* [patch] umount loopback filessystems on autofs stop
@ 2004-09-02 20:14 Jeff Moyer
2004-09-02 21:20 ` Joseph V Moss
0 siblings, 1 reply; 7+ messages in thread
From: Jeff Moyer @ 2004-09-02 20:14 UTC (permalink / raw)
To: raven; +Cc: autofs
Hello,
The netfs scripts for Red Hat go through and umount any loopback devices
before attempting to umount the encompassing filesystem. This seems a sane
thing to do (and I have a bug on said issue), so I've added similar
functionality to the autofs init script. Attached is a patch. Comments
encouraged. Testing by debian folks would be nice, too.
-Jeff
--- autofs-4.1.3/samples/rc.autofs.in.umount-loopback 2004-09-02 15:55:21.722705296 -0400
+++ autofs-4.1.3/samples/rc.autofs.in 2004-09-02 16:16:09.985940440 -0400
@@ -351,6 +351,31 @@ function signal_automounts()
return $RETVAL
}
+function umount_loopback()
+{
+ loops=`LC_ALL=C awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts`
+ automounts=`LC_ALL=C awk '!/^#/ && $1 ~ /^automount/ {print $2}' /proc/mounts`
+
+ for l in $loops
+ do
+ for a in $automounts
+ do
+ command="echo $l | grep -E '^$a'"
+ match=`echo $command | sh`
+ if [ -n "$match" ]; then
+ echo ""
+ echo -n $"Unmounting loopback filesystem $match: "
+ umount -d $match
+ if [ $? -eq 0 ]; then
+ echo "done"
+ else
+ echo "failed ($?)"
+ fi
+ fi
+ done
+ done
+}
+
# Redhat start/stop function.
#
function redhat()
@@ -386,6 +411,7 @@ case "$1" in
if [ -z "`pidofproc $prog`" -a -z "`getmounts`" ]; then
RETVAL=0
else
+ umount_loopback
signal_automounts
RETVAL=$?
fi
@@ -477,6 +503,7 @@ case "$1" in
;;
stop)
echo 'Stopping automounter.'
+ umount_loopback
start-stop-daemon --stop --quiet --signal USR2 --exec $DAEMON
;;
reload|restart)
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [patch] umount loopback filessystems on autofs stop
2004-09-02 20:14 [patch] umount loopback filessystems on autofs stop Jeff Moyer
@ 2004-09-02 21:20 ` Joseph V Moss
2004-09-03 2:31 ` Mike Waychison
0 siblings, 1 reply; 7+ messages in thread
From: Joseph V Moss @ 2004-09-02 21:20 UTC (permalink / raw)
To: jmoyer; +Cc: autofs, raven
Using /proc/mounts could be an issue for anyone trying to use this with a kernel
that truncates /proc/mounts.
> Hello,
>
> The netfs scripts for Red Hat go through and umount any loopback devices
> before attempting to umount the encompassing filesystem. This seems a sane
> thing to do (and I have a bug on said issue), so I've added similar
> functionality to the autofs init script. Attached is a patch. Comments
> encouraged. Testing by debian folks would be nice, too.
>
> -Jeff
>
> --- autofs-4.1.3/samples/rc.autofs.in.umount-loopback 2004-09-02 15:55:21.722705296 -0400
> +++ autofs-4.1.3/samples/rc.autofs.in 2004-09-02 16:16:09.985940440 -0400
> @@ -351,6 +351,31 @@ function signal_automounts()
> return $RETVAL
> }
>
> +function umount_loopback()
> +{
> + loops=`LC_ALL=C awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts`
> + automounts=`LC_ALL=C awk '!/^#/ && $1 ~ /^automount/ {print $2}' /proc/mounts`
> +
> + for l in $loops
> + do
> + for a in $automounts
> + do
> + command="echo $l | grep -E '^$a'"
> + match=`echo $command | sh`
> + if [ -n "$match" ]; then
> + echo ""
> + echo -n $"Unmounting loopback filesystem $match: "
> + umount -d $match
> + if [ $? -eq 0 ]; then
> + echo "done"
> + else
> + echo "failed ($?)"
> + fi
> + fi
> + done
> + done
> +}
> +
> # Redhat start/stop function.
> #
> function redhat()
> @@ -386,6 +411,7 @@ case "$1" in
> if [ -z "`pidofproc $prog`" -a -z "`getmounts`" ]; then
> RETVAL=0
> else
> + umount_loopback
> signal_automounts
> RETVAL=$?
> fi
> @@ -477,6 +503,7 @@ case "$1" in
> ;;
> stop)
> echo 'Stopping automounter.'
> + umount_loopback
> start-stop-daemon --stop --quiet --signal USR2 --exec $DAEMON
> ;;
> reload|restart)
>
> _______________________________________________
> autofs mailing list
> autofs@linux.kernel.org
> http://linux.kernel.org/mailman/listinfo/autofs
>
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [patch] umount loopback filessystems on autofs stop
2004-09-02 21:20 ` Joseph V Moss
@ 2004-09-03 2:31 ` Mike Waychison
2004-09-03 9:03 ` James Pearson
0 siblings, 1 reply; 7+ messages in thread
From: Mike Waychison @ 2004-09-03 2:31 UTC (permalink / raw)
To: Joseph V Moss; +Cc: autofs, raven
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Joseph V Moss wrote:
> Using /proc/mounts could be an issue for anyone trying to use this
with a kernel
> that truncates /proc/mounts.
>
Which kernel truncates /proc/mounts?
Any recent kernel uses the seq_file interface which guarantees that
nothing gets truncated. I have a system up and running that works fine
with about 2500 mounts.
>
>>Hello,
>>
>>The netfs scripts for Red Hat go through and umount any loopback devices
>>before attempting to umount the encompassing filesystem. This seems a
sane
>>thing to do (and I have a bug on said issue), so I've added similar
>>functionality to the autofs init script. Attached is a patch. Comments
>>encouraged. Testing by debian folks would be nice, too.
>>
>>-Jeff
>>
>>--- autofs-4.1.3/samples/rc.autofs.in.umount-loopback 2004-09-02
15:55:21.722705296 -0400
>>+++ autofs-4.1.3/samples/rc.autofs.in 2004-09-02 16:16:09.985940440 -0400
>>@@ -351,6 +351,31 @@ function signal_automounts()
>> return $RETVAL
>> }
>>
>>+function umount_loopback()
>>+{
>>+ loops=`LC_ALL=C awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/"
{print $2}' /proc/mounts`
>>+ automounts=`LC_ALL=C awk '!/^#/ && $1 ~ /^automount/ {print $2}'
/proc/mounts`
>>+
>>+ for l in $loops
>>+ do
>>+ for a in $automounts
>>+ do
>>+ command="echo $l | grep -E '^$a'"
>>+ match=`echo $command | sh`
>>+ if [ -n "$match" ]; then
>>+ echo ""
>>+ echo -n $"Unmounting loopback filesystem $match: "
>>+ umount -d $match
>>+ if [ $? -eq 0 ]; then
>>+ echo "done"
>>+ else
>>+ echo "failed ($?)"
>>+ fi
>>+ fi
>>+ done
>>+ done
>>+}
>>+
>> # Redhat start/stop function.
>> #
>> function redhat()
>>@@ -386,6 +411,7 @@ case "$1" in
>> if [ -z "`pidofproc $prog`" -a -z "`getmounts`" ]; then
>> RETVAL=0
>> else
>>+ umount_loopback
>> signal_automounts
>> RETVAL=$?
>> fi
>>@@ -477,6 +503,7 @@ case "$1" in
>> ;;
>> stop)
>> echo 'Stopping automounter.'
>>+ umount_loopback
>> start-stop-daemon --stop --quiet --signal USR2 --exec $DAEMON
>> ;;
>> reload|restart)
>>
>>_______________________________________________
>>autofs mailing list
>>autofs@linux.kernel.org
>>http://linux.kernel.org/mailman/listinfo/autofs
>>
>
>
> _______________________________________________
> autofs mailing list
> autofs@linux.kernel.org
> http://linux.kernel.org/mailman/listinfo/autofs
- --
Mike Waychison
Sun Microsystems, Inc.
1 (650) 352-5299 voice
1 (416) 202-8336 voice
http://www.sun.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
NOTICE: The opinions expressed in this email are held by me,
and may not represent the views of Sun Microsystems, Inc.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFBN9d9dQs4kOxk3/MRAuv7AJ9RwuxkWHgwUkHsW/+C+Qoy/5hzhgCdFoSW
fM9teS9ID1cEs0ZzVJfgO+g=
=qIAQ
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [patch] umount loopback filessystems on autofs stop
2004-09-03 2:31 ` Mike Waychison
@ 2004-09-03 9:03 ` James Pearson
2004-09-03 12:30 ` Mike Waychison
0 siblings, 1 reply; 7+ messages in thread
From: James Pearson @ 2004-09-03 9:03 UTC (permalink / raw)
To: Mike Waychison; +Cc: autofs
Mike Waychison wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Joseph V Moss wrote:
>
>>Using /proc/mounts could be an issue for anyone trying to use this
>
> with a kernel
>
>>that truncates /proc/mounts.
>>
>
>
> Which kernel truncates /proc/mounts?
>
> Any recent kernel uses the seq_file interface which guarantees that
> nothing gets truncated. I have a system up and running that works fine
> with about 2500 mounts.
seq_file support for /proc/mounts was introduced with 2.4.19
James Pearson
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [patch] umount loopback filessystems on autofs stop
2004-09-03 9:03 ` James Pearson
@ 2004-09-03 12:30 ` Mike Waychison
0 siblings, 0 replies; 7+ messages in thread
From: Mike Waychison @ 2004-09-03 12:30 UTC (permalink / raw)
To: James Pearson; +Cc: autofs, raven, jmoss
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
James Pearson wrote:
> Mike Waychison wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Joseph V Moss wrote:
>>
>>> Using /proc/mounts could be an issue for anyone trying to use this
>>
>>
>> with a kernel
>>
>>> that truncates /proc/mounts.
>>>
>>
>>
>> Which kernel truncates /proc/mounts?
>>
>> Any recent kernel uses the seq_file interface which guarantees that
>> nothing gets truncated. I have a system up and running that works fine
>> with about 2500 mounts.
>
>
>
> seq_file support for /proc/mounts was introduced with 2.4.19
>
Heh, which is only over two years old :)
Seriously, if your kernel doesn't do this right, the problem doesn't lie
with Jeff's initscript patch: please consider upgrading.
- --
Mike Waychison
Sun Microsystems, Inc.
1 (650) 352-5299 voice
1 (416) 202-8336 voice
http://www.sun.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
NOTICE: The opinions expressed in this email are held by me,
and may not represent the views of Sun Microsystems, Inc.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFBOGPpdQs4kOxk3/MRAkGlAJkBLHG2rYJZS9IbPn/qLP2mPAitPwCfcVv3
NegYJQX2jtqQwNV9STncXnY=
=k5GF
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 7+ messages in thread
* RE: [patch] umount loopback filessystems on autofs stop
@ 2004-09-03 2:46 Lever, Charles
2004-09-03 14:17 ` Jeff Moyer
0 siblings, 1 reply; 7+ messages in thread
From: Lever, Charles @ 2004-09-03 2:46 UTC (permalink / raw)
To: Mike Waychison; +Cc: autofs
i know that some RHEL AS 2.1 kernels truncate /proc/mounts. they have a
single page buffer for /proc/mounts and do not use seq_file. i'm not
sure if this has been addressed in recent updates; maybe jeff knows.
> -----Original Message-----
> From: Mike Waychison [mailto:Michael.Waychison@Sun.COM]
> Sent: Thursday, September 02, 2004 10:31 PM
> To: Joseph V Moss
> Cc: autofs@linux.kernel.org; raven@themaw.net
> Subject: Re: [autofs] [patch] umount loopback filessystems on
> autofs stop
>
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Joseph V Moss wrote:
> > Using /proc/mounts could be an issue for anyone trying to use this
> with a kernel
> > that truncates /proc/mounts.
> >
>
> Which kernel truncates /proc/mounts?
>
> Any recent kernel uses the seq_file interface which
> guarantees that nothing gets truncated. I have a system up
> and running that works fine with about 2500 mounts.
>
>
> >
> >>Hello,
> >>
> >>The netfs scripts for Red Hat go through and umount any loopback
> >>devices before attempting to umount the encompassing
> filesystem. This
> >>seems a
> sane
> >>thing to do (and I have a bug on said issue), so I've added similar
> >>functionality to the autofs init script. Attached is a patch.
> >>Comments encouraged. Testing by debian folks would be nice, too.
> >>
> >>-Jeff
> >>
> >>--- autofs-4.1.3/samples/rc.autofs.in.umount-loopback
> 2004-09-02
> 15:55:21.722705296 -0400
> >>+++ autofs-4.1.3/samples/rc.autofs.in 2004-09-02
> 16:16:09.985940440 -0400
> >>@@ -351,6 +351,31 @@ function signal_automounts()
> >> return $RETVAL
> >> }
> >>
> >>+function umount_loopback()
> >>+{
> >>+ loops=`LC_ALL=C awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/"
> {print $2}' /proc/mounts`
> >>+ automounts=`LC_ALL=C awk '!/^#/ && $1 ~ /^automount/
> {print $2}'
> /proc/mounts`
> >>+
> >>+ for l in $loops
> >>+ do
> >>+ for a in $automounts
> >>+ do
> >>+ command="echo $l | grep -E '^$a'"
> >>+ match=`echo $command | sh`
> >>+ if [ -n "$match" ]; then
> >>+ echo ""
> >>+ echo -n $"Unmounting loopback filesystem $match: "
> >>+ umount -d $match
> >>+ if [ $? -eq 0 ]; then
> >>+ echo "done"
> >>+ else
> >>+ echo "failed ($?)"
> >>+ fi
> >>+ fi
> >>+ done
> >>+ done
> >>+}
> >>+
> >> # Redhat start/stop function.
> >> #
> >> function redhat()
> >>@@ -386,6 +411,7 @@ case "$1" in
> >> if [ -z "`pidofproc $prog`" -a -z "`getmounts`" ]; then
> >> RETVAL=0
> >> else
> >>+ umount_loopback
> >> signal_automounts
> >> RETVAL=$?
> >> fi
> >>@@ -477,6 +503,7 @@ case "$1" in
> >> ;;
> >> stop)
> >> echo 'Stopping automounter.'
> >>+ umount_loopback
> >> start-stop-daemon --stop --quiet --signal USR2 --exec $DAEMON
> >> ;;
> >> reload|restart)
> >>
> >>_______________________________________________
> >>autofs mailing list
> >>autofs@linux.kernel.org
> >>http://linux.kernel.org/mailman/listinfo/autofs
> >>
> >
> >
> > _______________________________________________
> > autofs mailing list
> > autofs@linux.kernel.org
> > http://linux.kernel.org/mailman/listinfo/autofs
>
>
> - --
> Mike Waychison
> Sun Microsystems, Inc.
> 1 (650) 352-5299 voice
> 1 (416) 202-8336 voice
> http://www.sun.com
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> NOTICE: The opinions expressed in this email are held by me,
> and may not represent the views of Sun Microsystems, Inc.
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
>
> iD8DBQFBN9d9dQs4kOxk3/MRAuv7AJ9RwuxkWHgwUkHsW/+C+Qoy/5hzhgCdFoSW
> fM9teS9ID1cEs0ZzVJfgO+g=
> =qIAQ
> -----END PGP SIGNATURE-----
>
> _______________________________________________
> autofs mailing list
> autofs@linux.kernel.org
> http://linux.kernel.org/mailman/listinfo/autof> s
>
^ permalink raw reply [flat|nested] 7+ messages in thread* RE: [patch] umount loopback filessystems on autofs stop
2004-09-03 2:46 Lever, Charles
@ 2004-09-03 14:17 ` Jeff Moyer
0 siblings, 0 replies; 7+ messages in thread
From: Jeff Moyer @ 2004-09-03 14:17 UTC (permalink / raw)
To: Lever, Charles; +Cc: autofs, Mike Waychison
==> Regarding RE: [autofs] [patch] umount loopback filessystems on autofs stop; "Lever, Charles" <cel@netapp.com> adds:
cel> i know that some RHEL AS 2.1 kernels truncate /proc/mounts. they have
cel> a single page buffer for /proc/mounts and do not use seq_file. i'm
cel> not sure if this has been addressed in recent updates; maybe jeff
cel> knows.
AS2.1 does not support autofs4, so there's really no problem there. ;)
-Jeff
>> -----Original Message-----
From> Mike Waychison [mailto:Michael.Waychison@Sun.COM]
Sent> Thursday, September 02, 2004 10:31 PM
To> Joseph V Moss
Cc> autofs@linux.kernel.org; raven@themaw.net
Subject> Re: [autofs] [patch] umount loopback filessystems on
>> autofs stop
>>
>>
>> -----BEGIN PGP SIGNED MESSAGE-----
Hash> SHA1
>> Joseph V Moss wrote: > Using /proc/mounts could be an issue for anyone
>> trying to use this with a kernel > that truncates /proc/mounts.
>> >
>>
>> Which kernel truncates /proc/mounts?
>>
>> Any recent kernel uses the seq_file interface which guarantees that
>> nothing gets truncated. I have a system up and running that works fine
>> with about 2500 mounts.
>>
>>
>> >
>> >>Hello,
>> >>
>> >>The netfs scripts for Red Hat go through and umount any loopback
>> >>devices before attempting to umount the encompassing filesystem. This
>> >>seems a sane >>thing to do (and I have a bug on said issue), so I've
>> added similar >>functionality to the autofs init script. Attached is a
>> patch. >>Comments encouraged. Testing by debian folks would be nice,
>> too.
>> >>
>> >>-Jeff
>> >>
>> >>--- autofs-4.1.3/samples/rc.autofs.in.umount-loopback 2004-09-02
15> 55:21.722705296 -0400
>> >>+++ autofs-4.1.3/samples/rc.autofs.in 2004-09-02
16> 16:09.985940440 -0400
>> >>@@ -351,6 +351,31 @@ function signal_automounts() >> return $RETVAL >>
>> }
>> >>
>> >>+function umount_loopback() >>+{ >>+ loops=`LC_ALL=C awk '!/^#/ && $1
>> ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts` >>+
>> automounts=`LC_ALL=C awk '!/^#/ && $1 ~ /^automount/ {print $2}'
>> /proc/mounts` >>+ >>+ for l in $loops >>+ do >>+ for a in $automounts
>> >>+ do >>+ command="echo $l | grep -E '^$a'" >>+ match=`echo $command |
>> sh` >>+ if [ -n "$match" ]; then >>+ echo "" >>+ echo -n $"Unmounting
>> loopback filesystem $match: " >>+ umount -d $match >>+ if [ $? -eq 0 ];
>> then >>+ echo "done" >>+ else >>+ echo "failed ($?)" >>+ fi >>+ fi >>+
>> done >>+ done >>+} >>+ >> # Redhat start/stop function.
>> >> #
>> >> function redhat() >>@@ -386,6 +411,7 @@ case "$1" in >> if [ -z
>> "`pidofproc $prog`" -a -z "`getmounts`" ]; then >> RETVAL=0 >> else >>+
>> umount_loopback >> signal_automounts >> RETVAL=$? >> fi >>@@ -477,6
>> +503,7 @@ case "$1" in >> ;; >> stop) >> echo 'Stopping automounter.'
>> >>+ umount_loopback >> start-stop-daemon --stop --quiet --signal USR2
>> --exec $DAEMON >> ;; >> reload|restart)
>> >>
>> >>_______________________________________________ >>autofs mailing list
>> >>autofs@linux.kernel.org
>> >>http://linux.kernel.org/mailman/listinfo/autofs
>> >>
>> >
>> >
>> > _______________________________________________ > autofs mailing list
>> > autofs@linux.kernel.org >
>> http://linux.kernel.org/mailman/listinfo/autofs
>>
>>
>> - --
>> Mike Waychison Sun Microsystems, Inc. 1 (650) 352-5299 voice 1 (416)
>> 202-8336 voice
http> //www.sun.com
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
NOTICE> The opinions expressed in this email are held by me,
>> and may not represent the views of Sun Microsystems,
>> Inc. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> -----BEGIN PGP SIGNATURE-----
Version> GnuPG v1.2.4 (GNU/Linux)
>> iD8DBQFBN9d9dQs4kOxk3/MRAuv7AJ9RwuxkWHgwUkHsW/+C+Qoy/5hzhgCdFoSW
>> fM9teS9ID1cEs0ZzVJfgO+g= =qIAQ -----END PGP SIGNATURE-----
>>
>> _______________________________________________ autofs mailing list
>> autofs@linux.kernel.org
http> //linux.kernel.org/mailman/listinfo/autof> s
>>
cel> _______________________________________________ autofs mailing list
cel> autofs@linux.kernel.org
http> //linux.kernel.org/mailman/listinfo/autofs
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2004-09-03 14:17 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-09-02 20:14 [patch] umount loopback filessystems on autofs stop Jeff Moyer
2004-09-02 21:20 ` Joseph V Moss
2004-09-03 2:31 ` Mike Waychison
2004-09-03 9:03 ` James Pearson
2004-09-03 12:30 ` Mike Waychison
-- strict thread matches above, loose matches on Subject: below --
2004-09-03 2:46 Lever, Charles
2004-09-03 14:17 ` Jeff Moyer
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.