27 Apr
2001
27 Apr
'01
1:21 p.m.
In fact... I seem to remember that the Ultrix-11 installation
manual specified this action during the installation process.
No urban legend.. I remember doing it on a home -11...
sync; sync; sync; halt
Megan Gentry
Former RT-11 Developer
+--------------------------------+-------------------------------------+
| Megan Gentry, EMT/B, PP-ASEL | Internet (work): gentry!zk3.dec.com |
| Unix Support Engineering Group | (home): mbg!world.std.com |
| Compaq Computer Corporation | addresses need '@' in place of '!' |
| 110 Spitbrook Rd. ZK03-2/T43 | URL: http://world.std.com/~mbg/ |
| Nashua, NH 03062 | "pdp-11 programmer - some assembler |
| (603) 884 1055 | required." - mbg KB1FCA |
+--------------------------------+-------------------------------------+
27 Apr
27 Apr
1:41 p.m.
In fact... I seem to remember that the Ultrix-11
installation
manual specified this action during the installation process.
No urban legend.. I remember doing it on a home -11...
sync; sync; sync; halt
Years ago (so far back in another year it will be ontopic), this was the
correct way to shutdown Linux as well.
Zane
3:02 p.m.
On Fri, 27 Apr 2001 healyzh(a)aracnet.com wrote:
Doing that is the only thing that prevents my Sun Ultra running
Solaris 8 from needing a manual fsck... I tend to issue three syncs on
all of my systems just to be on the safe side (see, the magic number
three is not superstition), and taught students in a UNIX class that
they should do this.
BTW, in case anyone's interested, I've still got a very limited number
of spare copies of the textbook "A Quick and Simple Introduction to
UNIX for MS-DOS Users" around here somewhere (no, I didn't mention
syncing in the text), which was typeset on a classic PERQ T2 graphics
workstation using the Mint document processing software. US$25 plus
shipping (these were produced in small runs, and that's what it cost
me per copy).
I only began writing the book as I needed to write something in order
to figure out how to use Mint. :-)
It's my understanding that IDOCS Mint was ported to Sun workstations;
is there anyone here who knows anything about this port? Somewhere in
some old e-mail, I've got a letter from the person who wrote it, but
he never got back to me with any more information.
--
Copyright (C) 2001 R. D. Davis The difference between humans & other animals:
All Rights Reserved an unnatural belief that we're above Nature &
rdd(a)perqlogic.com 410-744-4900 her other creatures, using dogma to justify such
http://www.perqlogic.com/rdd beliefs and to justify much human cruelty.
No urban
legend.. I remember doing it on a home -11...
sync; sync; sync; halt
Years ago (so far back in another year it will be ontopic), this was the
correct way to shutdown Linux as well. 
28 Apr
28 Apr
6:36 a.m.
New subject: Shutting down cleanly ... sync; sync; sync;y
> In fact... I seem to remember that the Ultrix-11
installation
> manual specified this action during the installation process.
>
> No urban legend.. I remember doing it on a home -11...
>
> sync; sync; sync; halt
This could trash the disks on PDP11/45's and 11/70's with AT&T Unix.
You see the trick with the three syncs was the time it took
for the operator to type them in by hand -- each on it's own line
nwith the carriage returns and fill characters for them to echo back.
sync <return>
sync <return>
sync <return>
and push the halt switch down
generated enough sync's and time delays for even the older Dec
controllers to flush all the buffers and then halt.
(and this was at LA36 DECwriter II's at 300 baud with LA120's
sync;sync;sync; ad 1200 baud was probably more risky... ).
It's said one person did
sync;sync;sync <carriage return>
and reach over quickly and halted the front panel switch -- only to find
he trashed some of the file system on the next boot.
of course this depended on system speed, memory buffers amount and size
and the slowness of the disks and controllers involved.
Zane
Bill
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GIT d? s+:+ a+ C+++ UB++++ P+ L+ E--- W++ N++ o K++ w---
O++ M+ V++ PS+ PE Y+ PGP t 5 X R- tv++ b+ DI++ D-
G e++ h---- r+++ y++++
------END GEEK CODE BLOCK------
--
"Bill Gates is a Persian cat and a monocle away from being a villain
in a James Bond movie" - Dennis Miller || bpechter(a)shell.monmouth.com
9:02 a.m.
New subject: Shutting down cleanly ... sync; sync; sync;y
On Sat, 28 Apr 2001, Bill Pechter wrote:
sync <return>
sync <return>
sync <return>
and push the halt switch down
generated enough sync's and time delays for even the older Dec
controllers to flush all the buffers and then halt.
Why was it necessary to run sync three times in a row rather than just
running it once and waiting for a few seconds?
--
------- Alex Holden -------
http://www.linuxhacker.org/
http://www.robogeeks.org/
10:15 a.m.
New subject: Shutting down cleanly ... sync; sync; sync;y
While the 1-missippi, 2-miss... 3... idea of the sync's
sounds OK to me, I always thought it was due to the
possibility of a 3-level cache and a 1 level sync, that is,
that there could be multiple memory buffers in the
kernal, O/S and (any) applications that would each
make it "up" one level with each sync, So you had to do
it that many times to get all the way out to the disk.
Strictly a gut feel thing based on having to write my own
(app level) memory managers a couple of times.
John A.
29 Apr
29 Apr
11 a.m.
New subject: Shutting down cleanly ... sync; sync; sync;y
The first sync wrote all the 'dirty' buffers to disc.
The second sync writes the new 'dirty' buffers (with changed
disc usage bitmaps).
The third sync was just for good measure.
clint
On Sat, 28 Apr 2001, Alex Holden wrote:
On Sat, 28 Apr 2001, Bill Pechter wrote:
sync <return>
sync <return>
sync <return>
and push the halt switch down
generated enough sync's and time delays for even the older Dec
controllers to flush all the buffers and then halt.
Why was it necessary to run sync three times in a row rather than just
running it once and waiting for a few seconds?
--
------- Alex Holden -------
http://www.linuxhacker.org/
http://www.robogeeks.org/
27 Apr
27 Apr
5:13 p.m.
In fact... I seem to remember that the Ultrix-11 installation
manual specified this action during the installation process.
No urban legend.. I remember doing it on a home -11...
sync; sync; sync; halt
I'm not sure when the halt and shutdown programs first appeared.
I do recall that at least on V6 (mid 1970s) that the sync
sequence was required whenever the system was shutdown. If the
halt program existed then it only changed run level or issued
a processor halt instruction. On the 11/45 I recall using the
front panel to halt the processor after the sync sequence. THe
halt program may not have appeared (or be commonly used) until
after front panel started to disappear. I do vaguelly recall some
discussion of this when we moved V6 to the 11/03.
--
Dr. Mark Green mark(a)cs.ualberta.ca
McCalla Professor (780) 492-4584
Department of Computing Science (780) 492-1071 (FAX)
University of Alberta, Edmonton, Alberta, T6G 2H1, Canada
29 Apr
29 Apr
5:41 p.m.
On the 11/45 I recall using the fgront panel to halt
the
processor after the sync sequence.
My notes on an early UNIX were
"System Shutdown
As the operating system keeps some critical data in
core to speed up disk operations, it is necessary
to use the sync(1) command to force this data to be
written to disk before shutting down the system. Due
to interactions with the scheduler it may be necessary
to type sync two or three times to insure that it has
actually been completed before depressing HALT."
--Chuck
8643
days inactive
8646
days old
8 comments
9 participants
participants (9)
-
alex＠linuxhacker.org -
allain＠panix.com -
cmcmanis＠mcmanis.com -
healyzh＠aracnet.com -
mark＠cs.ualberta.ca -
mbg＠world.std.com -
pechter＠pechter.dyndns.org -
rdd＠smart.net -
vaxman＠qwest.net