>  ...    SDA> SHOW PORT) SDA> SHOW PORT/ADDR=PE_PDT/VC/NODE=MALMV3  ...  SDA> READ SCSDEFE SDA> EXA VC_MALMV3+VC$A_BUF_FIRST ! should be zero if XMT cache empty   ? The address of the VCRP shown should be one of the VCRPs in the  Transmit Pending Queue.   A Both TIMVCFAIL and LAN_FLAGS are at their default values as well.   F I can't see how changing TIMVCFAIL would influence this situation. But9 setting LAN_FLAGS bit 12 and/or 13 might change things...   F We have not logged a call, as the systems are not up-to-date regardingF OpenVMS patches. But based on our mutual experience, we may reconsider this.   D Let's see the TIMVCFAIL and LAN_FLAGS settings suggestions from Roy.   --- , Volker Halle, Invenate GmbH, OpenVMS Support  # An OpenVMS crashdump analysis a day $ makes the Windows headaches go away.   ------------------------------    Date: 28 Nov 2006 10:34:03 -0800/ From: "Volker Halle" <volker_halle@hotmail.com> 9 Subject: Re: Cluster connection lost when one link fails? C Message-ID: <1164738843.280218.250210@h54g2000cwb.googlegroups.com>    Malcolm,  A please also provide your driver, driver version and RECNXINTERVAL  setting.   In our case:   RECNXINTERVAL=60   SDA> SHOW LAN/FULL/DEV=EWA  @                -- EWA Device Information 23-NOV-2006 10:38:21 -- ..< Driver version     00000001.08020111    Device1 code address 00000000< Device1 version    00000000.00000000    Device2 code address 000000008 Device2 version    00000000.00000000    LAN code address 8162FA000 LAN version        00000001.08020172    DLL type CSMACD0 Device name                 EW_DE500    MOP name  EWAC MOP ID                           203    HW serial               Not  supplied8 HW version                  02000041    Promiscuous mode  OFF    ! $ anal/ima/sel=(ident,link,build) , sys$loadable_images:sys$ewdriver_de500ba.exe. SYS$COMMON:[SYS$LDR]SYS$EWDRIVER_DE500BA.EXE;1 "X-4"   9-FEB-2005 13:35:57.58  "XA99-0060111008"    Volker.    ------------------------------  % Date: Tue, 28 Nov 2006 10:50:03 -0800 , From: "Malcolm Dunnett" <dunnett@mala.bc.ca>9 Subject: Re: Cluster connection lost when one link fails?  Message-ID: <456c8485$1@flight>   ; "Volker Halle" <volker_halle@hotmail.com> wrote in message  < news:1164734492.295427.215300@45g2000cws.googlegroups.com...  
 > Malcolm, > ' > I have done some additional research:  > A > The code to close the VC will wait, until all packets have been H > transmitted. In our case, I still find packets in the Transmit PendingG > queue of the failing LAN interface. As long as they won't be finished F > or aborted, the VC will not leave the 'Notif_Pend Closing' state and4 > therefore can't re-form using the working channel. >  > SDA> SHOW LAN/FULL/DEV=EWB > ... > > Transmit pending queue  88AFA3D8  Status:  Valid, 4 elements > ...   ;  Yes, I see all the same things (In my case the queue has 7 3 elements in it and the VCRP points to one of them).   =   I suppose bits 12 and 13 of the lan flags may help. They're 9 not documented in VMS 7.3-2 but I guess that doesn't mean  they aren't honored.  9   Is the relationship between TIMVCFAIL and RECNXINTERVAL : significant, eg would lowering the first or increasing the second help?  8   If this requires a call to be logged it would probably: be better if you do it, you clearly understand the details& of what's happening better than I do.    ------------------------------    Date: 28 Nov 2006 08:21:16 -0800% From: "RJDurkee" <RJDurkee@yahoo.com> % Subject: Default sequential file size B Message-ID: <1164730876.091173.233380@14g2000cws.googlegroups.com>  E My OpenVMS is a little rusty as I haven't worked on it much lately. I G was wondering how VMS decides how big to allocate a sequential file. Is A there a default size parameter somewhere that is used  if I don't  specify anything? # For example, if I used this command   + search dummyfile.dat abcdef /out=result.lis   G How does VMS decide on the number of blocks to allocate for result.lis?    Thanks Rich   ------------------------------    Date: 28 Nov 2006 10:26:25 -0800< From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>) Subject: Re: Default sequential file size B Message-ID: <1164738385.392131.148950@45g2000cws.googlegroups.com>  E It uses the default RMS settings, looking at the process value first,  the system value next.D If neither is set it uses 2 buffers worth, rounded up to the nearest cluster of course.G The buffersize default "MULTI BLOCK COUNT" was 16 blocks, and now is 32  blocks. 3 There is also a device default ( $SHOW DEV / FULL )   	 $SHOW RMS    hth, Hein van den Heuvel.           RJDurkee wrote: G > My OpenVMS is a little rusty as I haven't worked on it much lately. I I > was wondering how VMS decides how big to allocate a sequential file. Is C > there a default size parameter somewhere that is used  if I don't  > specify anything? % > For example, if I used this command  > - > search dummyfile.dat abcdef /out=result.lis  > I > How does VMS decide on the number of blocks to allocate for result.lis?  >  > Thanks > Rich   ------------------------------  % Date: Tue, 28 Nov 2006 13:33:01 -0500 ' From: Dave Froble <davef@tsoft-inc.com> ) Subject: Re: Default sequential file size 9 Message-ID: <G7ydnbeUdP1xHfHYnZ2dnUVZ_q6dnZ2d@libcom.com>    RJDurkee wrote: G > My OpenVMS is a little rusty as I haven't worked on it much lately. I I > was wondering how VMS decides how big to allocate a sequential file. Is C > there a default size parameter somewhere that is used  if I don't  > specify anything? % > For example, if I used this command  > - > search dummyfile.dat abcdef /out=result.lis  > I > How does VMS decide on the number of blocks to allocate for result.lis?  >  > Thanks > Rich >   F A disk has a default EXTEND size, in 512 byte blocks.  Each time more G storage is needed, a chunk of this size will be allocated to the file.  I when the file is closed, any unused storage will be released.  Note that  ? allocation is by clustersize.  If the extendsize is 5, and the  I clustersize is 2, you'd start with 5 blocks.  If only 3 blocks are used,  C then when the file is closed, the final file allocation would be 4  + blocks.  The fifth block would be released.    --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 28 Nov 2006 05:31:26 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ! Subject: Future of Mozilla on VMS 6 Message-ID: <77c3$456c0fe6$cef8887a$3777@TEKSAVVY.COM>  K With Mozilla officially "mature", has VMS management announced any project  2 with regards to continued browser support or VMS ?  C Will VMS engineering be tasked to port SeaMonkey, the "unofficial"  N continuation of Mozilla suite, or will they focus on Firefox and Thunderbird ?    N Looks to me like Mosaic is the only viable browser on VMS in the long term :-)   ------------------------------    Date: 28 Nov 2006 02:59:53 -0800! From: "Ian Miller" <gxys@uk2.net> % Subject: Re: Future of Mozilla on VMS C Message-ID: <1164711593.680470.162030@j72g2000cwa.googlegroups.com>   C Back on Nov 1st I noted the following update to the OpenVMS Roadmap   $ OpenVMS eBusiness roadmap (slide 23)B - SWB 1.7-13, based on Mozilla 1.7.13, is now available on OpenVMSG Alpha and Integrity. Shown starting in 2007 is Secure Web Browser based  on Firefox and Thunderbird  9 http://www.openvms.org/stories.php?story=06/11/01/8671580   F You could contact someone in HP to see if they will say anything else.   ------------------------------    Date: 28 Nov 2006 09:17:21 -0800+ From: "madcrow" <madcrow.maxwell@gmail.com> % Subject: Re: Future of Mozilla on VMS C Message-ID: <1164734241.175517.164200@l39g2000cwd.googlegroups.com>    Ian Miller wrote: E > Back on Nov 1st I noted the following update to the OpenVMS Roadmap  > & > OpenVMS eBusiness roadmap (slide 23)D > - SWB 1.7-13, based on Mozilla 1.7.13, is now available on OpenVMSI > Alpha and Integrity. Shown starting in 2007 is Secure Web Browser based  > on Firefox and Thunderbird > ; > http://www.openvms.org/stories.php?story=06/11/01/8671580  > H > You could contact someone in HP to see if they will say anything else.  E Hmm. I believe that the X11 versions of Mozilla (or at least Firefox) 0 use GTK. Has GTK+ 2 ever been ported to OpenVMS?   ------------------------------    Date: 28 Nov 2006 02:13:00 -0800! From: "Ian Miller" <gxys@uk2.net> " Subject: Re: interested in wombatsC Message-ID: <1164708780.280245.230260@l12g2000cwl.googlegroups.com>    Lee K. Gleason wrote:   : >   I hear they also live in epergnes (whatever they are). >   C An epergne is described as a centerpiece which may be used to serve  food, as well.  G For more than you want to know see http://www.thegavel.net/Mayappr.html    or for Pictures , http://www.cranberryglass.co.uk/epergnes.htm  H I think it unlikely that wombats are often found in an unbroken epergne.   ------------------------------  % Date: Tue, 28 Nov 2006 09:09:48 -0500 - From: "John E. Malmberg" <wb8tyw@qsl.network> = Subject: Re: libXrender 0.9.0 for OpenVMS ALPHA 8.2 or later. ; Message-ID: <_7ednaAGLP-j3vHYnZ2dnUVZ_umdnZ2d@adelphia.com>    JOUKJ wrote: > John E. Malmberg wrote:  > H >> libXrender is the client side library for the Xrender extension.  It I >> is required for building the current Cairo graphics package, which is  4 >> used by GTK+ 2.x for displaying on X-11 displays. >>K > Do you have a working version of cairo and gtk2? I'm a little stuck with  J > cairo giving me wrong results except for the ones that use Xrender (but , > I cannot do that on a "native" VMS-screen.  I I have cairo working what appears to be good enough for my purposes.  It  F will just take me some time to get all the components packaged up for 
 distribution.   G I am seeing some tests fail because a few pixels are out of place.  It  H looks like rounding errors.  A search on google shows that LINUX on x86 G runs IEEE floating point at a different precision then the rest of the  1 world.  I do not know if that is the real reason.   D And at least one test does not agree with the library source and is  failing because of that.  D GTK+ 2.x requires true-color X11 displays on any platform.  I could 7 display them fine on a DS10 and a Radeon graphics card.   / My last tests with cairo were with Cairo 1.0.2.    > [snip] > J >> It requires the renderext package if you want to build it from sources.J >> The build enviroment also requires the current GNV and OpenVMS 8.2 and  >> the HP C 7.x compiler.  >>  >> Currently this package is at: >>Y >> http://encompasserve.org/~malmberg/GNV/JEM-AXPVMS-LIBXRENDER-V0009--1.PCSI$COMPRESSED   >> > I > Appart for the building procedure (GNV<->native VMS descrip.mms files)  = > Did you make other changes than what I did in my version at C >       http://nchrem.tnw.tudelft.nl/openvms/software1.html#XRENDER   G About the only changes that I can remember is that I had to modify the  I #include directives to not have a path, because otherwise the C compiler  G   was looking for them only in the X11 subdirectory of the GDK package.   E Also, I added a return 0 statement to a routine so that the compiler    would stop issuing a diagnostic.  9 I do not have time now to do a more extensive comparison.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------    Date: 28 Nov 2006 11:28:55 +0100R From: pmoreau@ath.cena.fr (Patrick MOREAU, DTI Athis ex CENA, Tel: 01.69.57.68.40) Subject: MMOV under OpenVMS 8.3 ! Message-ID: <I$tCnxtuN74g@sinead>    Hi all,   I Have you tested MMOV under Alpha VMS 8.3 ? I suppose it is not supported   but is it working ?    TIA    Regards,   Patrick  --O =============================================================================== N pmoreau@ath.cena.fr              ______      ___   _          (Patrick MOREAU)4 DSNA/DTI/SDER (ex CENA)         / /   /     / /|  /|J Athis-Mons France              / /___/     / / | / |   __   __   __   __  N BP 205                        / /         / /  |/  |  |  | |__| |__  |__| |  |N 94542 ORLY AEROGARE CEDEX    / /   ::    / /       |  |__| | \  |__  |  | |__|N http://www.ath.cena.fr/~pmoreau/              http://membres.lycos.fr/pmoreau/O ===============================================================================    ------------------------------  % Date: Tue, 28 Nov 2006 05:33:07 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> # Subject: Re: MMOV under OpenVMS 8.3 7 Message-ID: <ac435$456c104a$cef8887a$3777@TEKSAVVY.COM>   = Patrick MOREAU, DTI Athis ex CENA, Tel: 01.69.57.68.40 wrote: K > Have you tested MMOV under Alpha VMS 8.3 ? I suppose it is not supported   > but is it working ?   L Where can I download the latest release ? And is it pre-built, of not, what L do I need to build it ? I would be more than willing to test it. Is there a # test movie I can use as reference ?    ------------------------------  % Date: Tue, 28 Nov 2006 16:44:00 +0100 ( From: JOUKJ <joukj@hrem.nano.tudelft.nl># Subject: Re: MMOV under OpenVMS 8.3 3 Message-ID: <456C5940.1020708@hrem.nano.tudelft.nl>   = Patrick MOREAU, DTI Athis ex CENA, Tel: 01.69.57.68.40 wrote: 	 > Hi all,  > K > Have you tested MMOV under Alpha VMS 8.3 ? I suppose it is not supported   > but is it working ?  > I On my XP1000 running VMS8.3 and MMOV2.2-1 I can run MMOV$DECSOUND load a   .wav file and play it.                     Jouk   ------------------------------  # Date: Tue, 28 Nov 2006 15:02:11 GMT # From: "FredK" <fred@nospam.dec.com> & Subject: Re: Mouse: thumbwheel support1 Message-ID: <TbYah.2901$RH6.531@news.cpqcorp.net>   ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  2 news:a5ec4$456c22ce$cef8887a$27273@TEKSAVVY.COM... > FredK wrote:L >> By default in Motif CTRL-up/down arrow activates the scroll bar.  Try it I >> with the normal KB and see.  All I did was to effectively implement a  F >> Linux hack from a while back - to make the thumbwheel generate the  >> up/down arrow keys. > F > How dangerous would it be to change the driver to generate the page K > up/page down instead of up/down arrow key events ? (by dangerous, I mean  * > if one coudl see user interface issues). > L > The impression I have so far (this is my first time using a computer with M > a working thumbwheel) is that moving just one line up/down doesn't seem to  1 > provide "action" to make it really that useful.   K The USB version has/had this in it.  If you moved the wheel fast enough it  J generated a page up/down instead of up/down.  I didn't find it useful.  I G imagine that an accelleration routine could be implemented to generate  F increasingly large numbers of up/down rather than the 1:1 it does now.   ------------------------------    Date: 28 Nov 2006 08:39:59 +01006 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)* Subject: Re: public-key ssh into VMS 7.3-1, Message-ID: <456bf5df$1@news.langstoeger.at>  ] In article <ekfkou$o3j$1@shell.monmouth.com>, colonel@monmouth.com (Hoary Hairy Hoax) writes: D >I have installed SSH 1.3 (OpenSSH 0.97) on an Alpha running OpenVMS >7.3-2.   B What for? TCPIP has its own SSH (server and also client) built in.I And PSC (Multinet, TCPware) has also its own SSH (server and also client) K built in and even a SSH standalone product (to be run on top of TCPIP/UCX).   C >         Try as I may, I cannot set up a public-key login into the B >Alpha.  The documentation and web pages on VMS SSH are patchy andE >sometimes contradictory.  The VMS OPENSSL tool lets me generate keys < >but not to put them in the right format in the right files.  J Forget them for a while. Concentrate on TCPIP (if login is what you want).  B >Would someone please explain to me what files are needed, in what! >directories, and in what format?   I TCPIP V5.4 (IIRC ECO3) and above have SSH working out of the box. Earlier H version are known to require fiddling in the config files (but work too)G On TCPIP V5.3 you need a special TCPIP_SSH add on product (alpha only?) J which you need to remove before you can upgrade TCPIP to V5.4 (or higher).  F Enter TCPIP$CONFIG and enable the SSH server (which does also generate* a key pair if there is not already one)...   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------    Date: 28 Nov 2006 09:15:05 -05008 From: colonel@monmouth.com (A Deaf Heart, A Loose Liver)* Subject: Re: public-key ssh into VMS 7.3-1- Message-ID: <ekhg99$b3o$1@shell.monmouth.com>   B In <06112715393723_2020028D@antinode.org>, sms@antinode.org wrote: > I >    TCPIP V5.4 includes SSH.  I've installed OpenSSL (not relevant here, , > I believe), but not OpenSSH.  Around here: >  > alp $ tcpip show version > = >   HP TCP/IP Services for OpenVMS Alpha Version V5.4 - ECO 6 F >   on a COMPAQ Professional Workstation XP1000 running OpenVMS V7.3-2  G We too have 5.4, no E.C.O.  I got the SSH 1.3 from Compaq, which is now  Hewlett-Packard.   > alp $ ssh "-V"H > alp$dka0:[sys0.syscommon.][sysexe]tcpip$ssh_ssh2.exe: SSH Secure ShellG   OpenVMS (V5.5) 3.2.0 on COMPAQ Professional Workstation  - VMS V7.3-2   ( Substantially the same on our Alpha 600.  ? >    Yes, the key file formats differ between the TCPIP SSH and  > OpenSSH/OpenSSL. > I >    From which kind(s) of system(s) are you trying to "ssh" into the VMS 7 > system?  (And what sort of "ssh" is available there?)   E Mainly from Linux.  We also use Solaris, AIX, and Cygwin.  We have no C trouble with public-key login between different systems except with  OpenVMS.  F >    Were you planning to generate keys on the VMS system, or were youB > planning to use keys which were generated on a foreign/differentE > system?  (I generated my keys on a VMS system and converted them on G > Solaris systems using "ssh-keygen -X" or "ssh-keygen -i" or by manual ? > editing, but it should be possible to go the other way, too.)   C I tried generating the keys on VMS with the "openssl" tool. (I have D not found any ssh-keygen tool on our VMS system.) But I need to knowA which directories to put the files in, how the files should look, C and what permissions they should have.  I have experimented freely.   J >    Knowing nothing originally, I generated keys on both VMS and Solaris,J > and compared the key file formats to discover how they differed, and howI > to convert from one to the other.  Later, I learned of the "ssh-keygen" F > options which allow the Solaris SSH package to convert the VMS-styleE > ("SSH2-compatible") key files.  I don't see anything obvious in the E > TCPIP "ssh_keygen -h" output which suggests a conversion capability : > similar to that in the Solaris (OpenSSL) implementation.  E On our Solaris and Linux systems, ssh-keygen has no -h option.  But I E think I could put the keys in a format that works if I knew the names C of the files that are to hold them, and the directories that are to 8 hold the files!  The manual does not seem to mention it.   -:-  	Nothing changes,  	  But nothing's the same....  	Dance and love, 	  Eat and fight 	At the DISCO, 	  Disco Delight!    			--Bob Burden  --   Col. G. L. Sicherman home: colonel@mail.monmouth.com  work: sicherman@att.com ( web: <http://www.monmouth.com/~colonel/>   ------------------------------  % Date: Tue, 28 Nov 2006 08:45:26 -0500 - From: "John E. Malmberg" <wb8tyw@qsl.network> 8 Subject: Re: renderext 0.9 for OpenVMS (X11 definitions); Message-ID: <rIWdnQaB_cQWoPHYnZ2dnUVZ_radnZ2d@adelphia.com>    JOUKJ wrote: > John E. Malmberg wrote:  >  >>                 Render  >>           X Render Extension  >>             Version 0.8 >>               2003-4-21 >> > [snip] > E >> If you do not have DECWindows-Motif installed and the current GNV  7 >> installed, then this library is probably not of use.  >>I > And the code generated with it can only be applied on an X-server with  J > the XRender extension. None of the DECWindows-Motif have this extension I > (AFAIK). To use it you need an other X-server (i.e. a linux system) on  / > which the extension is more or less standard.   H Actually Cairo, a component of GTK+ 2.x will not product any X11 output H even on systems that do not support the Xrender unless it is built with  libXrender.   > I could not find any way to remove this dependency from Cairo.  J And with out X11 output on Cairo, GTK+ is useless for graphics on OpenVMS.  ? And libXrender will not build with out renderext being present.    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------   Date: 28 Nov 2006 06:55:20 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>= Subject: Re: Thoughts on the book: DEC is dead, long live DEC ? Message-ID: <DTiotGxQ0bj6-pn2-EU4BBohVt2nL@dave2_os2.home.ours>   D On Mon, 27 Nov 2006 23:10:49 UTC, Dave Froble <davef@tsoft-inc.com>  wrote:   > Bob Koehler wrote:h > > In article <4t0pp4F11i1kvU3@mid.individual.net>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:D > >> There was a person who posted a very big explanation of how hisD > >> company was buying VAXen for internal use in Switzerland.  OnceI > >> the machines arrived in Switzerland the company would do an internal I > >> corporate transfer to a branch in India.  India had never signed the H > >> trade restriction agreement against the USSR.  The Indian branch ofG > >> the company would then sell the box to Russia.  The guy was rather L > >> proud of the fact that they were pulling this off in front of everyone.L > >> So, makes you wonder why the Russians would even consider cloning whichD > >> was bound to cost more and always leave them behind the currentG > >> technology in the rest of the world.  Of course, it also brings to I > >> mind the quote, "With friends like these....." in regards India. :-) G > >> Couldn't be trusted then, can't be trusted now.  Some things never  > >> change. > > L > >    We can all wonder why, but we can't change the record which shows the& > >    USSR did build clones of VAXen. > > K > >    We can also assume it was in part to keep the smuggled-in VAXen to a 5 > >    sufficiently low volume to keep off the radar.  > >  > J > Once you become dependent upon something, you cannot chance having your I > supply cut off.  Yeah, they were getting them.  But that could change,  I > and they couldn't take that chance.  That's fine.  One more expense to  I > run them into the ground.  In the end, it was money, not weapons, that  9 > decided things.  As it should be.  (Somebody tell 'W').   A When you emulate/clone something, it's always useful to have the  D original to test against. Plus, you learn a lot more about how it's % done when you re-implement something.    --   Cheers - Dave W.   ------------------------------  # Date: Tue, 28 Nov 2006 18:31:40 GMT ' From: ChrisQuayle <nospam@devnul.co.uk> = Subject: Re: Thoughts on the book: DEC is dead, long live DEC 8 Message-ID: <gg%ah.58248$163.32303@newsfe6-gui.ntli.net>   Stephen Hoffman wrote: > ChrisQuayle wrote: >  >> Bob Koehler wrote:  >  > ...  > F >>>    You never heard of IBM?  Sometimes to link two of their systemsG >>>    together, they'd sell you software for both of their systems and I >>>    a VAX to sit between.  They sure as heck didn't include networking G >>>    with their mainframe OS, they didn't even include timesharing or 2 >>>    batch processing, everything was an option. >> >>J >> Irrespective of what was happening with mainframes, it was quite clear H >> to anyone with a clue that once the internet achieved critical mass, 8 >> tcp/ip networking would become the dominant standard. >  >  > H >    Irrespective of what is now obvious with hindsight, at the time it K > wasn't clear if IP or OSI (or X.25 or...) was going to be the choice for  D > the network.  It also wasn't clear that the Internet was going to G > explode -- well, prior to Tim Berners-Lee's little invention, though  G > once we saw that...  It also wasn't clear that the widespread use of  J > DECnet would be replaced by IP -- there were private networks with tens G > of thousands of DECnet nodes around at that time.  (This is also the  E > downside of an installed base that I've mentioned -- as for when a  K > product transitions from an extra-cost product to a table-stakes product  J > is always an interesting choice and a tricky decision.)  OpenVMS TCP/IP J > got its start in the fall of 1988, which was just after the time of the H > big-seven TLD transition.   (I'm old enough to have had an ARPA email D > address, something now long unknown to most Internet users.)  The J > Internet marketing started in earnest around 1985 -- SNMP didn't appear I > until 1987.  There were all of 100000 hosts in 1989, up from 28,000 in  C > 1987.  Major Internet backbones were T1 lines.  Though yes, once  K > something reaches critical mass, it's catch-up time -- which is itself a  H > real problem -- if you're not already apace in the game.  OpenVMS was L > obviously late to the game, first having TCP/IP (UCX) in the fall of 1988. >   I That may be true, but the only place I ever used Decnet seriously was at  I Dec, where the corporate network was said to be the largest in the world  H at the time. The functionality was amazing - copy a file on the command H line from anywhere in the world. Vax notes etc, so, so good. Everywhere I else I worked ran tcp/ip networking. Experimented with decnet at home on  G an ancient 730, vms 4.3, flaky R80 etc, but never got comfortable with  F all the concepts (terminology ?) and setup in the same way that I did I later with ip networking. It just seemed overcomplex and the real killer  F was that it didn't talk to anything else in the lab, other than other F Vaxen, or via kermit and serial lines. By that time, I had a scrapped I Sun 3/50 box, coerced into life and that was really the beginning of the  D end, so far as Vax home hackery was concerned. When the R80 finally F died, it was replaced with a qniverter, qbus box, rqdx controller and B rd53 to run vms, so we didn't lack enthusiasm. The last vax was a D uVaxII/GPX, but the Sun 3 did 12 ppm on Tex, against 4 ppm with the E uVaxII. The uVax / X interactive response time was a real problem. I  F guess i should have tried the mips based boxes, but it was not to be. H While systems like pdp were genuine hackers machines and built Dec into B a major force in terms of customer goodwill, the vax was always a ' corporate system, with prices to match.   @ I think unix mail was the first ip networking application, 1989 A timescale, with vt220 terminals, but by then, running on Sun 3...    Chris    ------------------------------    Date: 28 Nov 2006 08:56:18 +01006 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)& Subject: Re: [sFTP] Current situation?, Message-ID: <456bf9b2$1@news.langstoeger.at>  V In article <op.tjo8ememtte90l@hyrrokkin>, "Tom Linden" <tom@kednos-remove.com> writes:J >             psftp does not drop the version number as does native ftp on
 >W2K or XP  - RUsure that this is a function of the client? 8 I tend to think, that the server doesn't send them then.  F At least, I've seen a TCPIP$FTP_NO_VERSION logical, but I so far don't= know if it affects the FTP or also the sFTP server or client.   ? I've seen a TCPIP sFTP server sending the whole file attributes I (like saying ls -l in the TCPIP sftp client) but I don't know why for the H same PC client, one TCPIP server does it and the other TCPIP server not.  D >I have discovered.  I am using hgftp as the server if that matters.  4 Does hgftp do sFTP now? I didn't know this so far...   --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  % Date: Tue, 28 Nov 2006 06:20:20 -0800 * From: "Tom Linden" <tom@kednos-remove.com>& Subject: Re: [sFTP] Current situation?) Message-ID: <op.tjqcr6h3tte90l@hyrrokkin>   ? On Mon, 27 Nov 2006 23:56:18 -0800, Peter 'EPLAN' LANGSTOEGER    <peter@langstoeger.at> wrote:   : > In article <op.tjo8ememtte90l@hyrrokkin>, "Tom Linden"  ! > <tom@kednos-remove.com> writes: K >>             psftp does not drop the version number as does native ftp on  >> W2K or XP > / > RUsure that this is a function of the client? : > I tend to think, that the server doesn't send them then.  = That would imply that hgftp thinks psftp on W2K is a VMS box.  > H > At least, I've seen a TCPIP$FTP_NO_VERSION logical, but I so far don't? > know if it affects the FTP or also the sFTP server or client.  > A > I've seen a TCPIP sFTP server sending the whole file attributes K > (like saying ls -l in the TCPIP sftp client) but I don't know why for the J > same PC client, one TCPIP server does it and the other TCPIP server not. > F >> I have discovered.  I am using hgftp as the server if that matters. > 6 > Does hgftp do sFTP now? I didn't know this so far... >        --  E Using Opera's revolutionary e-mail client: http://www.opera.com/mail/    ------------------------------   End of INFO-VAX 2006.655 ************************