Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 29 Nov 2017 21:25:53 -0800
From:      Mark Millard <markmi@dsl-only.net>
To:        FreeBSD Current <freebsd-current@freebsd.org>, freebsd-hackers <freebsd-hackers@freebsd.org>
Subject:   FYI: poudriere bulk -a ended up in a "all STOP" status. . . Turns out to be: /usr/bin/nohup in background vs. tty output issue
Message-ID:  <8D4E08D3-17B3-4F22-9330-16123EA4DAD2@dsl-only.net>

next in thread | raw e-mail | index | archive | help
I attempted the experiment of building
all ports via poudriere (run via
"/usr/bin/nohup . . . &" and watched via
tail -f nohup.out ). It got to:

[16:06:54] [14] [00:03:00] Finished net-mgmt/icinga-core | =
icinga-core-1.13.3_1: Success
[16:06:57] [14] [00:00:01] Building devel/elixir-crontab | =
elixir-crontab-1.1.2
[16:06:58] [08] [00:03:38] Finished archivers/php71-zlib | =
php71-zlib-7.1.11: Success
[16:07:00] [08] [00:00:00] Building mail/p5-Mail-IMAPTalk | =
p5-Mail-IMAPTalk-4.04
[16:07:10] [26] [00:08:01] Finished devel/geany-plugin-commander | =
geany-plugin-commander-1.31: Success
[16:07:12] [26] [00:00:00] Building net/spread | spread-3.17.4_5

and then seemed to hang before whatever would have
been the next message, not that I noticed it at
that time. (Other things seem to be working fine.)
top showed most everything in a STOP state. It
had built 7014 ports.

Context: head -r326192 inside a Windows 10 Pro Hyper-V
virtual machine.

dmesg -a shows a couple of lines with a type of
message that I've not seen before:

sonewconn: pcb 0xfffff811e2da6570: Listen queue overflow: 1 already in =
queue awaiting acceptance (1 occurrences)
sonewconn: pcb 0xfffff8175d260cb0: Listen queue overflow: 1 already in =
queue awaiting acceptance (2 occurrences)

(The machine is basically dedicated to this build and
my monitoring of it.)

But I do not know the relative timing of those 2 messages
or just what they were tied to. they could be some
independent issue for all I know.

Turns out that the following started things going again:

^C the tail -f nohup.out and fg the background process.

This resulted in the following text:

^C
[1] + Stopped (tty output)    /usr/bin/nohup poudriere bulk -j =
FBSDFSSDjail -w -C -a
# fg
/usr/bin/nohup poudriere bulk -j FBSDFSSDjail -w -C -a

Then things were no longer in the STOP state. No visible
tty output appeared after the fg-command material shown
above, despite supposedly being stopped for "tty output".

Looking separately at the nohup.out file, it continued with:

. . .
[16:07:10] [26] [00:08:01] Finished devel/geany-plugin-commander | =
geany-plugin-commander-1.31: Success
[16:07:12] [26] [00:00:00] Building net/spread | spread-3.17.4_5
[16:51:28] [23] [00:48:53] Finished =
multimedia/gstreamer1-plugins-gnonlin | =
gstreamer1-plugins-gnonlin-1.4.0: Success
[16:51:34] [23] [00:00:00] Building devel/p5-MooseX-App-Cmd | =
p5-MooseX-App-Cmd-0.32
. . .

So I have an upper bound how long it was STOP'd before I
noticed.

Looks like:

/usr/bin/nohup poudriere bulk -j FBSDFSSDjail -w -C -a &

is currently a bad idea. Next time I will omit the "&".

=3D=3D=3D
Mark Millard
markmi at dsl-only.net




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8D4E08D3-17B3-4F22-9330-16123EA4DAD2>