6.0-RC1/amd 6.1.2.1: mount stuck

From: Danny Braniss <danny_at_cs.huji.ac.il>
Date: Mon, 17 Oct 2005 13:50:30 +0200
problem: /cs/share is atomounted, but mount(NFSv3/UDP) is stuck

 ps alwx | grep amd
    0   395     1   0  96  0  1764  1784 select Ss    ??    0:00.20 /SBIN/amd 
-p -a /a -l /var/log/amd -x all -D all /usr/local hesiod.lo ...
    0   511   395 208  -1  0  1764  1784 nfsreq D     ??    0:00.00 /SBIN/amd 
-p -a /a -l /var/log/amd -x all -D all /usr/local hesiod.lo ...

a gcore of 511, shows that it's stuck in the mount syscall.

this seems to be some kernel timming/race problem, because:
	1- it does not happen with a slower cpu.
	2- it does not happen if the mount sequence is slight different
	3- it does not happen with the older amd  6.0.10-2004051.

btw, kern.smp.disabled=1, the problem still exists.

I don't think it's the server, a NetAPP, because all other mounts are ok.
danny
Received on Mon Oct 17 2005 - 09:50:31 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:45 UTC