> I've noticed this as well, although I haven't had time to go into it much.
> The strange thing is that attaching to the amd process with strace seems to
> get it going again.
Ditto here. And strace shows, that it hangs before a stat(). Seems to be
one of the changes in linux/fs. Unfortunately I've not enough time to
check this more thouroughly.
--Swen
PS: here are some traces of hung amds:
# strace -p 1093
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1639, ...}) = 0
open("/etc/mtab", O_RDWR) = 9
flock(9, LOCK_EX) = 0
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1639, ...}) = 0
fstat(9, {st_mode=S_IFREG|0644, st_size=1639, ...}) = 0
...
and on another machine:
# strace -p 505
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1512, ...}) = 0
open("/etc/mtab", O_RDWR) = 9
SYS_143(0x9, 0x2, 0xfbad2480, 0x2, 0x9) = 0
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1512, ...}) = 0
fstat(9, {st_mode=S_IFREG|0644, st_size=1512, ...}) = 0
...
# strace -p 339
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1414, ...}) = 0
open("/etc/mtab", O_RDWR) = 9
SYS_143(0x9, 0x2, 0xfbad2480, 0x2, 0x9) = 0
stat("/etc/mtab", {st_mode=S_IFREG|0644, st_size=1414, ...}) = 0
fstat(9, {st_mode=S_IFREG|0644, st_size=1414, ...}) = 0
...