[TUHS] [tuhs] The Unix shell: a 50-year view

[Michael Kjörling]

On 8 Jul 2021 17:47 -0400, from tytso at mit.edu (Theodore Ts'o):
>> I'm going to stick my neck out here by saying that the VSZ and RSS
>> values reported by ps, at least for Firefox, are largely meaningless.
> 
> No, VSZ is correct, but it's confusing since it includes things that
> you might not expect.  VSZ includes *all* virtual memory space
> consumed by a process.

You may very well be correct. That would, however, seem to me to make
it _correct_, but still not necessarily _meaningful_ for trying to
determine the amount of memory used by a _specific_ process; and
similarly for the RSS.


>> I started my usual Firefox instance, which has a handful of plugins,
>> about a metric gazillion bookmarks, and has been my main web browser
>> profile for years (so it probably has collected some crud over time).
>> `ps auxw` reported that process as having a total RSS of a whopping
>> 374 GB.
> 
> I don't think that's right.  Are you sure it's not 374MB?  I started
> firefox-esr on my Debian machine, and the PS output is:

On this, I believe I must stand corrected.

For my previous post, I checked the man page for ps, and noted that
RSS and VSZ are both reported in units of 1 KiB. For an attempt just
now, the ps output is similarly:

USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
michael  29890 51.8  0.7 2728696 255456 tty2   Sl+  21:46   0:06 firefox-esr
michael  29942  4.7  0.3 2415008 106448 tty2   Sl+  21:46   0:00 /usr/lib/firefox-esr/firefox-esr -contentproc -childID 1 -isF
michael  29994 38.3  0.7 34043160 240612 tty2  Sl+  21:46   0:04 /usr/lib/firefox-esr/firefox-esr -contentproc -childID 2 -isF
michael  30168  3.2  0.2 2400636 72992 tty2    Sl+  21:46   0:00 /usr/lib/firefox-esr/firefox-esr -contentproc -childID 3 -isF

On my system, unless I need to get a different calculator, 0.7% of
memory is either 229 ± <16 or 688 ± <49 MiB, depending on whether it
counts only RAM or also includes swap space.

The 249 MiB RSS of the main process also much more closely matches the
difference that was reported by `free` in my previous experiment.
Oddly, when trying the same thing again now, the difference as
reported by `free` is 423 MiB, so I'm not quite sure what's going on
there, but at least all of those numbers are considerably more
_plausible_, both in isolation and in relation to other relevant
measurements.

I'm honestly not sure how I went from the `ps` output (which wasn't
particularly dissimilar from the above, though of course I don't know
what the exact numbers were since I didn't record them) and the unit
of measurement being KiB, to going from KiB straight to GiB.


> That's not because the hello world program actually used half a
> megabyte worth of the C library; but rather, almost half a megabyte of
> the C library has been paged in to support all of the processes
> currently running in the system.  It also follows that every process
> using shared library is going to have an RSS which is at least 512k.

This is definitely a point on which I agree, and which I tried to
point out with the mention of shared libraries in my previous post.

-- 
Michael Kjörling • https://michael.kjorling.se • michael at kjorling.se
 “Remember when, on the Internet, nobody cared that you were a dog?”