https://bugzilla.novell.com/show_bug.cgi?id=707898
https://bugzilla.novell.com/show_bug.cgi?id=707898#c3
Neil Brown changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEEDINFO |NEW
InfoProvider|nfbrown@suse.com |
--- Comment #3 from Neil Brown 2011-10-10 00:18:29 UTC ---
There are a few different issues here.
Firstly it is not clear to me that an NFS mount request should time out
particularly quickly.
The fact that you list a mountpoint in /etc/fstab suggests that you need that
filesystem mounted for successful operation, so there might be no point
continuing if it is not mounted.
If you don't really need the filesystem but would just like it there if
possible, then the correct way to request that is to use the "bg" mount option,
or to use an automounter.
So it probably makes sense fort YaST to suggest, or at least allow, the 'bg'
option to be set for mountpoints that are created -- or perferably it should
configure an automounter and use that for NFS.
That said, there is a bug in the kernel relating to NFSv4 mounts that I haven't
managed to get to the bottom of yet. If the server that is not responding is
on a different subnet, the ICMP error message which comes back to say "cannot
get there from here" gets ignored and we get a long timeout. Also NFSv4 mounts
don't seem to reduce the timeout when 'bg' is set.
However as you are not using V4 you can avoid this by explicitly requesting
'v3'.
So you should be able to reduce the timeout significantly by specifying options
"bg,vers=3".
Also I think 11.4 has some better timeouts than 11.3, so if changing the
options doesn't help, an upgrade might.
YaST could help by allowing, or even encouraging, the 'bg' option to be
specified (if it doesn't already) and also explicitly setting the NFS version
to be used.
Bug #679059 covers the issues that NFSv4 mounts don't respond to ENETUNREACH
properly, though I'm not sure if this is the case here.
If you still have problems and want more analysis, please run mount with "-v"
and report the output, and also use "tcpdump -w /tmp/somefile -s 0 host
$SERVERIP" to collect a network trace, and attach that.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.