Linux NFS development
 help / color / mirror / Atom feed
From: Banu Cosmin <sharpcoder@xnet.ro>
To: nfs@lists.sourceforge.net
Subject: local unmount problem after starting or using NFS
Date: Wed, 25 Dec 2002 23:06:30 +0200	[thread overview]
Message-ID: <200212252306.30993.sharpcoder@xnet.ro> (raw)

Hello!

=09I am posting this since I searched about it on the internet and it see=
ms like=20
there are many people having these types of problems, yet no good solutio=
n=20
found.
=09I want to show you some scenarios and then pose the questions, and hop=
e=20
someone can help:

=09-- Local UnMounting doesn't work --

=091. consider the following:
=09=09- the linux sistem boots
=09=09- some local filesystems are automounted under `/mnt` (or some othe=
r=20
directory)
=09=09- the NFS service starts up
=09=09- the machine finishes loading the linux system
=09=09- now, when the user logs in, if he tries to unmount a local filesy=
stem=20
(`umount /mnt/drive`) mounted by automount, he gets `umount: /mnt/drive:=20
device busy`; it seems like, although the user didn't launch any process=20
which uses /mnt/drive, the NFS service somehow locked it and it cannot be=
=20
locally unmounted
=09=09- note: this scenario can be avoided by not letting automount to mo=
unt local=20
partitions at system startup; the NFS service must be started before any=20
local mounting takes place; however this doesn't work in scenario 2
=092. the previous scenario problem can be obtained also like this:
=09=09- NFS Server boots, no automount takes place
=09=09- NFS Client boots (another machine)
=09=09- NFS Server locally mounts /mnt/drive; please note that in this si=
tuation,=20
if we want to unmount /mnt/drive, we can do so, we don't get any errors, =
as=20
long as the following step takes place
=09=09- the Client mounts Server:/mnt/drive
=09=09- the Client gets its job done, then unmounts Server:/mnt/drive
=09=09- if the Server tries to unmount the local filesystem (`umount /mnt=
/drive`),=20
he gets `umount: /mnt/drive: device busy`; it is the same result as in=20
scenario 1
=09I did a little research and found that these two situations can be res=
ovled=20
with the following steps:
=09=09a. `exportfs -u Client:/mnt/drive` for each client that has access =
to that=20
drive (or a `exportfs -ua` for all exports)
=09=09b. unmount the local partition `umount /mnt/drive`, which succeedes=
 this=20
time
=09=09c. `exportfs Client:/mnt/drive` for each client that has access to =
that=20
drive (or a `exportfs -ra` for all exports)
=09OR
=09=09a. the NFS service on Server is stopped
=09=09b. unmount the local partition `umount /mnt/drive`
=09=09c. the NFS service on Server is started

=09Anyway, it seems like this problem is valid only for the local mounted=
=20
filesystems (partitions, cdrom, etc.) which appear in the '/etc/exports'=20
config file, thus shared to the outside world. It would be great if someo=
ne=20
could tell us what really goes on when NFS starts up, or when the Client=20
nfs-mounts the Server's local mounted drive. I'm kind of new to linux but=
 it=20
seems like on startup all entries in `/etc/exports` are somehow locked...=
 we=20
need gurus here :)

=09Thanks in advance for any help,
=09=09Banu Cosmin.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

                 reply	other threads:[~2002-12-25 21:08 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200212252306.30993.sharpcoder@xnet.ro \
    --to=sharpcoder@xnet.ro \
    --cc=nfs@lists.sourceforge.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox