Skip to contentSkip to conversation

pre6 weird ps output

DISCUSSION Jeff Garzik 2 participants 3 messages Oct 31, 2001 – Nov 4, 2001

lore
Share & tools
lore
#0 lore raw

alpha running 2.4.14-pre6. The "...log" you see is where output is redirected. /proc/filesystems and /proc/mounts at end of output.

A devpts problem? I usually don't enable devpts.

[jgarzik@brutus rpm]$ ps xf

  PID TTY      STAT   TIME COMMAND
16013 pts/1    S      0:00 -bash
32105 pts/1    R      0:00 ps xf
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder

30660 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/usr/lib/rpm/rpmb
30670 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh -e /tm
30671 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh ./
31942 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00
\_ /bin/s
32098 pts/0 S 0:00 \_ /bin/sh configure
--prefix=/usr
32099 pts/0 S 0:00 \_ gcc -O -o conftest
conftest.c
32103 pts/0 S 0:00 \_
/usr/lib/gcc-lib/alpha-ma
32104 pts/0 R 0:00 \_ /usr/bin/ld
-m elf64a
[jgarzik@brutus rpm]$ ps xf

  PID TTY      STAT   TIME COMMAND
16013 pts/1    S      0:00 -bash
  518 pts/1    R      0:00 ps xf
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder

30660 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/usr/lib/rpm/rpmb
30670 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh -e /tm
30671 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh ./
31942 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00
\_ /bin/s
515 pts/0 S 0:00 \_ /bin/sh configure
--prefix=/usr

  516 pts/0    S      0:00                      \_ gcc -O -c conftest.c
  517 pts/0    D      0:00                          \_
/usr/lib/gcc-lib/alpha-ma
[jgarzik@brutus rpm]$ ps xf
  PID TTY      STAT   TIME COMMAND
16013 pts/1    S      0:00 -bash
 5275 pts/1    R      0:00 ps xf
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder

30660 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/usr/lib/rpm/rpmb
30670 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh -e /tm
30671 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh ./
5004 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00
\_ make
5201 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S
0:00 \_ /b
5202 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S
0:00 \
5203 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S
0:00
5272 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S
0:00
5274 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log R
0:03

[jgarzik@brutus rpm]$ cat /proc/filesystems

nodev   rootfs
nodev   bdev
nodev   proc
nodev   sockfs
nodev   tmpfs
nodev   shm
nodev   pipefs
        ext2
nodev   devfs
nodev   nfs
nodev   devpts

[jgarzik@brutus rpm]$ cat /proc/mounts /dev/ide/host0/bus0/target0/lun0/part2 / ext2 rw 0 0 /proc /proc proc rw 0 0
none /dev/pts devpts rw 0 0
none /dev/shm tmpfs rw 0 0
bum:/g /g nfs rw,v3,rsize=8192,wsize=8192,hard,udp,lock,addr=bum 0 0 cum:/cooker /cooker nfs rw,v3,rsize=8192,wsize=8192,hard,udp,lock,addr=cum 0 0

Signature · 3 lines
Jeff Garzik      | Only so many songs can be sung
Building 1024    | with two lips, two lungs, and one tongue.
MandrakeSoft     |         - nomeansno
#1 lore raw
Quoted 6 lines · Jeff Garzik

[jgarzik@brutus rpm]$ ps xf

  PID TTY      STAT   TIME COMMAND
16013 pts/1    S      0:00 -bash
32105 pts/1    R      0:00 ps xf
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder
Quoted 7 lines

30660 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/usr/lib/rpm/rpmb
30670 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh -e /tm
30671 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/bin/sh ./
31942 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00

Further clue: kill ssh session and log back in later, and ps output [from a different pty] is normal.

Quoted 21 lines

[jgarzik@brutus jgarzik]$ ps xf

  PID TTY      STAT   TIME COMMAND
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder
21855 pts/0    S      0:00  \_ /usr/lib/rpm/rpmb -bb --clean --rmspec --rmsource
21863 pts/0    S      0:00      \_ /bin/sh -e /tmp/rpm/tmp/rpm-tmp.70455
25822 pts/0    S      0:00          \_ make
25823 pts/0    S      0:00              \_ make all-recursive
25824 pts/0    S      0:00                  \_ /bin/sh -c set fnord w; amf=$2;  
25840 pts/0    S      0:00                      \_ /bin/sh -c set fnord w; amf=$
25841 pts/0    S      0:00                          \_ make all
25842 pts/0    S      0:00                              \_ /bin/sh -c set fnord 
 2530 pts/0    S      0:00                                  \_ /bin/sh -c set fn
 2531 pts/0    S      0:00                                      \_ make all
 3802 pts/0    S      0:00                                          \_ /bin/sh .
 4041 pts/0    S      0:00                                              \_ gcc -
 4043 pts/0    R      0:01                                                  \_ /
 4044 pts/0    S      0:00                                                  \_ a
 2801 pts/2    S      0:00 -bash
 2777 pts/1    S      0:00 -bash
 4045 pts/1    R      0:00 ps xf
Signature · 3 lines
Jeff Garzik      | Only so many songs can be sung
Building 1024    | with two lips, two lungs, and one tongue.
MandrakeSoft     |         - nomeansno
#2 lore raw
Quoted 9 lines · Jeff Garzik
Quoted 8 lines · Jeff Garzik

[jgarzik@brutus rpm]$ ps xf

  PID TTY      STAT   TIME COMMAND
16013 pts/1    S      0:00 -bash
32105 pts/1    R      0:00 ps xf
15858 pts/0    S      0:00 -bash
15889 pts/0    S      0:02 /bin/sh /usr/bin/rpm-rebuilder

30660 /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log S 0:00 \_
/usr/lib/rpm/rpmb

Your TTY is /tmp/rpm/other_clisp-2.27-1mdk.src.rpm.log isn't it? (it is a char device that matches what /proc/*/stat reports)

Quoted 2 lines

Further clue: kill ssh session and log back in later, and ps output [from a different pty] is normal.

This may be because ps from a different pty is unable to see the real pathname in /proc/*/fd/ and must guess via other means.

The whole name resolution process is really gross. It would be wonderful to have a /proc/*/tty symlink to examine.