INFO-VAX Sun, 21 Jan 2007 Volume 2007 : Issue 41 Contents: Re: BACKUP - how much of the tape has been used Re: File rename fault tolerance?? Re: How long to really setup a VMS system ? Re: How long to really setup a VMS system ? Re: RUN SYS$SYSTEM:DCL on Itanium VMS 8.2-1 Re: SIMH hints and tips please ---------------------------------------------------------------------- Date: Sat, 20 Jan 2007 23:17:25 +0100 From: Paul Sture Subject: Re: BACKUP - how much of the tape has been used Message-ID: In article <87ejpp7r9w.fsf@k9.prep.synonet.com>, prep@k9.prep.synonet.com wrote: > David J Dachtera writes: > > > The short answer is no, there is no data provided back to the system > > to say how far in which direction, etc. the tape was moved in > > total. That's about the only way to figure what would have been left > > at the end of a save operation. > > It is some years since I read the tech manual, but with the DEC DLTs > you could get the drive to return the position `down the tape'. I > suspect it is not used at all in the driver though. In the early '90s I looked at a time limited demo of a software product which sped up tape backups by caching. It also gave tape info not available with standard VMS, including how long the tape was, and how much of it had been written (in feet IIRC). I only tried it on a TK50, which of course didn't support hardware compression. I can't remember what the product was called. -- Paul Sture ------------------------------ Date: 20 Jan 2007 19:45:10 -0800 From: "Hein RMS van den Heuvel" Subject: Re: File rename fault tolerance?? Message-ID: <1169351110.275607.299440@l53g2000cwa.googlegroups.com> Paul Sture wrote: > In article <1169184824.339739.262890@11g2000cwr.googlegroups.com>, > "Hein RMS van den Heuvel" wrote: : > > RMS does a remove and an enter operation, so the failure mode is that > > the file appears neither with the old nor the new name. > > > > While this seems strange, it prevents some really awful problems of > > deleting important files, such as when version limits exist. : > Curiosity kicks in here. Say that A.A has a version limit of 5, and > there are currently versions 1 to 5. You now rename another file to A.A. > > At what point in the renaming process is A.A;1 deleted? Before or after > A.A;6 is entered in the directory? That's why the remove is done first. Imagine a.a;1 is renamed to ;6 If it was entered before removing, then the file should be deleted due to version limit. This would surprise users, and surprises are to be avoided. As it is first removed, then entered, so there are never more than 5 versions. No limit exceeded. This is the only 'interesting case'. If an other file is entered/in the directory with a newer version, whether through create or rename ($enter) then a.a.1 should be deleted. The user deserves whatever they get on failure (imho). Hein. ------------------------------ Date: Sat, 20 Jan 2007 17:20:12 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: How long to really setup a VMS system ? Message-ID: <45b2959c$0$49204$14726298@news.sunsite.dk> JF Mezei wrote: > OK, I know, insert CD, boot the alpha from CD and follow the prompts and > you have VMS running in about an hour. > > But in reality, how long does it REALLY take to get a virgin system > fully loaded with all the prerequisite stuff, configured, tested and > ready to go into production without any more improvements being done to > it ? > > (I am talking about full IP stack, extra utilities like ethermon loaded, > SNMP/syslog configured to talk to the rest of the corporate network, > email, various automated system monitoring tasks , handling if system > failures and ensuring the systartup_vms and shyshutdown procedures are > complet and robost etc etc. > > In other words, how long does it take to get a system from fully naked > to fully dressed in formal wear, realy to go a lights out fully > automated state ? > > (as opposed to a customer starting to be able to use it even though the > system may still need nightly reboots to test new procedures etc) Ignoring shortcuts like adding a node to a cluster or creating a system based on a clone disk, then I would say that it would take 3-10 work days to get everything setup and quick tested depending on how much and what needs to be installed. Arne ------------------------------ Date: Sat, 20 Jan 2007 18:15:13 -0500 From: JF Mezei Subject: Re: How long to really setup a VMS system ? Message-ID: <45b2a2f5$0$8745$c3e8da3@news.astraweb.com> Arne Vajh=F8j wrote: > Ignoring shortcuts like adding a node to a cluster or > creating a system based on a clone disk, then I would say > that it would take 3-10 work days to get everything > setup and quick tested depending on how much and what > needs to be installed. But this would assume that you have all media/files handy and all manuals= =20 downloaded etc, right ? If you need to start hunting for patches, MIBs,=20 find out where to donwload UNZIP/ZIP and all other essentials to get your= =20 system properly populated, you could still do it in 10 days ? Just trying to find Mozilla on the VMS web site isn't too obvious. ------------------------------ Date: Sat, 20 Jan 2007 17:15:21 -0500 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: RUN SYS$SYSTEM:DCL on Itanium VMS 8.2-1 Message-ID: <45b29479$0$49204$14726298@news.sunsite.dk> > DEC provided at least four at different times: DCL, MCR, DECShell (a > Bourne shell), and a POSIX shell (subset of ksh). Documentation on > how to write one does exist, I've messed around with the idea of > writing one, but I don't think it was in the standard doc set, nor > was it a cookbook: "how to write a VMS CLI" document. > > I think GNV is providing a bash, but I don't know if it's as a full > blown CLI, or as a program that can be started from DCL. It could even be argued that a program running in P0 space was more equivalent to a *NIX shell than a CLI running in P1 space. Arne ------------------------------ Date: Sat, 20 Jan 2007 19:24:08 +0100 From: martin@radiogaga.harz.de (Martin Vorlaender) Subject: Re: SIMH hints and tips please Message-ID: <45b25e48.524144494f47414741@radiogaga.harz.de> Paul Sture wrote: > koehler@eisner.nospam.encompasserve.org (Bob Koehler) wrote: >> martin@radiogaga.harz.de (Martin Vorlaender) writes: >> > I have SimH running on my WinXP notebook with local DECwindows. [...] >> > I had it on a (now dead) notebook that ran Linux, too - look for >> > "TUN/TAP bridge". >> >> Ahh, cool. I'll have to try that. > > With your Mac? Though I don't have a Mac, IMHO OS X should be unixy enough... Googling for "tun tap Mac OS X" yields as the first hit http://www-user.rhkh.uni-kl.de/~nissler/tuntap/ titled "tun/tap driver for Mac OS X", so it should be doable. http://www.retrocomputinggeek.com/index/retrowiki/SIMHNetworking/ describes how to use the tun/tap devices. cu, Martin -- | Martin Vorlaender | OpenVMS rules! UNIX is user friendly. | work: mv@pdv-systeme.de It's just selective about | http://www.pdv-systeme.de/users/martinv/ who its friends are. | home: martin@radiogaga.harz.de ------------------------------ End of INFO-VAX 2007.041 ************************