Skip to contentSkip to conversation

automount doesn't reread file map, automount terminates on empty file map

DISCUSSION Thomas Jahns 3 participants 9 messages Dec 11, 2004 – Dec 20, 2004 Kernel automounter

lore
Share & tools
lore
#0 lore raw

Hi,

I'm currently experimenting with a script-created map and face the following two problems:

- After the script rewrites the map, I need to kill -HUP the corresponding automount to reread the map, the documentation says the daemon would read a changed map on next access of the directory?

- Unfortunately there exists the possibility, that the script might leave the map empty. The corresponding automount process will then exit. Is there perhaps some comment convention or anything else I could do to keep the automount process running?

How can I solve this? The man-pages don't document the exit on empty map (or at least I haven't found a corresponding statement). What exactly triggers the rereading of file maps?

Thanks for any input on how to improve on the described situations.

Regards,
Thomas Jahns

Signature · 3 lines
"Computers are good at following instructions,
but not at reading your mind."
D. E. Knuth, The TeXbook, Addison-Wesley 1984, 1986, 1996, p. 9
maintainer
#1 lore raw
Quoted 8 lines · Thomas Jahns

Hi,

I'm currently experimenting with a script-created map and face the following two problems:

- After the script rewrites the map, I need to kill -HUP the corresponding automount to reread the map, the documentation says the daemon would read a changed map on next access of the directory?

Known problem with 4.1.
But what vesion are you running?

Quoted 5 lines

- Unfortunately there exists the possibility, that the script might leave the map empty. The corresponding automount process will then exit. Is there perhaps some comment convention or anything else I could do to keep the automount process running?

Known problem but not yet fixed.

Ian

#2 lore raw

Hi,

Quoted 11 lines · Ian Kent
Quoted 7 lines · Thomas Jahns

I'm currently experimenting with a script-created map and face the following two problems:

- After the script rewrites the map, I need to kill -HUP the corresponding automount to reread the map, the documentation says the daemon would read a changed map on next access of the directory?

Known problem with 4.1.
But what vesion are you running?

sorry I forgot to mention some essential information it seems, I use the 4.1.3-6 debian package.

Quoted 7 lines
Quoted 5 lines

- Unfortunately there exists the possibility, that the script might leave the map empty. The corresponding automount process will then exit. Is there perhaps some comment convention or anything else I could do to keep the automount process running?

Known problem but not yet fixed.

Okay, could I perhaps add some dummy entry? Any device/location I could specify that would never ever get mounted?

Regards,
Thomas Jahns

Signature · 3 lines
"Computers are good at following instructions,
but not at reading your mind."
D. E. Knuth, The TeXbook, Addison-Wesley 1984, 1986, 1996, p. 9
maintainer
#3 lore raw
Quoted 12 lines · Thomas Jahns

Hi,

Quoted 9 lines · Ian Kent
Quoted 7 lines · Thomas Jahns

I'm currently experimenting with a script-created map and face the following two problems:

- After the script rewrites the map, I need to kill -HUP the corresponding automount to reread the map, the documentation says the daemon would read a changed map on next access of the directory?

Quoted 5 lines
Quoted 2 lines

Known problem with 4.1.
But what vesion are you running?

sorry I forgot to mention some essential information it seems, I use the 4.1.3-6 debian package.

I can't remember the patches that Arthur included in the deb (Arthur?). I'll check if we can't raise him.

A further difficulty with this is that to fix this problem a small kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

Quoted 11 lines
Quoted 8 lines
Quoted 6 lines

- Unfortunately there exists the possibility, that the script might leave the map empty. The corresponding automount process will then exit. Is there perhaps some comment convention or anything else I could do to keep the automount process running?

Known problem but not yet fixed.

Okay, could I perhaps add some dummy entry? Any device/location I could specify that would never ever get mounted?

Any bogus entry should be OK.

Must admint I've known about this for a while but it slipped through the cracks. Now is difficult as there have been so many changes, making patches is going to start to be difficult to get them to apply. So I'm likely not to add much to the current code until after I make a new release.

Ian

#4 lore raw

Hi again,

Quoted 2 lines · Ian Kent

A further difficulty with this is that to fix this problem a small
kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

I'm currently running a kernel compiled from debian kernel-source-2.6.9_2.6.9-2 package.

Thomas Jahns

Signature · 3 lines
"Computers are good at following instructions,
but not at reading your mind."
D. E. Knuth, The TeXbook, Addison-Wesley 1984, 1986, 1996, p. 9
#5 lore raw
Quoted 7 lines · Thomas Jahns

Hi again,

Quoted 2 lines · Ian Kent

A further difficulty with this is that to fix this problem a small
kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

I'm currently running a kernel compiled from debian kernel- source-2.6.9_2.6.9-2 package.

I'm not good with Debian packages but I had a go at applying the map update patch to the source deb.

I've placed the source deb package files in
http://www.kernel.org/pub/linux/kernel/people/raven/deb
if your interested.

I had to back out an LDAP entries limit patch as the map update patch changes the LDAP lookup module. If that's not a problem for you try it out.

I haven't tested this at all except to build a deb from it on my system. The patch itself has been tested previously.

The kernel patch that is needed for it to function properly is there also.

Ian

#6 lore raw
Quoted 16 lines · On Mon, 20 Dec 2004 [email protected] wrote:
Quoted 7 lines · Thomas Jahns

Hi again,

Quoted 2 lines · Ian Kent

A further difficulty with this is that to fix this problem a small
kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

I'm currently running a kernel compiled from debian kernel- source-2.6.9_2.6.9-2 package.

I'm not good with Debian packages but I had a go at applying the map update patch to the source deb.

I've placed the source deb package files in
http://www.kernel.org/pub/linux/kernel/people/raven/deb
if your interested.

Quoted 5 lines

I had to back out an LDAP entries limit patch as the map update patch changes the LDAP lookup module. If that's not a problem for you try it out.

I haven't tested this at all except to build a deb from it on my system. The patch itself has been tested previously.

The kernel patch that is needed for it to function properly is there also.

btw kernel.org signes and creates bz2s so you need

autofs-4.1.3-7.dsc
autofs-4.1.3-7.diff.gz
autofs-4.1.3-7.orig.tar.gz

and

linux-2.6.7-autofs4-lookup.patch

only for this.

Quoted 5 lines

Ian

--

   ,-._|\    Ian Kent
  /      \   Perth, Western Australia
  *_.--._/   E-mail: [email protected]
        v    Web: http://themaw.net/
#7 lore raw
In reply to Ian Kent
Quoted 7 lines · Ian Kent
Quoted 3 lines

sorry I forgot to mention some essential information it seems, I use the 4.1.3-6 debian package.

I can't remember the patches that Arthur included in the deb (Arthur?). I'll check if we can't raise him.

Package is currently in adoption, Arthur missed time to continue with proper maintainance. Recently, me and Steinar Gunderson worked to have the package in decent shape for sarge.

Quoted 4 lines

A further difficulty with this is that to fix this problem a small kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

That's an interesting hint. I suspect our kernel team could be not aware of it. Do you have a pointer to the patch somewhere?

Signature · 1 line
Francesco P. Lovergine
maintainer
#8 lore raw
Quoted 13 lines · Francesco P. Lovergine
Quoted 7 lines · Ian Kent
Quoted 3 lines

sorry I forgot to mention some essential information it seems, I use the 4.1.3-6 debian package.

I can't remember the patches that Arthur included in the deb (Arthur?). I'll check if we can't raise him.

Package is currently in adoption, Arthur missed time to continue with proper maintainance. Recently, me and Steinar Gunderson worked to have the package in decent shape for sarge.

Quoted 7 lines
Quoted 4 lines

A further difficulty with this is that to fix this problem a small kernel patch was needed. But what kernel are you using (ie. are the latest kernel updates included in it)?

That's an interesting hint. I suspect our kernel team could be not aware of it. Do you have a pointer to the patch somewhere?

Probably not.

And there's another problem in that compatible ioctls for archs that are 32bit <-> 64bit are not present.

I'll be submiting a bug report for both soonish.

Mean time I'll put the patch on kernel.org in the usual place tomorrow.

Ian