🚀 go-pugleaf

In article <CM1nAp.37L@world.std.com> RAID7@world.std.com (John OBrien) writes:
>bob@bdofed.UUCP writes:
>
>>I have the need to spin a very large file system.  In the area of 8-10 Gb
>>in size.  Two questions:
>
>>          1) What is the best way to accomplish this?
>>                    - RAID technology.........
>
>Our customers typically find that the new fast RAID disk array platforms are
>particularly useful in the range of 3 GB up to 150 GB and well beyond.  I 
>have had a couple of technical reports on standards-based RAID 
>architecture and performance published in the literature.  Technical 
>people needing copies with graphs, charts, and diagrams can call us up on 
>the phone for copies, either here or at our European headquarters.  I 
>often do not get to my Email very promptly because of my being out at 
>customer sites planning installations, so it is a lot faster to just call 
>us up on the phone.  Your understanding is appreciated.
>
>John O'Brien                Tel. 603-880-3005
>Storage Computer Corp.      Fax. 603-889-7232
>11 Riverside Street
>Nashua, NH 03062            European headquarters: Wiesbaden, Germany
>United States               Tel. 49 611 30 82 30 6   
>
>-----------------------"Protect Your Writes"-------------------------
PR>
PR>	OK, I've got to get the straight story on HP-UX/9.0 files whose
PR>	size exceeds 2^31-1.  I guess a file could be bigger than 2GB if
PR>	the user of the file can live with having to `lseek' more than
PR>	once to get to the desired offset, but the man page (lseek(2))
PR>	seems a bit confused.  Firstly, it refers to the return value
PR>	as an _integer_ whereas it's clearly printed as an `off_t'.  OK,
PR>	that's nitpicking.  Then it goes on to say (in a circumlocutory
PR>	way) that `lseek' returns an _unsigned_ "integer" when a file's
PR>	> 2GB.  Finally, it says that "-1 always indicates an error re-
PR>	turn, even when encountered on greater that 2 gigabyte disk sec-
PR>	tions."  OK, I'll bite: how can I distinguish between a successful
PR>	`lseek' to 2^32-1 and any unsuccessful seek?
PR>
PR>	Also, was my postulation about the multiple `lseek's for big files
PR>	correct?  Thanks all.
PR>							Paul Ruel
PR>