[Info-vax] Anyone interested in another public access system
Bill Pechter
pechter at bandit.pechter.dyndns.org.pechter.dyndns.org
Tue Apr 14 12:29:47 EDT 2009
In article <74jo28F13th6fU1 at mid.individual.net>,
Bill Gunshannon <billg999 at cs.uofs.edu> wrote:
>In article <gs0quj$8en$1 at pechter.motzarella.org>,
> pechter at bandit.pechter.dyndns.org.pechter.dyndns.org (Bill Pechter) writes:
>> In article <rdSdnTee6OfhV37UnZ2dnUVZ_jGdnZ2d at giganews.com>,
>> Dennis Boone <drb at ihatespam.msu.edu> wrote:
>>> > Unix provides no means to create a contiguous file; just splatter it
>>> > all over the disk!. Even <obligatory retching noises> Windows has
>>> > a utility to make your files and free space contiguous. Unix just
>>> > doesn't care.
>>>
>>>All over the disk? Nonsense. Unices have always tried to keep files
>>>(and their related inodes) contiguous or close together on the disk.
>>>Defrag utilities exist for some filesystems, and though coverage could
>>>be better, fixing it later with a defrag tool is a bass-ackwards fix.
>>>In any event, the idea that the system has to *force* the file to be
>>>contiguous can't be useful very often -- surely a best effort is better
>>>than an abort.
>>>
>>
>> Really... Not true in AT&T based Unixes using the SystemV 1k filesystem.
>> Fragmentation was an issue. Before that it was even worse.
>> I remember doing ncheck, icheck etc before there was FSCK.
>
>Yeah, and I remember doing SQUEEZE on my DEC disks. Do we still have to
>do that on VMS, I wonder?
Nope... Never did although DSC (Disk Save and Compress) was used before
VMS Backup. IIRC DSC may have been compatibility mode. It was used for
the initial install and compressing backups in the VMS 2.x days.
It did the equivalent of Dump/Restore...
Squeeze was RT11 which did only contiguous allocation for Real Time
fast disk access.
>
>>
>> Mount verification in progress and the bitmap check on VMS was a whole
>> lot better. BSD very different than the AT&T based Unix filesystems.
>>
>> VMS and the BSD's were a bit better in trying to keep the allocations
>> together.
>
>Can't speak for VMS, but as regards BSD, that is a bit of an understatement.
>From some of my systems:
>
>1746 files, 44420 used, 10109779 free (2635 frags, 1263393 blocks, 0.0%
>fragmentation)
>19 files, 25064 used, 20286338 free (26 frags, 2535789 blocks, 0.0%
>fragmentation)
>6878 files, 289010 used, 25099511 free (5143 frags, 3136796 blocks, 0.0%
>fragmentation)
>789103 files, 79801403 used, 156703416 free (197312 frags, 19563263
>blocks, 0.1% fragmentation)
>
>User fileserver. Last one is the users directory. 1400 users filesystem
>shared thru NFS on Unix systems and Samba on Windows system. All faculty
>and student files are here.
>
>
>132495 files, 2163445 used, 54601550 free (41366 frags, 6820023 blocks,
>0.1% fragmentation)
>
>Server set up for literacy students to put up web pages. 4100 students.
>
>
>562 files, 2945723 used, 34908981 free (293 frags, 4363586 blocks, 0.0%
>fragmentation)
>
>Our email spool. 1400 users. ~40000 connections evey 24 hours.
>
>Note the percentage of fragmentation. And, no, there are no de-frag
>programs for BSD Unix. The question does come up from a windows weenie
>periodically in the BSD Newsgroups.
>
Don't need 'em.
>>
>> The BSD Fast Filesystem and it's decendants were a huge improvement over
>> the AT&T filesystems which looked and acted like they were written for
>> RK05's. (and they were...)
>
>But then, look at when AT&T actually stopped developing SYSV.
Uh... late 80's or so. Spun off to USL in about 1990. AT&T still
had an interest back then...
Don't think the Bell Labs folks were involved -- having started Plan9
around then.
http://plan9.bell-labs.com/plan9/about.html
Remember AT&T Unix used to have to have error-free RP06 packs
to run. Misco cleaned their clock on the very expensive packs.
A friend hacked in a bad block table sticking them in a special inode.
This was back in my Ft. Monmouth Contractor Days... post DEC
(around 1987 or so).
>
>bill
>
>--
>Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
>billg999 at cs.scranton.edu | and a sheep voting on what's for dinner.
>University of Scranton |
>Scranton, Pennsylvania | #include <std.disclaimer.h>
Bill
--
--
Digital had it then. Don't you wish you could buy it now!
pechter-at-pechter.dyndns.org
More information about the Info-vax
mailing list