INFO-VAX Fri, 06 Apr 2007 Volume 2007 : Issue 190 Contents: Re: OpenVMS V8.3 BACKUP /ENCRYPT Re: OpenVMS V8.3 BACKUP /ENCRYPT Re: OpenVMS V8.3 BACKUP /ENCRYPT Re: OpenVMS V8.3 BACKUP /ENCRYPT Re: relative directories Re: relative directories ---------------------------------------------------------------------- Date: 6 Apr 2007 09:48:18 -0700 From: "AEF" Subject: Re: OpenVMS V8.3 BACKUP /ENCRYPT Message-ID: <1175878098.575267.27010@y80g2000hsf.googlegroups.com> On Apr 6, 11:55 am, "Richard B. gilbert" wrote: > Bob Koehler wrote: > > In article , bri...@encompasserve.org writes: > > >>In article <++4Tcz8MZ...@eisner.encompasserve.org>, koeh...@eisner.nospam.encompasserve.org (Bob Koehler) writes: > > >>>In article , Dan Foster writes: > > >>>>I know it stores the key used in the saveset header. > > >>>>I'm just curious if the .BCK format supports multiple backup (no pun > >>>>here) saveset headers? BACKUP creation of a single file and then DUMP of > >>>>the .BCK seems to suggest that yes, it does. > > >>> I've never seen BACKUP do anything with multiple savesets in a > >>> single BACKUP file. What else would you use multiple headers > >>> for? > > >>It's not the "headers" that we're talking about per se. The ANSI > >>labels on the save set are not encrypted and carry no information > >>relevant to the encryption. > > > ANSI tape label headers apply only to tape backup sets, which I > > did not assume. BACKUP is perfectly capable at dealing with more > > than one saveset separated by appropriate multiple ANSI tape labels. > > >>Assuming that what I've just read is accurate and that I've understood > >>it properly, what we're talking about is encryption-related information > >>stored in the backup summary record that is part of the save set itself. > > >>The "session key" used to encrypt the data in a BACKUP save set is > >>generated randomly. It is stored in the backup summary record. > >>However, that key is itself encrypted using the key the user provided. > > > So I assume you want to store more than one copy of the session key > > by aving more than one header in a single saveset. BACKUP might > > have to look at your /blocksize to determine when it can write > > that second header and be sure its in a physically different block. > > >>This means that if the summary record is lost, the "session key" is > >>gone and the backup is toast. > > > Like any part of the data, BACKUP has some very good capabiities at > > restoring data from physically bad blocks that should significantly > > reduce this if not turned off when the savest is made. I've > > acutally seen them work when reading critical data from a ratty > > 9-track. > > Ever see those capabilities work with DLT? Hint: they don't!! So I guess you might as well use /GROUP=0. OTOH, twice I recall seeing a message about either recoverable or irrecoverable blocks -- once from 8mm tape and once from 4mm tape but unfortunately I didn't write it down at the time! ARGH! So I think once in a very blue moon it might actually work, but 99.99% (or so) of the time your SOOL (or SOL according to some). AEF ------------------------------ Date: 6 Apr 2007 12:19:33 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: OpenVMS V8.3 BACKUP /ENCRYPT Message-ID: In article <1175871394.447411.135070@q75g2000hsh.googlegroups.com>, "AEF" writes: > On Apr 6, 8:34 am, koeh...@eisner.nospam.encompasserve.org (Bob > Koehler) wrote: >> >> $ backup test1.file a.bck/save >> $ backup test2.file b.bck/save >> $ append b.bck a.bck > > Yes, I thought of that, but why would anyone expect something like > this to work? (See JF's response to my question.) It's what the original post sounded like, but not what I think was intended. I was just trying to get clarification on what was intended. ------------------------------ Date: 6 Apr 2007 12:21:09 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: OpenVMS V8.3 BACKUP /ENCRYPT Message-ID: In article <46166D60.3010701@comcast.net>, "Richard B. gilbert" writes: > > Ever see those capabilities work with DLT? Hint: they don't!! I know for some DLT they won't work without special firmware that some data recovery specialists have put together. But then I don't use DLT when I have a choice. ------------------------------ Date: 6 Apr 2007 10:28:58 -0700 From: "AEF" Subject: Re: OpenVMS V8.3 BACKUP /ENCRYPT Message-ID: <1175880538.683663.36850@q75g2000hsh.googlegroups.com> On Apr 6, 1:19 pm, koeh...@eisner.nospam.encompasserve.org (Bob Koehler) wrote: > In article <1175871394.447411.135...@q75g2000hsh.googlegroups.com>, "AEF" writes: > > > On Apr 6, 8:34 am, koeh...@eisner.nospam.encompasserve.org (Bob > > Koehler) wrote: > > >> $ backup test1.file a.bck/save > >> $ backup test2.file b.bck/save > >> $ append b.bck a.bck > > > Yes, I thought of that, but why would anyone expect something like > > this to work? (See JF's response to my question.) > > It's what the original post sounded like, but not what I think was > intended. I was just trying to get clarification on what was intended. OK. Good point. AEF ------------------------------ Date: 6 Apr 2007 10:07:26 -0700 From: "AEF" Subject: Re: relative directories Message-ID: <1175879246.785418.172340@n76g2000hsh.googlegroups.com> On Apr 6, 10:59 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > In article <1175822909.973611.248...@e65g2000hsc.googlegroups.com>, > "AEF" writes: > > > > > I have plenty of symbols. I wrote TO.COM to take as much pain out of > > changing defaults as possible. I use WH for SHOW DEFAULT, one less > > letter than pwd and it does not contain two consecutive letters for > > the same hand! :-) > > Explain to me again why a command like "WH" is more intuitive than > "ls" or "cd"? > > (For those who have forgotten, just remember back to all the comments > made here about the lack of intuitiveness of Unix commands!!) > > bill > > -- > Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves > b...@cs.scranton.edu | and a sheep voting on what's for dinner. > University of Scranton | > Scranton, Pennsylvania | #include Hi Bill! Hmmmm. Show me where I said WH was more intuitive. I didn't say that. I was claiming it was easier to type, which was the whole point of using PWD vs. SHOW DEFAULT, no? Furthermore, I don't write operating systems to use such commands; I inflict them only upon myself! Anyway, since you almost asked, it comes from this: $ WH*ERE :== SHOW DEFAULT But when I am writing "code", I use SHOW DEFAULT. The spelled-out "intuitive" commands from DCL are best in DCL command procedures, but at the interactive DCL prompt, I use my symbols, many of which run DCL command procedures or executables. Hey, *my* biggest complaint about Unix is the "man pages". They vary in quality, but many of them are really, really dense. As for the commands like rm and mv I have to remember NOT to think of their English origins because when I do that I tend to type "rem" and "mov". Yuck! The command "ls" is okay because I think of that as "ls" for some reason (the origin is list?). I need to think of 'rm' and 'mv' as ARE-EM and EM-VEE, but for some reason I think of REMOVE and MOVE instead. Yuck! (I tend to think in sounds -- I'm not sure how others think.) Ah, a quick check on Wikipedia for ls says it comes from List Segments (segments?). That's fine as it's an obvious acronym. Hmmmmm. I think I'll be able to learn more about Unix from Wikipedia than from its own man pages! AEF ------------------------------ Date: 6 Apr 2007 12:14:56 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: relative directories Message-ID: In article , "Tom Linden" writes: > It was intended to show similar behaviour on Unix, ../../../somedir > which doen't have anything like [---.somedir] Ah yes, thank you for another example of eunichs being a PITA. ------------------------------ End of INFO-VAX 2007.190 ************************