xfs
[Top] [All Lists]

Stability of the Red Hat 7.1 + XFS system

To: linux-xfs@xxxxxxxxxxx
Subject: Stability of the Red Hat 7.1 + XFS system
From: Micah Yoder <yodermk@xxxxxxxx>
Date: Mon, 11 Jun 2001 14:17:04 -0400
Sender: owner-linux-xfs@xxxxxxxxxxx
Hi,

I just installed RH 7.1 w/SGI's CD and the kernel on it.  The server is 
co-located, about 400 miles away...

I am running a process (slashd of Slashcode) that hung in the "R" state using 
99% of a CPU (there are two).  That was after it was running well over a day. 
 I can't kill it, even with -9.  And every time I try to access the directory 
of its log file or any file in that directory, the process hangs in the "D" 
state and I can't kill it.

I couldn't reboot using normal methods -- when typing 'reboot', the 
'shutdown' process was unkillable in "R" and kept hogging CPU cycles.  When I 
typed '/sbin/init 0' the same happened with a 'umount' process.

I had the hosting center reset it.  Now they're saying there are filesystem 
errors on the restart and I can't get to it.  They're saying they can't use 
locate and aren't finding ifconfig in the normal spot.  They're looking at it 
now.

Is this XFS's fault?  Any ideas on what to do?  I noticed that the FAQ said 
that processes hanging while accessing an FS could be fixed by recompiling 
with egcs.  Was the kernel on SGI's CD compiled with egcs or 2.96?

Need help!

Thanks,
Micah

<Prev in Thread] Current Thread [Next in Thread>