The Unix Heritage Society mailing list
 help / color / mirror / Atom feed
From: "Theodore Ts'o" <tytso@mit.edu>
To: Dan Cross <crossd@gmail.com>
Cc: tuhs@minnie.tuhs.org
Subject: Re: [TUHS] [tuhs] The Unix shell: a 50-year view
Date: Sat, 3 Jul 2021 13:37:06 -0400	[thread overview]
Message-ID: <YOCgQkwbz8Xf8ap9@mit.edu> (raw)
In-Reply-To: <CAEoi9W4NXd-yajWk9bcb1X9eK_Ws17TBJOpDE6M-Qf+uLQ2v6A@mail.gmail.com>

On Sat, Jul 03, 2021 at 09:20:57AM -0400, Dan Cross wrote:
> 
> Systemd is both good and bad.

I agree with most of what you assess as "good" and "bad" with systemd.
However...

> It uses ersatz data formats that are abjectly
> horrible. It uses binary logging, which means that one cannot use the full
> complement of Unix text filters to inspect logs....

To be fair, systemd isn't the first.  BSD's sar program uses a binary
log file.

I'm not sure whether it's BSD's fault or only after it was pulled into
the systat package for Linux, but the binary format for the sar file
is (a) not self-describing, (b) not backwards compatible, and (c) not
stable between different versions, such that if you copy /var/log/sa/saNN
the files from one system and try to interpret it on another system, you
need to make sure that other system has the same version of sar installed.

This can be a headache if you are trying to debug an enterprise system
which is using a seven-year-old version of RHEL, and your laptop is
running something a bit more modern.

Things like Ingres also came out of Berkely, and modern databases all
use binary format files.  Is anything running a database, whether it
be Ingress, Postgres, Oracle Enterprise Database, DB2, no longer "Unix"?

And then there's AIX, which uses binary config files which are
manipulated using "smit", and mixes per-machine specific configs with
more general configs, all in a single config (or should I say,
"registry") file, so $DEITY help you if you try to copy the smit
database from one system to another if you are trying to do some kind
of large scale administration setup.  And granted there are many
people would dispute whether AIX is actually "Unix", it technically
speaking qualified to use the "Unix"(tm) trademark.

Similarly, Digital Equipment Corpartion's Ultrix used a binary log
file which you had to transmogrify using the uerf ("Ultrix Error
Report Formatter").  Like systemd, Ultrix uses a structured logging so
you can pull out specific types of logs, and Ultrix's uerf works much
systemd's journalctl.

So the concept of using binary logs and binary files predates systemd,
and there are more than a few examples that can be found in historical
systems that most people would have no trouble calling "Unix", and
even qualify for the Unix(tm) trademark.

					- Ted

P.S. For the most part, Linux systems don't qualify for the Unix(tm)
trademark, although it's not clear most people care about that these
days.  (A long time ago distributions paid $$$ to Posix Compliance
labs, probably because the government procurement requirements
required Unix(tm) systems, much like OSI stacks were simlarly required
by government procurement requirements.)



  reply	other threads:[~2021-07-03 17:45 UTC|newest]

Thread overview: 109+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-02 21:24 Nelson H. F. Beebe
2021-07-02 21:36 ` Larry McVoy
2021-07-02 21:56   ` Henry Bent
2021-07-02 23:12     ` Steve Nickolas
2021-07-02 23:49       ` Steffen Nurpmeso
2021-07-03 13:34         ` Steffen Nurpmeso
2021-07-03 13:56           ` Richard Salz
2021-07-03 12:04       ` Thomas Paulsen
2021-07-03 13:20         ` Dan Cross
2021-07-03 17:37           ` Theodore Ts'o [this message]
2021-07-03 17:57             ` Warner Losh
2021-07-03 18:10               ` Theodore Ts'o
2021-07-03 20:02                 ` Dan Cross
2021-07-04  0:47           ` Tomasz Rola
2021-07-04  4:36             ` Larry McVoy
2021-07-04 14:56               ` Dan Cross
2021-07-04 16:07               ` Theodore Ts'o
2021-07-04 20:10               ` David Barto
2021-07-05  0:25                 ` Larry McVoy
2021-07-05  1:23                 ` John Cowan
2021-07-04 12:48             ` Dan Cross
2021-07-05  7:14               ` Tomasz Rola
2021-07-05 16:26                 ` John Cowan
2021-07-06 23:17                   ` Tomasz Rola
2021-07-06 23:47                     ` Steve Nickolas
2021-07-06 23:49                       ` Warner Losh
2021-07-06 23:48                     ` John Cowan
2021-07-07  0:46                     ` Theodore Ts'o
2021-07-07  0:58                       ` George Michaelson
2021-07-07  2:48                         ` Larry McVoy
2021-07-07 18:32                       ` Tomasz Rola
2021-07-07 20:50                         ` Michael Kjörling
2021-07-08  6:46                           ` [TUHS] Overgrown ffox (was: The Unix shell: a 50-year view) Tomasz Rola
2021-07-08 13:59                             ` Derek Fawcus
2021-07-08 19:25                               ` Steffen Nurpmeso
2021-07-08 19:37                                 ` Steffen Nurpmeso
2021-07-08 20:40                                 ` Steffen Nurpmeso
2021-07-08 22:23                             ` Kevin Bowling
2021-07-08 21:47                           ` [TUHS] [tuhs] The Unix shell: a 50-year view Theodore Ts'o
2021-07-09 20:14                             ` Michael Kjörling
2021-07-07 13:54                     ` Tony Finch
2021-07-06 16:05                 ` Clem Cole
2021-07-09 22:19                   ` Tomasz Rola
2021-07-04 20:10           ` Tony Finch
2021-07-05  3:59             ` Theodore Ts'o
2021-07-05 15:08               ` Steffen Nurpmeso
2021-07-05  3:52           ` Bakul Shah
2021-07-04 18:17     ` John Dow via TUHS
2021-07-04 19:46       ` Clem Cole
2021-07-05  1:33         ` Noel Hunt
2021-07-05  2:38           ` Clem Cole
2021-07-05  2:51             ` Warner Losh
2021-07-05  3:03               ` Clem Cole
2021-07-05  3:01             ` Clem Cole
2021-07-05  5:22             ` Noel Hunt
2021-07-06  5:10           ` Nevin Liber
2021-07-06 13:30             ` Clem Cole
2021-07-06 16:23               ` Theodore Ts'o
2021-07-07  1:57                 ` Dan Cross
2021-07-07  2:52                   ` Larry McVoy
2021-07-07  5:19                     ` Andrew Warkentin
2021-07-07 18:28                   ` Jon Steinhart
2021-07-10 11:51                     ` [TUHS] " Ralph Corderoy
2021-07-10 13:54                       ` Henry Bent
2021-07-10 14:12                         ` Ralph Corderoy
2021-07-10 16:57                           ` [TUHS] Death by bug [formerly The Unix shell: a 50-year view] Jon Steinhart
2021-07-11  8:53                             ` [TUHS] Death by bug Ralph Corderoy
2021-07-11  9:04                               ` arnold
2021-07-12  1:42                                 ` Theodore Y. Ts'o
2021-07-12  2:57                                   ` Jon Steinhart
2021-07-12  6:39                                   ` arnold
2021-07-12  9:56                                   ` Ralph Corderoy
2021-07-11 16:10                               ` Jon Steinhart
2021-07-12 10:37                                 ` Ralph Corderoy
2021-07-06 13:40             ` [TUHS] [tuhs] The Unix shell: a 50-year view John Cowan
2021-07-06 14:12             ` Chet Ramey
2021-07-07  0:53               ` Nevin Liber
2021-07-07 13:08                 ` Chet Ramey
2021-07-07 15:15                   ` Richard Salz
2021-07-03  0:09   ` Andrew Warkentin
2021-07-03 15:49   ` Andy Kosela
2021-07-04 23:24     ` [TUHS] Is C obsolete? (was Re: [tuhs] The Unix shell: a 50-year view) Derek Fawcus
2021-07-04 23:50       ` Nemo Nusquam
2021-07-05  0:15         ` Dan Stromberg
2021-07-05  0:21       ` Larry McVoy
2021-07-05  2:36         ` John Cowan
2021-07-05  2:59           ` Richard Salz
2021-07-05  3:47           ` Larry McVoy
2021-07-05  4:02             ` Dan Stromberg
2021-07-05 13:45               ` Steffen Nurpmeso
2021-07-05 20:15                 ` Dan Stromberg
2021-07-05 21:05                   ` Larry McVoy
2021-07-05 21:29                   ` Clem Cole
2021-07-05 22:22                     ` Brantley Coile
2021-07-06  4:35                     ` Dan Stromberg
2021-07-06  4:44                       ` Warner Losh
2021-07-06  5:58                       ` Rico Pajarola
2021-07-06 13:05                       ` Clem Cole
2021-07-05 12:11         ` Thomas Paulsen
2021-07-05  4:08       ` Dan Stromberg
2021-07-05  4:23         ` George Michaelson
2021-07-05 14:43           ` Larry McVoy
2021-07-05 15:17             ` Steffen Nurpmeso
2021-07-05 15:36               ` Steffen Nurpmeso
2021-07-05 15:53       ` Mike Markowski
2021-07-05 16:39       ` Warner Losh
2021-07-05 19:02         ` Clem Cole
2021-07-02 22:27 ` [TUHS] [tuhs] The Unix shell: a 50-year view Chet Ramey
2021-07-02 23:09 ` Steve Nickolas

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YOCgQkwbz8Xf8ap9@mit.edu \
    --to=tytso@mit.edu \
    --cc=crossd@gmail.com \
    --cc=tuhs@minnie.tuhs.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).