Re: [Cooker] df -x supermount does not work
borsenkow == Borsenkow Andrej [EMAIL PROTECTED] writes:
borsenkow {pts/1}% rpm -q fileutils
borsenkow fileutils-4.1.5-1mdk
borsenkow df -x supermount still stats all mount points every time.
in 2.4.18-16mdk (to be released).
df will not enter any supermounted device, except in somebody else is
accesing it.
Later, Juan.
borsenkow execve(/bin/df, [df, -k, -x, supermount], [/* 32 vars */]) = 0
borsenkow ...
borsenkow open(/proc/mounts, O_RDONLY) = 3
borsenkow fstat64(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
borsenkow old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
borsenkow -1, 0) = 0x40193000
borsenkow read(3, /dev/root / reiserfs rw 0 0\n/dev..., 4096) = 483
borsenkow stat64(/, {st_mode=S_IFDIR|0755, st_size=544, ...}) = 0
borsenkow stat64(/dev, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
borsenkow close(3)= 0
borsenkow ...
borsenkow open(/proc/mounts, O_RDONLY) = 3
borsenkow fstat64(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
borsenkow old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
borsenkow -1, 0) = 0x40193000
borsenkow read(3, /dev/root / reiserfs rw 0 0\n/dev..., 4096) = 434
borsenkow stat64(/, {st_mode=S_IFDIR|0755, st_size=544, ...}) = 0
borsenkow stat64(/dev, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
borsenkow stat64(/proc, {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0
borsenkow stat64(/dev/pts, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
borsenkow stat64(/dev/shm, {st_mode=S_IFDIR|S_ISVTX|0777, st_size=0, ...}) = 0
borsenkow stat64(/mnt/cdrom, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
borsenkow stat64(/mnt/floppy, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
borsenkow stat64(/mnt/hd, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
borsenkow stat64(/mnt/win_c, {st_mode=S_IFDIR|0777, st_size=22016, ...}) = 0
borsenkow stat64(/mnt/win_d, {st_mode=S_IFDIR|0777, st_size=16384, ...}) = 0
borsenkow stat64(/mnt/hdb7, {st_mode=S_IFDIR|0755, st_size=432, ...}) = 0
borsenkow stat64(/proc/bus/usb, {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0
borsenkow stat64(/proc/sys/fs/binfmt_misc, {st_mode=S_IFDIR|0755, st_size=0,
borsenkow ...}) = 0
borsenkow close(3)= 0
borsenkow ...
borsenkow -andrej
--
In theory, practice and theory are the same, but in practice they
are different -- Larry McVoy
Re: [Cooker] df -x supermount does not work
Borsenkow Andrej [EMAIL PROTECTED] writes:
{pts/1}% rpm -q fileutils
fileutils-4.1.5-1mdk
df -x supermount still stats all mount points every time.
I've been looking at that problem (because it takes forever on my
test machine with SCSI cdrom).
It seems that `df -x supermount' does work (e.g. in
df.c:show_dev(), excluded_fstype(fstype) yields true when a
supermount mountpoint is considered, and the function does return
before displaying the entry), the problem is with sys/statvfs.h's
statvfs(), which systematically stats all the mountpoints,
although we call it with a given mountpoint only.
So I've looked inside glibc's code, but I failed to find where
the glibc would enforce stating all the mountpoints, even if
stracing the df seems to show that there are multiple syscalls,
so maybe it's the kernel itself that generates multiple syscalls
to itself?
--
Guillaume Cottenceau - http://people.mandrakesoft.com/~gc/
Re: [Cooker] df -x supermount does not work
Guillaume Cottenceau [EMAIL PROTECTED] writes: So I've looked inside glibc's code, but I failed to find where the glibc would enforce stating all the mountpoints, even if stracing the df seems to show that there are multiple syscalls, so maybe it's the kernel itself that generates multiple syscalls to itself? Ok I've been caught by their habit of including a .c file inside another .c file. Black magic seems to be inside sysdeps/unix/sysv/linux/internal_statvfs.c where they stat all filesystems present in /proc/mounts, in order to discover the mount flags :-(. -- Guillaume Cottenceau - http://people.mandrakesoft.com/~gc/
[Cooker] df -x supermount does not work
{pts/1}% rpm -q fileutils
fileutils-4.1.5-1mdk
df -x supermount still stats all mount points every time.
execve(/bin/df, [df, -k, -x, supermount], [/* 32 vars */]) = 0
...
open(/proc/mounts, O_RDONLY) = 3
fstat64(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x40193000
read(3, /dev/root / reiserfs rw 0 0\n/dev..., 4096) = 483
stat64(/, {st_mode=S_IFDIR|0755, st_size=544, ...}) = 0
stat64(/dev, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
close(3)= 0
...
open(/proc/mounts, O_RDONLY) = 3
fstat64(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x40193000
read(3, /dev/root / reiserfs rw 0 0\n/dev..., 4096) = 434
stat64(/, {st_mode=S_IFDIR|0755, st_size=544, ...}) = 0
stat64(/dev, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
stat64(/proc, {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0
stat64(/dev/pts, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
stat64(/dev/shm, {st_mode=S_IFDIR|S_ISVTX|0777, st_size=0, ...}) = 0
stat64(/mnt/cdrom, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
stat64(/mnt/floppy, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
stat64(/mnt/hd, {st_mode=S_IFDIR|0777, st_size=0, ...}) = 0
stat64(/mnt/win_c, {st_mode=S_IFDIR|0777, st_size=22016, ...}) = 0
stat64(/mnt/win_d, {st_mode=S_IFDIR|0777, st_size=16384, ...}) = 0
stat64(/mnt/hdb7, {st_mode=S_IFDIR|0755, st_size=432, ...}) = 0
stat64(/proc/bus/usb, {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0
stat64(/proc/sys/fs/binfmt_misc, {st_mode=S_IFDIR|0755, st_size=0,
...}) = 0
close(3)= 0
...
-andrej
Re: [Cooker] df
On 20010925 Fabrice FACORAT wrote: Le Mardi 25 Septembre 2001 01:50, Yura Gusev scribit : none192212 0192212 0% /dev/shm ^^ Isn't it tmpfs? Filesystem means for the device. As in fstab we have : none /dev/shm tmpfs defaults 0 0 df shows : none 125M 0 124M 0% /dev/shm That is a problem whe you have several 'none' device filsystems: none /proc proc defaults 0 0 nome /dev/pts devpts gid=5,mode=0620 0 0 nome /dev/shm tmpfs defaults,size=128M 0 0 procfs /proc proc defaults 0 0 devpts /dev/pts devpts gid=5,mode=0620 0 0 tmpfs /dev/shm tmpfs defaults,size=128M 0 0 -- J.A. Magallon # Let the source be with you... mailto:[EMAIL PROTECTED] Mandrake Linux release 8.1 (Cooker) for i586 Linux werewolf 2.4.10-beo #1 SMP Sun Sep 23 23:30:20 CEST 2001 i686
Re: [Cooker] df (this time complete)
(sorry for previous post, Ctrl-s too fast...) On 20010925 Fabrice FACORAT wrote: Le Mardi 25 Septembre 2001 01:50, Yura Gusev scribit : none192212 0192212 0% /dev/shm ^^ Isn't it tmpfs? Filesystem means for the device. As in fstab we have : none /dev/shm tmpfs defaults 0 0 df shows : none 125M 0 124M 0% /dev/shm That is a problem whe you have several 'none' device filsystems: none /proc proc defaults 0 0 nome /dev/pts devpts gid=5,mode=0620 0 0 nome /dev/shm tmpfs defaults,size=128M 0 0 Many filesystem utils just report 'problem with none', and you get confused (they report the device, and it is 'none', as 'could not mount none' ). All that filesystems accept also its filesystem type instead of 'none'. I even read it is good practice: procfs /proc proc defaults 0 0 devpts /dev/pts devpts gid=5,mode=0620 0 0 tmpfs /dev/shm tmpfs defaults,size=128M 0 0 werewolf:~ df Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda1 248895100894135151 43% / tmpfs 131072 0131072 0% /dev/shm /dev/sda2 3099292 2038984902872 70% /usr /dev/sda3 4095488 1091012 2796436 29% /home /dev/sda5 102776816975544 1% /toast See Configure.help in kernel-source/Documentation. Look for tmpfs. -- J.A. Magallon # Let the source be with you... mailto:[EMAIL PROTECTED] Mandrake Linux release 8.1 (Cooker) for i586 Linux werewolf 2.4.10-beo #1 SMP Sun Sep 23 23:30:20 CEST 2001 i686
[Cooker] df
none192212 0192212 0% /dev/shm ^^ Isn't it tmpfs? -- 7:50pm up 31 days, 8:54, 4 users, load average: 0.11, 0.08, 0.08 __ | / \ |Iouri Goussev// \\ \_\\ //_/ [EMAIL PROTECTED] _\\()//_ .'/()\'. Foo-Bar / // \\ \ jgs\\ // http://foobar.irc-unix.net| \__/ | I am not 31337. But I can use the Vi editor... ;-0
Re: [Cooker] df
On Tue, 25 Sep 2001, Fabrice FACORAT wrote: Filesystem means for the device. As in fstab we have : none /dev/shm tmpfs defaults 0 0 [will@bastard will]$ df -h | grep boot /dev/sdb1 235M 3.9M 219M 2% /boot Oh ok sorry. -- 9:42pm up 31 days, 10:46, 3 users, load average: 0.29, 0.13, 0.10 __ | / \ |Iouri Goussev// \\ \_\\ //_/ [EMAIL PROTECTED] _\\()//_ .'/()\'. Foo-Bar / // \\ \ jgs\\ // http://foobar.irc-unix.net| \__/ | I am not 31337. But I can use the Vi editor... ;-0
Re: [Cooker] df always shows Use% 100% even when 1GB free
On Sun, 22 Jul 2001 13:44:22 -0230 Juhan Leemet [EMAIL PROTECTED] wrote: I recently ran out of space, and got a few error messages that programs could not write output files. So I deleted a bunch, and cleared /var/tmp/ and /tmp. At that point the machine hung?!? Now, when I use the df command I always get the output: bash-2.05# df Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 17465100 16649144 0 100% / /dev/sdb5 16982780 13498584 2621508 84% /home1 bash-2.05# Seems to say that my root partition is full, but there's about 1GB free?!? Huh? I know df (at least on Sun) does its calculations in a weird way, but I would have expected to see at least SOME free space. Any hints appreciated. Some space is reserved for root. If it is filled for the real 100% you might run into real problems. I assume root is then not even able to login to remove some files. As a user you won't and shouldn't be able to fill it for the real 100%. The 100% shown here might actually be 95% or there about. -- Marcel Pol [EMAIL PROTECTED]
[Cooker] df always shows Use% 100% even when 1GB free
I recently ran out of space, and got a few error messages that programs could not write output files. So I deleted a bunch, and cleared /var/tmp/ and /tmp. At that point the machine hung?!? Now, when I use the df command I always get the output: bash-2.05# df Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 17465100 16649144 0 100% / /dev/sdb5 16982780 13498584 2621508 84% /home1 bash-2.05# Seems to say that my root partition is full, but there's about 1GB free?!? Huh? I know df (at least on Sun) does its calculations in a weird way, but I would have expected to see at least SOME free space. Any hints appreciated. -- Juhan Leemet Logicognosis, Inc.
Re: [Cooker] 'df' resets the Music CDROM
lonely == Lonely [EMAIL PROTECTED] writes: lonely I don't know if this is normal or not. lonely If I check the usage of my harddisk with 'df', the playing music CD stops. lonely Can this be fixed? lonely Lonely Could you give me more info about your hardware? This could happen if you have: - slow disks - a lot of memory - masking interrupts during ide writes. - a very slow processor. And this is only from the top of my memory, there could be other configurations with the same problem. If you can send me a way to reproduce that stops, I will look at it. Later, Juan. -- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy
Re: [Cooker] 'df' resets the Music CDROM
Juan Quintela [EMAIL PROTECTED] writes: This could happen if you have: - supermount (probably).
[Cooker] 'df' resets the Music CDROM
I don't know if this is normal or not. If I check the usage of my harddisk with 'df', the playing music CD stops. Can this be fixed? Lonely -- 2B or not 2B, that's a question. :)
Re: [Cooker] 'df' resets the Music CDROM
So sprach Lonely am Thu, Apr 19, 2001 at 01:20:16AM +0200: If I check the usage of my harddisk with 'df', the playing music CD stops. hehe, it reminds you to start working again :) SCNR Alexander Skwar -- How to quote: http://learn.to/quote (german) http://quote.6x.to (english) Homepage: http://www.digitalprojects.com | http://www.iso-top.de iso-top.de - Die gnstige Art an Linux Distributionen zu kommen Uptime: 9 hours 29 minutes
