On Tuesday 26 May 2009 18:51:28 Ritesh Raj Sarraf wrote: > I sshfs mounted and was able to write to /mount/point/proc/sysrq-trigger > successfully, if /proc/kmsg was not being read by cat. /proc/kmsg was being > read directly on the server using cat. This way, no hang. > > Still, is sshfs/fuse still involved for a bug here ?
Finally, got it working other ways:
Machine 1: The SSH Server
Machine 2: The SSH Client. Where I use sshfs to access Machine 1's filesystem.
On Machine 1, I run,
[r...@localhost ~]# cat sock.pl
#!/usr/bin/perl
open kmesg, "</proc/kmsg";
use IO::Socket;
my $sock = new IO::Socket::INET (
PeerAddr => '192.168.122.239',
PeerPort => '7070',
Proto => 'tcp',
);
die "Could not create socket: $!\n" unless $sock;
while (<kmesg>) {
print $sock "$_";
}
On Machine 2:
Before running sock.pl on Machine 1, I run "netcat -l -p 7070" on machine 2.
Then here, I do an sshfs mount.
And then I trigger as much SysRq data as I want. It mostly gets delivered to
Machine 2's port 7070.
So, thing that I omitted was, "cat".
Now, is this really an sshfs/fuse + cat bug ?
I'll leave this bug report open for the upstream dev and maintainers to
decide.
Ritesh
--
Ritesh Raj Sarraf
RESEARCHUT - http://www.researchut.com
"Necessity is the mother of invention."
signature.asc
Description: This is a digitally signed message part.
