proccess locks after touching large files over NFS (in first time only)

Alizee · 05-16-2006, 02:54 PM

Hello, I have two linux boxes. One with Slack 10.1 (client) and one with Slack 10.2 (server). Kernels on both are 2.6.16.16.
All proccesses on client are hang after touching not small files (with 1kb files all ok, but if i try to execute

Code:

# file file_in_nfs_dir.mp3

my bash on client computer hangs (if i try to read this file from samba, samba hangs too). When bash hangs in /var/log/messages appears:

Code:

May 16 23:15:46 server-games kernel: nfs: server 192.168.0.4 not responding, still trying
May 16 23:15:51 server-games last message repeated 2 times

. But connection with server seems to be okay. I have tryed to ping it with

Code:

# rpcinfo -u 192.168.0.4 nfs

. I have no firewall & my hosts.allow and hosts.deny are empty. After process hangs I can login from another shell and execute

Code:

root@server-games:~# umount -f -a -t nfs
umount2: Device or resource busy
umount: /mnt/smuzd5: device is busy

. After this command server works OK (!!!). It really not umount anything, but bash on hanged console start to work again (??).

Code:

root@server-games:/mnt/smuzd5# ls
angel_eyes.avi*  games/  music/  music_for_sort/  YO/
root@server-games:/mnt/smuzd5# strace file angel_eyes.avi
execve("/usr/bin/file", ["file", "angel_eyes.avi"], [/* 30 vars */]) = 0
brk(0)                                  = 0x8054000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=33052, ...}) = 0
old_mmap(NULL, 33052, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7f04000
close(3)                                = 0
open("/usr/lib/libz.so.1", O_RDONLY)    = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0000\27\0"..., 512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=66908, ...}) = 0
old_mmap(NULL, 69632, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7ef3000
old_mmap(0xb7f03000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0xf000) =
 0xb7f03000
close(3)                                = 0
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0 U\1\000"..., 512) = 512
fstat64(3, {st_mode=S_IFREG|0755, st_size=1357414, ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7ef2000
old_mmap(NULL, 1166612, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0xb7dd5000
mprotect(0xb7eeb000, 27924, PROT_NONE)  = 0
old_mmap(0xb7eec000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x116000
) = 0xb7eec000
old_mmap(0xb7ef0000, 7444, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0xb
7ef0000
close(3)                                = 0
mprotect(0xb7eec000, 4096, PROT_READ)   = 0
munmap(0xb7f04000, 33052)               = 0
brk(0)                                  = 0x8054000
brk(0x8075000)                          = 0x8075000
open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE) = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=293248, ...}) = 0
mmap2(NULL, 293248, PROT_READ, MAP_PRIVATE, 3, 0) = 0xb7d8d000
close(3)                                = 0
stat64("/root/.magic", 0xbfa21210)      = -1 ENOENT (No such file or directory)
open("/etc/file/magic.mgc", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory)
open("/etc/file/magic", O_RDONLY|O_LARGEFILE) = 3
old_mmap(NULL, 528384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7d0c000
fstat64(3, {st_mode=S_IFREG|0644, st_size=348394, ...}) = 0
old_mmap(NULL, 131072, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7cec000
read(3, "# Magic\n# Magic data for file(1)"..., 131072) = 131072
read(3, "e nor demand-paged:\n#\n#\tif it ha"..., 131072) = 131072
mremap(0xb7d0c000, 528384, 552960, MREMAP_MAYMOVE) = 0xb7c65000
read(3, "executables\n# From: Ignatios Sou"..., 131072) = 86250
mremap(0xb7c65000, 552960, 577536, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 577536, 602112, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 602112, 630784, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 630784, 655360, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 655360, 679936, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 679936, 704512, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 704512, 733184, MREMAP_MAYMOVE) = 0xb7bd8000
mremap(0xb7bd8000, 733184, 757760, MREMAP_MAYMOVE) = 0xb7bd8000
read(3, "", 131072)                     = 0
close(3)                                = 0
munmap(0xb7cec000, 131072)              = 0
fstat64(1, {st_mode=S_IFCHR|0720, st_rdev=makedev(136, 2), ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb7f0c000
lstat64("angel_eyes.avi", {st_mode=S_IFREG|0777, st_size=292683776, ...}) = 0
open("angel_eyes.avi", O_RDONLY|O_LARGEFILE) = 3
read(3,

Here system hangs. So i login via ssh and try to umount with -f option. And strace command completes successfull.

Code:

read(3, "RIFF(\3\246WAVI LISTr\"\0\0hdrlavih8\0\0\0"..., 65536) = 65536
--- SIGWINCH (Window changed) @ 0 (0) ---
_llseek(3, 0, [0], SEEK_SET)            = 0
close(3)                                = 0
write(1, "angel_eyes.avi: RIFF (little-end"..., 123angel_eyes.avi: RIFF (little-endian) data, AVI, 720 x 416, 25.00 fps, video: XviD, audio: Dolby AC3 (5 channels, 48000 Hz)
) = 123
munmap(0xb7f0c000, 4096)                = 0
exit_group(0)                           = ?
root@server-games:/mnt/smuzd5#

Sorry for english ^_^

Alizee · 05-16-2006, 03:07 PM

and i forgot to ask help

please help me with this issue

Alizee · 06-11-2006, 05:14 AM

after recompiling kernel (snapshot 2.6.17-rc6-git2) UNhang effect (with umount) is dissappeared. it's seems like something wrong on my local network hardware with UDP packets (may be with large only UDP packets, i haven't tested it). Problem solved by compiling kernel (on server and client) with TCP support and mounting with "-o tcp" option.

cwwilson721 · 06-11-2006, 09:16 AM

I like this thread.

OP gives the problem, logs, and other relevant information.

And then when they found the solution, they posted it.

Well done, Alizee.

Alizee · 06-11-2006, 02:13 PM

Quote:

Originally Posted by cwwilson721

Well done, Alizee.

Thx.
I was right:

Code:

root@server-games:/soft# ./netio -u 192.168.0.4

NETIO - Network Throughput Benchmark, Version 1.26
(C) 1997-2005 Kai Uwe Rommel

UDP connection established.
Packet size  1k bytes:  10790 KByte/s (0%) Tx,  11090 KByte/s (81%) Rx.
Packet size  2k bytes:  11005 KByte/s (0%) Tx,  10340 KByte/s (83%) Rx.
Packet size  4k bytes:  11413 KByte/s (0%) Tx,  9872 Byte/s (99%) Rx.
Packet size  8k bytes:  11401 KByte/s (0%) Tx,  2724 Byte/s (99%) Rx.
Packet size 16k bytes:  11338 KByte/s (0%) Tx,  0 Byte/s (100%) Rx.
Packet size 32k bytes:  11270 KByte/s (0%) Tx,  0 Byte/s (100%) Rx.
Done.