INFO-VAX Wed, 07 Feb 2007 Volume 2007 : Issue 76 Contents: adding Poly Center Archive Files Re: Anyone have a copy of the DCL book they'd part with? Re: Blank page when printing with setup module Re: Blank page when printing with setup module Re: DVD writer for DS10L ? Re: If you are attending the Encompass event in Canada Re: Intel prepares to kill off the Pentium 4 Re: Intel prepares to kill off the Pentium 4 Re: Intel prepares to kill off the Pentium 4 Re: Mark Daniel and/or Hein van den Huevel or anyone really Re: MSA1500 on VMS? MySQL on IA64 Problem Re: Odd Telnet output Program looping Re: Program looping Re: Searching FMS documentation in pdf Re: SPANNING BACKUP TAPES Re: SPANNING BACKUP TAPES Re: SPANNING BACKUP TAPES Re: SPANNING BACKUP TAPES Re: SPANNING BACKUP TAPES Re: SPANNING BACKUP TAPES Re: Wanted: Alisashare / Alisatalk for VMS / VAX ---------------------------------------------------------------------- Date: 7 Feb 2007 06:59:15 -0800 From: "Beach Runner" Subject: adding Poly Center Archive Files Message-ID: <1170860355.438000.186640@s48g2000cws.googlegroups.com> Greetings: I have a site that for some reason does not have the monthly and yearly .cpd files. Advise archive does not create them. The only answer I've received so far was to reinstall them, but there must be another alternative. How to recreate the archival files, Monthly and Yearly? Thanks ------------------------------ Date: 7 Feb 2007 04:13:29 -0800 From: "mb301@hotmail.com" Subject: Re: Anyone have a copy of the DCL book they'd part with? Message-ID: <1170850409.702886.43900@p10g2000cwp.googlegroups.com> On 7 Feb, 02:49, wins...@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) wrote: > In article , heal...@aracnet.com writes: > >DeanW wrote: > >> I gave my copy of "Writing Real Programs in DCL" to a geographically > >> remote co-worker, figuring I'd just pick up another copy. That book > >> turns out to be unobtanium. > > >> Does anyone have an extra copy that I might be able to purchase, or > >> know where there is one? > > >I'm afraid not, that book seems to be one of the most sought after VMS > >titles out there if the prices I've seen it offered at are any indication. > >Thankfully I own a copy, and I'm not about to part with it. > > >I think it's time for a Third Edition :^) Of course I'd also like to see a > >Second edition of Alan Winston's OpenVMS Webserver book! > > Thanks for the kind words. > > Digital Press has asked me for a second edition. > I'm not really sure I've got the time. > > Also, I keep finding problems with SWS 2.11, and I'd feel like a fraud if I > tried to update the book when the Apache installation I manage isn't in good > shape. I hope to get this sorted out soonish, though. > > -- Alan There appears to be a few on the second hand section. http://www.amazon.com/Vax-Vms-Writing-Programs-Digital/dp/1555580238 http://www.elsevier.com/wps/find/bookdescription.librarians/677449/description#description http://www.amazon.co.uk/gp/offer-listing/1555581919/ref=dp_olp_2/203-8424089-1971938 ------------------------------ Date: Wed, 07 Feb 2007 07:00:30 GMT From: John Santos Subject: Re: Blank page when printing with setup module Message-ID: Malcolm Dunnett wrote: > "Richard B. gilbert" wrote in message > news:45C91410.2080002@comcast.net... > | Malcolm Dunnett wrote: > | That's not quite the way it works! > | > | The symbiont issues a formfeed if it sees anything in the setup module > | that could change the initial print position. > > Yes, that is a more precise description. I still maintain that I should > be able to decide whether or not my setup modules change the initial > print position and I can then request a "top of form" in said setup > modules if I deem it necessary - I don't need the symbiont making > (what are in this case) incorrect decisions. > > | The symbiont does not > | understand HP PCL, only ANSI X3-64 escape sequences. > > That's a holdover from the days when DEC made printers which spoke > ANSI and wasn't particularly interested in making it easy to use > other peoples (especially HPs) printers with VMS. Thus my little > dig that HP should remove this vestige of proprietariness from > VMS (or at least add a setup option to disable it ) so that it > would be easier to use their printers with VMS. > > | There is a way to > | tiptoe around this but I have not needed it for several years now and I > | have forgotten the details. > > The shenanigans I mentioned at the beginning of this thread are > what I thought were the way to tiptoe around the symbiont, but they > no longer seem to work (I could swear I used them successfully some > years ago) > I'm only responding here because a bunch of other people have already responded and no on has mentioned this... I could be wrong, but I think the "shenanigans" are a two-parter. You need to start off with the initial escape sequence, containing the "VMS" part, then do all your other setup sequences, and then, right at the very end, immediately before your text starts (which could be ASCII, PCL, postscript, or whatever), finish up with the final part of the "shenanigans". The way it works, as I understand it, is the print symbiont sees the VMS escape sequence, strips it out, passes on everything else blindly and unaltered to the printer, and then when it sees the terminating sequence, strips that out too and goes back into normal mode. Since it starts the job at position (0,0), and doesn't see anything to change that (since it passes on everything else blindly), it is still there at the end of the setup module, and so it doesn't attempt to reset the printer position by sending an . HTH. > | > | And it's not really HP's O/S. As we all know far too well, it's an > orphan! > | > > Sadly that appears to be true, but that shouldn't stop me from > challenging HP to prove otherwise. > > ps. I found a workaround in this particular situation by putting the > "paper source select" escape sequence in a page setup module. The symbiont > apparently doesn't try to be cute about what appears in a page > setup module and in this case it doesn't hurt to specify the > paper source at the top of each page. However this wouldn't > work in all situations, sometimes you need to just do the setup > once per job. The workaround I usually use in that case is to > print the setup information as the first file in the print job, but > in this case I needed to make the tray select transparent because > I can't change the application that generates the print job. > -- John Santos Evans Griffiths & Hart, Inc. 781-861-0670 ext 539 ------------------------------ Date: Wed, 7 Feb 2007 08:29:13 -0800 From: "Malcolm Dunnett" Subject: Re: Blank page when printing with setup module Message-ID: <45c9fe4a$1@flight> "JF Mezei" wrote in message news:1191c$45c964c0$cef8887a$7276@TEKSAVVY.COM... > How is your computer connected to the printer ? Serial ? Lat ? TCPIP ? > Telnet ? > TCP/IP (Multinet stream symbiont) > Before blasting the symbiont, you may wish to try to dump your data "raw" > to the printer to see how it reacts to it. Works fine that way. Also works fine if the sequences are part of the file or in a file thats part of the print job (as I mentioned in the previous note) > It is quite possible that the printer automatically ejects a page that > contains "some" content when you issue the reset comand just after that. > Nope. > If the symbiont does not understand HP's proprietary sequences, then it > won't recognize ANY of the sequences, so the argument that the symbiont > would insert a form feed between two HP proprietary sequences seems > doubtful. I don't understand what you're trying to say here. The symbiont is programmed to recognize ANSI control sequences and know which ones move the print position. This is because DECs laser printer offerings (eg LN03) used ANSI control sequences. It doesn't recognize PCL control sequences (because DEC didn't use them) so when it encounters a "printable" character in a PCL sequence it wrongly assumes that character has caused the printer to move off of the home position so it sends a form feed before the first file in the job to reset things. It's not inserting a form feed between the PCL sequences it's inserting the FF after all of the characters in the setup module(s). If the symbiont could be told to trust the user that the setup modules wouldn't affect the home position ( or possibly they affect in a way the user desires ) then the problem would go away. Seems like a relatively simple programming change - just add a switch to turn this "feature" off, preferably within the form definition, eg: DEFINE/FORM/NOFF_AFTER_SETUP ------------------------------ Date: 1 Feb 2007 20:22:17 +0100 From: vaxinf@chclu.chemie.uni-konstanz.de (Eberhard Heuser-Hofmann) Subject: Re: DVD writer for DS10L ? Message-ID: <45c23de9$1@merkur.rz.uni-konstanz.de> In article , JF Mezei writes: >I have a DS10L with the right bay free. > >I have a DSL10L with a SCSI card, but both bays full. > >What would be the recommended solution to get some sort of DVD/CD writer >on >one of these machines ? One where I could burn CDs and DVDs preferably ? > >I looked in some of the on-line catalogues and couldn't find SCSI based >DVD >writers. (since those seemed to focus on consumer market). > >Are there any DVD writers that actually fit inside the DS10L ? Would those >>be the types designed for laptops ? > There are two solutions: 1.) A slimline DVD writer (I'm just testing one for DVDwrite/Cdrecord/ "copy/rec"). Does a DS10L have an IDE-Interface build in??? 2.) Use an external burner with an Acard-SCSI-IDE converter. Hope this helps Eberhard ------------ And now a word from our sponsor ------------------ For a quality usenet news server, try DNEWS, easy to install, fast, efficient and reliable. For home servers or carrier class installations with millions of users it will allow you to grow! ---- See http://netwinsite.com/sponsor/sponsor_dnews.htm ---- ------------------------------ Date: Wed, 7 Feb 2007 07:38:04 -0500 From: "Neil Rieck" Subject: Re: If you are attending the Encompass event in Canada Message-ID: <45c9c646$0$7429$9a6e19ea@news.newshosting.com> "Paul Sture" wrote in message news:paul.sture.nospam-923892.12370907022007@mac.sture.homeip.net... > In article , John Santos > wrote: > [...snip...] > > Well, if it's conspiracy theories you are after, I now have definitive > proof that the NASA moon landing was faked. > [...snip...] > > Paul Sture > Segue for a Lunar Lander simulation that I just stumbled across. Working AGC (Apollo Guidance Computer) and DSKY. http://www.eaglelander3d.com/ Neil Rieck Kitchener/Waterloo/Cambridge, Ontario, Canada. http://www3.sympatico.ca/n.rieck/ http://www3.sympatico.ca/n.rieck/docs/eagle_lander_3d_activities.html ------------------------------ Date: 7 Feb 2007 13:24:10 GMT From: bill@triangle.cs.uofs.edu (Bill Gunshannon) Subject: Re: Intel prepares to kill off the Pentium 4 Message-ID: <52u27qF1pircrU1@mid.individual.net> In article <45c769da$0$177$157c6196@dreader1.cybercity.dk>, "Dr. Dweeb" writes: > > "Bill Gunshannon" wrote in message > news:52n5t1F1o87deU3@mid.individual.net... >> In article , >> david20@alpha2.mdx.ac.uk writes: >>> In article <52kndeF1m9o2dU1@mid.individual.net>, bill@cs.uofs.edu (Bill >>> Gunshannon) writes: >>>>In article <45c47807$0$7466$9a6e19ea@news.newshosting.com>, >>>> "Neil Rieck" writes: >>>>> >>>>> "Tad Winters" wrote in message >>>>> news:Xns98CBE9BC1B3C5staffordnospamwinter@130.81.64.196... >>>>> [...snip...] >>>>>> >>>>>> So, wouldn't you think it would have been better to port to AMD to >>>>>> start >>>>>> with? AMD's 64 bit processor provides better performance for the >>>>>> 32-bit >>>>>> crowd at little additional cost over their 32-bit processor and, from >>>>>> my >>>>>> experience, at less cost than Intel's 32-bit offering. I would also >>>>>> believe their 64-bit processor sales are probably ahead of Itanium. >>>>>> >>>>> >>>>> Intel makes both the x86-64 and Itanium processors. Intel will save >>>>> tons of >>>>> money by killing Itanium and HP should be able to leverage this guilt >>>>> by >>>>> saying "we paid to port OpenVMS to Itanium and now you're killing it; >>>>> so now >>>>> you must pay for the OpenVMS port to x86-64". >>>> >>>>And Intel's response would be, "Hahahahahahahahahahahahaha". >>>> >>>>> >>>>> This is why it would be a good idea to sell OpenVMS to Intel. >>>> >>>>Why would they want it? An OS with no more future than the only >>>>processor >>>>it runs on. >>>> >>> Maybe because it is still the OS running their FABS. >>> >> >> If that is still truei (and, personally, I doubt it), it's a safe bet >> it isn't running on Itanium. >> > > Your doubt is unfounded. It runs on Alpha now. Irrelevant, really. Alpha is as dead as VAX as far as HP is concerned. HP says Itanium is VMS's future and Itanium is all but dead. > >> And the last version of VMS for the VAX that runs those FABs was >> completed long ago. >> bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 7 Feb 2007 18:00:32 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Intel prepares to kill off the Pentium 4 Message-ID: <52uie0F1poafpU1@mid.individual.net> In article <14c$45c9f96d$cef8887a$10256@teksavvy.com>, JF Mezei writes: > Bill Gunshannon wrote: >> Irrelevant, really. Alpha is as dead as VAX as far as HP is >> concerned. HP says Itanium is VMS's future and Itanium is >> all but dead. > > > Since Alpha sales were extended due to popular demand, I'd have to disagree > that Alpha is as dead as VAX. Have they started manufacturing them again? No? Looks like they are just trying to dump stock. And that says something in itself. Looks like people abandoned Alpha even faster than HP thought they would otherwise, they would not have found themselves with all that leftover stock they now have to try and get rid of. > The fact that HP had to make a decisioh to > postpone end of sales means that it is still on its radar, as much as it > may dislike having to admit to it. Nope, sorry. It just means that HP made yet another big mistake and it left them with a lot more junk to get rid of than they had originally planned. And as to the iddea of a future port to x86-64. Dumping the likes of Hoff and the others from engineering shows that HP's plan of action is still "burning boats" so that even if someone with the power to do it were to loose their nerve and change their minds, reversing course is just not going to be possible. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 7 Feb 2007 18:10:26 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Intel prepares to kill off the Pentium 4 Message-ID: <52uj0iF1poafpU2@mid.individual.net> In article <10e8$45c9fead$cef8887a$12779@teksavvy.com>, JF Mezei writes: > david20@alpha2.mdx.ac.uk wrote: >> Not irrelevent to Intel though. Presumably their default plan was to move >> their FABS to VMS on Itanium - the simplest solution > > Intel probably has a stockpile of alpha systems and alpha parts to keep > their critical vms systems running for years. And that does about as much for VMS and its future as all those VS3100's I have in my basement. > > This gives them plenty of time to port their application to Linux or MacOS, > or just run their app on Charon's 8086 based Alpha emulator. > > Since Intel and HP are in bed together with regards to the real future of > IA64, and hence VMS and HP-UX and NSK, Intel might know that VMS will be > ported to the 8086 soon and thus not bother with migration to IA64. It > would have enough alphas to last until VMS is available on the 8086. Who is going to do that port? The VMS fairies? HP "burned their boats" again when they got rid of the cream of VMS engineering. > > And if VMS is at a cul-de-sac with IA64, then Intel also has plenty of time > to migrate to a totally different platform/OS. Remember that once HP > announces that VMS is no longer developped, the timer starts before support > is pulled out. And Intel probably won't want to be running a truly mission > critical application on an unsupported platform. Exactly. And, as has been said repeatedly here, VMS sinks or swims with Itanium. Now, does that give everyone here a real warm and fuzzy feeling? If you run a serious business and you currently run it on VMS, it is really time to read the writting on the wall. I know there is nothing that can match some of the more esoteric features of VMS (like security, wether it is real or by obscurity) but reality is a hard mistress. It's really time to start looking at what can be done to acquire/devise/create a replacement. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 7 Feb 2007 08:18:30 -0600 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Mark Daniel and/or Hein van den Huevel or anyone really Message-ID: In article , Paul Sture writes: > > Device BG274:, device type unknown, is online, mounted, record-oriented > device, network device, mailbox device. "record-oriented device" is what's really significant here. COPY is basically going to use RMS to get records from the file and make send them as RMS records on the device. IIRC the OP had 512 byte fixed length records. That means RMS will present COPY with 512 byte records and COPY will present them back to RMS as 512 byte records. That's consitent with the problems with OP was seeing. Changing the rfm to stmlf means RMS will break the records at LF boundaries, which may be what the JRE on the other end is expecting (the JRE on VMS expects this as its file handling is unmodified to deal with other formats). Alternatively the JRE may be expecting TELNET style CRLF pairs, so CONVERTing to stm from stmlf might help. It's hard to tell exactly what the BG device is going to do, though, from the documentaion (not laid out as clearly as the I/O Users Guide is for other device types). I think it will just transmit bytes as presented by RMS which means both the LF and the CRLF will be stripped as meta-data. ------------------------------ Date: Wed, 7 Feb 2007 13:27:58 -0500 From: "Syltrem" Subject: Re: MSA1500 on VMS? Message-ID: <12sk6hfk29mdrb2@corp.supernews.com> wrote in message news:1170681812.661839.237470@s48g2000cws.googlegroups.com... >I know of a number of people who have MSA1000s on their AlphaServers > and one on an Integrity too with VMS. The biggest gotcha is that the > 1000s only do active-standby, not active-active if you have two > controllers. There is a (long awaited) firmware release I believe to > fix this, but I've no news of it yet. > > Steve > For all I know, if you have VMS 8.3 then you can upgrade the controller firmware and enable active-active. To be tested soon (in a few weeks), as we migrate to 8.3 and join an Itanium box to the cluster. Syltrem ------------------------------ Date: Wed, 7 Feb 2007 13:30:22 +0100 From: "Walter Kuhn" Subject: MySQL on IA64 Problem Message-ID: <45c9c65f$0$25609$91cee783@newsreader02.highway.telekom.at> Hello Group, when I tried to install JFP-I64VMS-MYSQL-V0401-14-1.PCSI (OpenVMS V8.3), the procedure START_MYSQL.COM terminates with the following errors: (Installation etc. was ok, no errors) .... .... InnoDB: cannot read directory ., error 17 InnoDB: Error: os_file_readdir_next_file() returned -1 in InnoDB: directory . InnoDB: Crash recovery may have failed for some .ibd files! InnoDB: Restoring possible half-written data pages from the doublewrite InnoDB: buffer... 070207 13:25:22 InnoDB: Starting log scan based on checkpoint at InnoDB: log sequence number 0 43780. InnoDB: Doing recovery: scanned up to log sequence number 0 43814 InnoDB: Last MySQL binlog file position 0 79, file name ./cssi02-bin.000005 070207 13:25:22 InnoDB: Flushing modified pages from the buffer pool... 070207 13:25:22 InnoDB: Started; log sequence number 0 43814 070207 13:25:22 [ERROR] After InnoDB crash recovery, checking if the binary log './cssi02-bin.000005' contains rolled back transactions which must be removed from it... 070207 13:25:22 [ERROR] Fatal error: Can't open privilege tables: Table 'mysql.host' doesn't exist The same thing works on Alpha OpenVMS 7.3-2 Any Idea(s)? Kind Regards / Mit freundlichen Grüssen Walter Kuhn KSG GesmbH Computerstrasse 6 A - 1101 Wien Austria ------------------------------ Date: Wed, 07 Feb 2007 07:40:55 -0600 From: pechter@pechter.dyndns.org (William Pechter) Subject: Re: Odd Telnet output Message-ID: In article <9b70e$45c97339$cef8887a$10607@TEKSAVVY.COM>, JF Mezei wrote: >Alpha 8.3, TCPIP Services 5.3 > >This is over a DSL line, and a IPV4 only NAT router > >$ telnet/port=25 gprs.fido.ca >%TELNET-I-TRYING, Trying ... 204.92.15.150 >%TELNET-I-TRYING, Trying ... ::6000:0:3b:7777 >%TELNET-E-CONNFAIL, Failed to connect to remote host >-SYSTEM-F-PROTOCOL, network protocol error Could this be this is the IPV6 address... Is there any IPV6 involved? > > >Can anyone explain what exactly is happening ? > >$ ping 204.92.15.150 >PING 204.92.15.150 (204.92.15.150): 56 data bytes >64 bytes from 204.92.15.150: icmp_seq=0 ttl=243 time=57 ms > > >Now, this smtp server is supposed to service mobile handsets, so it is >normal that it would be blocking calls from the outside of the GPRS >network. But I am puzzled by the second "telnet-i-trying" line, and what >would trigger it. Is that strange address an IPV6 one ? Bill -- -- "When I think back on all the crap I learned in Vax school It's a wonder I fixed anything at all." (to the tune of Kodachrome) pechter-at-ureach.com ------------------------------ Date: 7 Feb 2007 03:58:43 -0800 From: "mb301@hotmail.com" Subject: Program looping Message-ID: <1170849523.384149.255050@m58g2000cwm.googlegroups.com> Have a cobol program that's looping does anyone have any ideas or pointer to find out where in the coding its looping? How about setting the process to DUMP? $ SET PROC/DUMP I could crash the system (with 1000 users I'll be very popular) generate a crash dump but what could I look at? Regards Mark ------------------------------ Date: 7 Feb 2007 13:37:57 +0100 From: huber@NIRWANA-mppmu.mpg.de (Joseph Huber) Subject: Re: Program looping Message-ID: In article <1170849523.384149.255050@m58g2000cwm.googlegroups.com>, "mb301@hotmail.com" writes: > Have a cobol program that's looping does anyone have any ideas or > pointer to find out where in the coding its looping? How about setting > the process to DUMP? > > $ SET PROC/DUMP > I could crash the system (with 1000 users I'll be very popular) > generate a crash dump but what could I look at? $ SET PROC/DUMP=NOW could catch it at the loop PC if it's a tight loop. You can look at the "Current PC" SHOW PROC/CONT to get an idea where the program loops., i.e. which addresses appear repeatedly. Then look at the map and compiler listings to get an idea. -- Joseph Huber , Muenchen,Germany: http://www.huber-joseph.de/ ------------------------------ Date: 7 Feb 2007 07:38:09 -0600 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Searching FMS documentation in pdf Message-ID: <1hIut5S+TqnV@eisner.encompasserve.org> In article <1170781401.529579.12790@a34g2000cwb.googlegroups.com>, vancouvercancun@yahoo.ca writes: > On Feb 6, 11:34 am, JF Mezei wrote: >> vancouvercan...@yahoo.ca wrote: >> > Hi >> > I am looking for FMS Forms Management System documentation in pdf >> > format. Is it available somewhere ? Can anybody send it to me ? >> >> I have it in bookreader format. If that is acceptable, let me know. > > Thanks, JF but I already have it on VMS in DECW which I suppose is > bookreader format. I am really looking for the stuff in a nice pdf > file if it exists for consultation on a Windows PC. > > HP, are you there ?? There is a Windows application which will read bookreader format. it's been distributed on the documentation CDs for quite a few years now. ------------------------------ Date: Wed, 07 Feb 2007 12:34:42 +0100 From: Paul Sture Subject: Re: SPANNING BACKUP TAPES Message-ID: In article <45C94D83.9060800@tx.rr.com>, John wrote: > We currently use 35/70-GB tapes and the bulk of the disk drives are > 10-30-GB in size. > > I have always made it a policy not to span tapes during my backup > procedures. To clarify, do you mean ensuring that a full backup for each disk will fit on one tape? > I have done this for a couple of reasons... > > 1) auditors - one of the questions they ask is "do you span tapes" - at > least in the past they have I'll make an intelligent guess here that what the auditors are thinking of is that some backup systems write a catolog or index of files backed up at _the end_ of the backup, and this is used when restoring. Thus, if you lose or more likely can't read the last tape, then the data on the preceding tapes is next to useless. In fact, going back a decade or more, this was the case with the DOS utility PKZIP with backups spanning multiple floppies. The restore process would ask you to insert the last floppy first. No only was this a complete swine if that last floppy couldn't be read, but when trying to use it to transfer a large amount of data from one PC to another, a) needed enough spare floppies to contain the complete backup b) couldn't start the restore on the target PC until the backup on the source PC was complete. See why the auditors ask the question? VMS BACKUP does not suffer from this design fault. You _can_ restore from multiple tape savesets if one tape is bad or missing, though of course you'll be missing the files on that tape. That leads to the question of what happens with files which happen to span tapes in a multi-tape saveset, of course. :-) > 2) DR - in the event that we have a DR incident we can restore multiple > tapes at once True, and anything which keeps things simple in the event of DR has to be a Good Thing; the KISS principle really does come into its own when doing DR for real. > 3) loss of tape - in the event that the human factor plays her deadly > card we will not lose the second tape Ditto for damaged tapes. > My questions... In brief: > How much of an issue is this? Not the issue that the auditors are thinking of. > Is not spanning tapes a good practice? Yes, if you have the resources. But I'd rather let a backup continue to another tape than find the backup is incomplete. One nasty here is that if BACKUP requires a continuation tape, and that tape isn't available, it can happily overwrite the next tape physically mounted, irrespective of its label. > Should I care about this? YES. > What are your thoughts? See above ;-) -- Paul Sture ------------------------------ Date: 7 Feb 2007 06:47:21 -0600 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: SPANNING BACKUP TAPES Message-ID: <0QAf4cXZ$XMr@eisner.encompasserve.org> In article , "Colin Butcher" writes: > After all the purpose of a backup is to be able to get the system back from > the dead as swiftly as possible if you have to. Not necessarily. In a software development environment, accuracy is much more important than speed. It is also possible that the whole system is not impacted by one disk that needs to be restored. ------------------------------ Date: Wed, 07 Feb 2007 06:58:53 -0600 From: John Subject: Re: SPANNING BACKUP TAPES Message-ID: <45C9CD0D.3070402@tx.rr.com> >>We currently use 35/70-GB tapes and the bulk of the disk drives are >>10-30-GB in size. >> >>I have always made it a policy not to span tapes during my backup >>procedures. >> >> > >To clarify, do you mean ensuring that a full backup for each disk will >fit on one tape? > > > A full backup for each disk, yes. However depending on the size of the save-set we end up with multiple save-sets on one tape. >>Is not spanning tapes a good practice? >> >> > >Yes, if you have the resources. But I'd rather let a backup continue to >another tape than find the backup is incomplete. One nasty here is that >if BACKUP requires a continuation tape, and that tape isn't available, >it can happily overwrite the next tape physically mounted, irrespective >of its label. > > The backup process (DCL) does a pre-scan of the disk to be put to tape and after the backup is complete it does a post-synopsis of the blocks put to tape - using the lexical F$GETDVI(device,"OPCNT") which has been discussed many times in previous threads. Therefore the process knows if it can or can not put to tape a save-set. In the event that it can not, it e-mails, logs, etc., the admin's. In the event that the process does over-run the tape, that process is deleted and a backup of the failed save-set(s) is then executed. So yes, we may have a failed over-write however we perform another backup immediately following that failure. ------------------------------ Date: Wed, 07 Feb 2007 07:16:42 -0600 From: John Subject: Re: SPANNING BACKUP TAPES Message-ID: <45C9D13A.2090302@tx.rr.com> >>After all the purpose of a backup is to be able to get the system back from >>the dead as swiftly as possible if you have to. >> >> > >Not necessarily. In a software development environment, accuracy is >much more important than speed. > >It is also possible that the whole system is not impacted by one disk >that needs to be restored. > > Accuracy is important however in the event of a DR - both accuracy and speed are extremely important. Restoring 10 each 70-GB tapes rather than 1 each 800-GB tape ... well, have not done the math but my money lies with the 10 each. We have ~60 disk drives - user drives, system drives, application drives, data drives, database drives, etc. Some are dependent on other disks. I remember years back when I was asked to draw up plans for a DR and what drives would be necessary for a production restores. All of them was my response. However they wanted just the bare minimum. Yet the production environment is so tightly integrated into almost every facet of the hardware - duh, that is what it was purchased for... The same person asked me to consider and develop a plan for migrating the resources and applications of a fully loaded GS160 to a single AS 2100. Dumb a** ------------------------------ Date: Wed, 07 Feb 2007 07:16:51 -0600 From: John Subject: Re: SPANNING BACKUP TAPES Message-ID: <45C9D143.7020709@tx.rr.com> >>After all the purpose of a backup is to be able to get the system back from >>the dead as swiftly as possible if you have to. >> >> > >Not necessarily. In a software development environment, accuracy is >much more important than speed. > >It is also possible that the whole system is not impacted by one disk >that needs to be restored. > > Accuracy is important however in the event of a DR - both accuracy and speed are extremely important. Restoring 10 each 70-GB tapes rather than 1 each 800-GB tape ... well, have not done the math but my money lies with the 10 each. We have ~60 disk drives - user drives, system drives, application drives, data drives, database drives, etc. Some are dependent on other disks. I remember years back when I was asked to draw up plans for a DR and what drives would be necessary for a production restores. All of them was my response. However they wanted just the bare minimum. Yet the production environment is so tightly integrated into almost every facet of the hardware - duh, that is what it was purchased for... The same person asked me to consider and develop a plan for migrating the resources and applications of a fully loaded GS160 to a single AS 2100. Dumb a** ------------------------------ Date: Wed, 07 Feb 2007 06:58:44 -0600 From: John Subject: Re: SPANNING BACKUP TAPES Message-ID: <45C9CD04.2080700@tx.rr.com> >>We currently use 35/70-GB tapes and the bulk of the disk drives are >>10-30-GB in size. >> >>I have always made it a policy not to span tapes during my backup >>procedures. >> >> > >To clarify, do you mean ensuring that a full backup for each disk will >fit on one tape? > > > A full backup for each disk, yes. However depending on the size of the save-set we end up with multiple save-sets on one tape. >>Is not spanning tapes a good practice? >> >> > >Yes, if you have the resources. But I'd rather let a backup continue to >another tape than find the backup is incomplete. One nasty here is that >if BACKUP requires a continuation tape, and that tape isn't available, >it can happily overwrite the next tape physically mounted, irrespective >of its label. > > The backup process (DCL) does a pre-scan of the disk to be put to tape and after the backup is complete it does a post-synopsis of the blocks put to tape - using the lexical F$GETDVI(device,"OPCNT") which has been discussed many times in previous threads. Therefore the process knows if it can or can not put to tape a save-set. In the event that it can not, it e-mails, logs, etc., the admin's. In the event that the process does over-run the tape, that process is deleted and a backup of the failed save-set(s) is then executed. So yes, we may have a failed over-write however we perform another backup immediately following that failure. ------------------------------ Date: Wed, 07 Feb 2007 15:10:58 GMT From: Paul Anderson Subject: Re: Wanted: Alisashare / Alisatalk for VMS / VAX Message-ID: In article , healyzh@aracnet.com wrote: > The Alpha Version apparently stopped working at some point with > OpenVMS V7.3-2. Does anyone happen to know which patch breaks it? PATHWORKS for OpenVMS (Macintosh) stops working with OpenVMS Alpha V8.2. I don't know of any patches that break it. Paul -- Paul Anderson OpenVMS Engineering Hewlett-Packard Company ------------------------------ End of INFO-VAX 2007.076 ************************