Re: Giving up an NFS fs that wasn't available initially?

From: Shaul Karl (shaulk_at_nonexisting.hamakor.org.il)
Date: Wed 26 May 2004 - 03:00:57 IDT


On Tue, May 25, 2004 at 02:43:48PM +0300, Ehud Karni wrote:
> >
> > My expectation and understanding of the man page is that it should
> > retry. Yet this is not what happened here. Is that behavior expected? Can
> > it be that some other parameter involved? The client attempt to mount
> > the remote fs bg,soft,intr.
>
> That's not my experience. I can see the `mount' command running in the
> background if the nfs server was down on boot.
>
> do you have the `bg' parameter in your /etc/fstab entry ?
>

  Yes I do. However ps axf | grep mount shows nothing. Yet I believe
this is not the appropriate command to see what is happening since it
also shows nothing when the server went off after the fs got mounted;
even though in this case the client is able to `remount' it
automatically when the server gets back. Any suggestion what to look
for in order to have a better view of what is happening? I am using
Debian testing. What are you running?

> example from my /etc/fstab:
> linux:/ /linux nfs rsize=16384,wsize=16384,noac,soft,intr,bg 0 0
>

  The options from my /etc/fstab are:

     nosuid,bg,soft,intr,rw,rsize=4096,wsize=4096

(old hardware --> small wsize).

> As for `soft' Vs. `hard' parameter (see Oron mail) - When you use the
> `hard' option and the server goes down (after it was mounted) - your
> load may go to (almost) infinite value, and there is no way to umount
> it, OTOH, when it comes up again, any waiting I/O completes successfully
> as if no interruption occurred. Use this option CAREFULLY. I use it only
> on mounts of a critical file server.
>

  Indeed. I have reverted from hard to soft because hard seemed to
prevent unattended shutdown on critical power situations. The shutdown
process just hanged, and my, perhaps wrong, conclusion was that it
occurred because the server shutdown before the client did. This seems
to me excessive behavior on the part of hard. I mean, overriding root's
command to shutdown. On the other hand, I don't know what complications
would it take for the client to not hang only under extremely specific
circumstances.

  I have a long term plan to try another NFS project like coda. Have you
got experience with other NFS projects?

-- 
"If you have an apple and I have  an apple and we  exchange apples then
you and I will still each have  one apple. But  if you have an idea and I
have an idea and we exchange these ideas, then each of us will have two
ideas." -- George Bernard Shaw     (sent by  shaulk @ actcom . net . il)
=================================================================
To unsubscribe, send mail to linux-il-request_at_linux.org.il with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail linux-il-request_at_linux.org.il


This archive was generated by hypermail 2.1.7 : Wed 26 May 2004 - 03:10:18 IDT