Skip to contentSkip to conversation

autofs timeout and large map

DISCUSSION Dannie Stanley 3 participants 14 messages Oct 31, 2005 – Nov 11, 2005 2 messages not in the archive Kernel automounter

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#2 lore raw

Ian Kent has pointed me to this list. I have included my original message thread posted to linux-kernel (sorry for the incorrect posting). Here are my answers to Ian's specific questions:

Quoted 1 line

Distribution?

$ cat /etc/redhat-release
Red Hat Enterprise Linux AS release 3 (Taroon Update 6)
Quoted 1 line

Your version of autofs?

$ rpm -q autofs
autofs-4.1.3-154
Quoted 1 line

Kernel version?

$ uname -a
Linux ... 2.4.21-37.EL #1 Wed Sep 7 13:35:21 EDT 2005 i686 i686 i386 GNU/Linux
Quoted 1 line

Have any autofs4 patches been applied to your kernel?

Standard packaged RedHat kernel, no patches.

Quoted 1 line

What do you have in your master map?

$ cat /etc/auto.master
#/- /etc/auto.virtual --ghost
/- /etc/auto.virtual
Quoted 2 lines

Then you have 1345 offset mount (multi-mount) entries similar to the one above. Each with two offsets. Correct?

True.

Quoted 1 line

How many automount processes end up running after starting autofs?

Two...

$ ps -fC automount |cut -c49-200
CMD
/usr/sbin/automount --timeout=86400 /- file /etc/auto.virtual
/usr/sbin/automount --submount --timeout=86400 /virtual file /etc/auto.virtual

Thanks for your help!
Dannie

---------- Forwarded message ----------
From: Ian Kent <[email protected]>
Date: Oct 28, 2005 10:13 PM
Subject: Re: Fw: autofs timeout and large map
To: Dannie Stanley <[email protected]>
Cc: Andrew Morton <[email protected]>

Thanks for forwarding this Andrew.

Quoted 9 lines · Andrew Morton

fyi..

Begin forwarded message:

Date: Fri, 28 Oct 2005 15:28:05 -0500
From: Dannie Stanley <[email protected]>
To: [email protected]
Subject: autofs timeout and large map

Hi Dannie.

The best place for this query is the autofs list at:
http://linux.kernel.org/mailman/listinfo/autofs

We probably don't need to clutter Andrews inbox further with this thread. I'm sure he has more than enough to deal with.

Quoted 4 lines

I think I may have discovered a bug. But perhaps I overlooked something -- looking for advice.

Perhaps but lets work through it and see.

I some more information about your environment.

Distribution?
Your version of autofs?
Kernel version?
Have any autofs4 patches been applied to yuor kernel?

Quoted 4 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Quoted 8 lines
   /virtual/USERNAME \
      /public --bind 127.0.0.1:/usr/local/files/public \
      /private --bind 127.0.0.1:/usr/local/files/USERNAME

My /etc/auto.virtual file has 1345 entries very similar to the example provided (not wrapped like the example above). Everything works like expected for the first mount. When the mount timeout expires, the mount will not auto remount. When I try to ls, it gives the error:

First, the handling of the --bind option might not parse very well. I think that the option will end up being specified twice for the mount command. Likely not a problem but it is not needed as autofs checks if the mount is local and should automatically call mount with the bind option.

I'd like to understand you automount maps a little more clearly.

What do you have in your master map?
I guess a line that starts with "/-" and specifies "auto.virtual" as the map file. Correct me if I'm wrong.

Then you have 1345 offset mount (multi-mount) entries similar to the one above. Each with two offsets. Correct?

How many automount processes end up running after starting autofs? If I'm correct there should one process handling the /virtual directory.

I know that's not the way it should work but that's being addressed.

Quoted 11 lines

"No such file or directory"

When --ghost is enabled most directories look like this when autofs first starts:

drwxrwxr-x 2 virtual virtual 4096 Oct 28 14:45 private

However after the timeout they look like this:

dr-xr-xr-x 2 root root 0 Oct 28 14:45 private

Nothing wrong with this.

When a filesystem is mounted on a directory you should see the attributes of the mounted filesystem.

When the mount goes away you will see the attributes of the directory as it is within the autofs pseudo filesystem.

Quoted 9 lines

Once I restart autofs it mounts again fine the first time. For now I have set my timeout to 24 hours which functions as a work-around but clutters up my mounted filesystems.

My system is in production and I can't recreate the problem right now lest I interrupt users. I have done my best to recreate the actual directory listings and error messages that I saw but they may not 100% accurate as they are from memory.

That's cool but we will need to collect a debug log. I'll be able to offer more advice when I know what version of autofs you are running.

Perhaps we can duplicate this on another non production system.

Quoted 3 lines

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

I wish I could do that.

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

#3 lore raw

I performed some testing in the early morning hours and this is what I found. The problem seems to be with the --ghost option. When I specify this in my auto.master:

   /- /etc/auto.virtual --ghost

I get the originally reported errors. When I remove the --ghost option it seems to work as expected.

For good measure I also removed the --bind option from all of my entries per Ian's suggestion. He was right they still work without the bind option.

-- Dannie

Quoted 38 lines · Dannie Stanley

Ian Kent has pointed me to this list. I have included my original message thread posted to linux-kernel (sorry for the incorrect posting). Here are my answers to Ian's specific questions:

Quoted 1 line

Distribution?

$ cat /etc/redhat-release
Red Hat Enterprise Linux AS release 3 (Taroon Update 6)
Quoted 1 line

Your version of autofs?

$ rpm -q autofs
autofs-4.1.3-154
Quoted 1 line

Kernel version?

$ uname -a
Linux ... 2.4.21-37.EL #1 Wed Sep 7 13:35:21 EDT 2005 i686 i686 i386 GNU/Linux
Quoted 1 line

Have any autofs4 patches been applied to your kernel?

Standard packaged RedHat kernel, no patches.

Quoted 1 line

What do you have in your master map?

$ cat /etc/auto.master
#/- /etc/auto.virtual --ghost
/- /etc/auto.virtual
Quoted 2 lines

Then you have 1345 offset mount (multi-mount) entries similar to the one above. Each with two offsets. Correct?

True.

Quoted 1 line

How many automount processes end up running after starting autofs?

Two...

$ ps -fC automount |cut -c49-200
CMD
/usr/sbin/automount --timeout=86400 /- file /etc/auto.virtual
/usr/sbin/automount --submount --timeout=86400 /virtual file /etc/auto.virtual

Thanks for your help!
Dannie

maintainer
#4 lore raw
Quoted 9 lines · Dannie Stanley

I performed some testing in the early morning hours and this is what I found. The problem seems to be with the --ghost option. When I specify this in my auto.master:

   /- /etc/auto.virtual --ghost

I get the originally reported errors. When I remove the --ghost option it seems to work as expected.

Btw there are two kernel modules, autofs and autofs4, which do you have loaded?

You can find out with an lsmod. You'll see either autofs or autofs4.

Ian

#5 lore raw

Thanks for the feedback. Here is my answer to your last question:

$ lsmod |grep fs
autofs4                15864   1  (autoclean)

I am assuming it is safe to skip the debug steps since Jeff Moyer has assisted in this diagnosis -- Thanks Jeff!

-- Dannie

Quoted 12 lines · Ian Kent
Quoted 9 lines · Dannie Stanley

I performed some testing in the early morning hours and this is what I found. The problem seems to be with the --ghost option. When I specify this in my auto.master:

   /- /etc/auto.virtual --ghost

I get the originally reported errors. When I remove the --ghost option it seems to work as expected.

Quoted 8 lines

Btw there are two kernel modules, autofs and autofs4, which do you have loaded?

You can find out with an lsmod. You'll see either autofs or autofs4.

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

maintainer
#6 lore raw
In reply to Dannie Stanley
Quoted 11 lines · Dannie Stanley

Ian Kent has pointed me to this list. I have included my original message thread posted to linux-kernel (sorry for the incorrect posting). Here are my answers to Ian's specific questions:

Quoted 1 line

Distribution?

$ cat /etc/redhat-release
Red Hat Enterprise Linux AS release 3 (Taroon Update 6)
Quoted 1 line

Your version of autofs?

$ rpm -q autofs
autofs-4.1.3-154

That's good.

Quoted 4 lines
Quoted 2 lines

Kernel version?

$ uname -a
Linux ... 2.4.21-37.EL #1 Wed Sep 7 13:35:21 EDT 2005 i686 i686 i386 GNU/Linux

The RHEL kernel should be fine also.

Quoted 3 lines
Quoted 2 lines

Have any autofs4 patches been applied to your kernel?

Standard packaged RedHat kernel, no patches.

Yep. Shouldn't need any with RHEL.

Quoted 8 lines
Quoted 2 lines

What do you have in your master map?

$ cat /etc/auto.master
#/- /etc/auto.virtual --ghost
/- /etc/auto.virtual
Quoted 2 lines

Then you have 1345 offset mount (multi-mount) entries similar to the one above. Each with two offsets. Correct?

Quoted 9 lines

True.

Quoted 1 line

How many automount processes end up running after starting autofs?

Two...

$ ps -fC automount |cut -c49-200
CMD
/usr/sbin/automount --timeout=86400 /- file /etc/auto.virtual
/usr/sbin/automount --submount --timeout=86400 /virtual file /etc/auto.virtual

Now about the debug log we need to discover what's going on.

Please add --debug to your master map entry and forward the resulting log.

You need to check that your syslog configuration includes priority debug.

Quoted 135 lines

Thanks for your help!
Dannie

---------- Forwarded message ----------
From: Ian Kent <[email protected]>
Date: Oct 28, 2005 10:13 PM
Subject: Re: Fw: autofs timeout and large map
To: Dannie Stanley <[email protected]>
Cc: Andrew Morton <[email protected]>

Thanks for forwarding this Andrew.

Quoted 9 lines · Andrew Morton

fyi..

Begin forwarded message:

Date: Fri, 28 Oct 2005 15:28:05 -0500
From: Dannie Stanley <[email protected]>
To: [email protected]
Subject: autofs timeout and large map

Hi Dannie.

The best place for this query is the autofs list at:
http://linux.kernel.org/mailman/listinfo/autofs

We probably don't need to clutter Andrews inbox further with this thread. I'm sure he has more than enough to deal with.

Quoted 4 lines

I think I may have discovered a bug. But perhaps I overlooked something -- looking for advice.

Perhaps but lets work through it and see.

I some more information about your environment.

Distribution?
Your version of autofs?
Kernel version?
Have any autofs4 patches been applied to yuor kernel?

Quoted 4 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Quoted 8 lines
   /virtual/USERNAME \
      /public --bind 127.0.0.1:/usr/local/files/public \
      /private --bind 127.0.0.1:/usr/local/files/USERNAME

My /etc/auto.virtual file has 1345 entries very similar to the example provided (not wrapped like the example above). Everything works like expected for the first mount. When the mount timeout expires, the mount will not auto remount. When I try to ls, it gives the error:

First, the handling of the --bind option might not parse very well. I think that the option will end up being specified twice for the mount command. Likely not a problem but it is not needed as autofs checks if the mount is local and should automatically call mount with the bind option.

I'd like to understand you automount maps a little more clearly.

What do you have in your master map?
I guess a line that starts with "/-" and specifies "auto.virtual" as the map file. Correct me if I'm wrong.

Then you have 1345 offset mount (multi-mount) entries similar to the one above. Each with two offsets. Correct?

How many automount processes end up running after starting autofs? If I'm correct there should one process handling the /virtual directory.

I know that's not the way it should work but that's being addressed.

Quoted 11 lines

"No such file or directory"

When --ghost is enabled most directories look like this when autofs first starts:

drwxrwxr-x 2 virtual virtual 4096 Oct 28 14:45 private

However after the timeout they look like this:

dr-xr-xr-x 2 root root 0 Oct 28 14:45 private

Nothing wrong with this.

When a filesystem is mounted on a directory you should see the attributes of the mounted filesystem.

When the mount goes away you will see the attributes of the directory as it is within the autofs pseudo filesystem.

Quoted 9 lines

Once I restart autofs it mounts again fine the first time. For now I have set my timeout to 24 hours which functions as a work-around but clutters up my mounted filesystems.

My system is in production and I can't recreate the problem right now lest I interrupt users. I have done my best to recreate the actual directory listings and error messages that I saw but they may not 100% accurate as they are from memory.

That's cool but we will need to collect a debug log. I'll be able to offer more advice when I know what version of autofs you are running.

Perhaps we can duplicate this on another non production system.

Quoted 3 lines

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

I wish I could do that.

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

#7 lore raw

==> Regarding [autofs] Re: autofs timeout and large map; Ian Kent <[email protected]> adds:

Quoted 8 lines · Dannie Stanley
Quoted 8 lines

Ian Kent has pointed me to this list. I have included my original message thread posted to linux-kernel (sorry for the incorrect posting). Here are my answers to Ian's specific questions:

Quoted 1 line

Distribution? $ cat /etc/redhat-release Red Hat Enterprise Linux AS

release 3 (Taroon Update 6)

Quoted 1 line

Your version of autofs? $ rpm -q autofs autofs-4.1.3-154

raven> That's good.

Quoted 2 lines
Quoted 2 lines
Quoted 1 line

Kernel version? $ uname -a Linux ... 2.4.21-37.EL #1 Wed Sep 7

13:35:21 EDT 2005 i686 i686 i386 GNU/Linux

raven> The RHEL kernel should be fine also.

Quoted 2 lines
Quoted 2 lines
Quoted 1 line

Have any autofs4 patches been applied to your kernel? Standard

packaged RedHat kernel, no patches.

raven> Yep. Shouldn't need any with RHEL.

Quoted 11 lines
Quoted 11 lines
Quoted 1 line

What do you have in your master map? $ cat /etc/auto.master #/-

/etc/auto.virtual --ghost /- /etc/auto.virtual

Quoted 1 line

Then you have 1345 offset mount (multi-mount) entries similar to the >

one above. Each with two offsets. Correct? True.

Quoted 1 line

How many automount processes end up running after starting autofs?

Two... $ ps -fC automount |cut -c49-200 CMD /usr/sbin/automount --timeout=86400 /- file /etc/auto.virtual /usr/sbin/automount --submount --timeout=86400 /virtual file /etc/auto.virtual

raven> Now about the debug log we need to discover what's going on.

raven> Please add --debug to your master map entry and forward the raven> resulting log.

raven> You need to check that your syslog configuration includes priority raven> debug.

This is quite easy to reproduce. Basically, after the expire event, accesses to the /virtual/USERNAME directory do not trigger callouts to the daemon. Here is the expiry message from the logs:

automount[2495]: sigchld: exp 2513 finished, switching from 2 to 1 automount[2495]: get_pkt: state 2, next 1
automount[2495]: st_ready(): state = 2
automount[2495]: sig 14 switching from 1 to 2
automount[2495]: get_pkt: state 1, next 2
automount[2495]: st_expire(): state = 1
automount[2495]: expire_proc: exp_proc=2515
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 22, name jmoyer/private automount[2516]: expiring path /virtual/jmoyer/private
automount[2516]: umount_multi: path=/virtual/jmoyer/private incl=1 automount[2516]: umount_multi: unmounting dir=/virtual/jmoyer/private automount[2516]: expired /virtual/jmoyer/private
automount[2495]: handle_child: got pid 2516, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2516: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=22
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 23, name jmoyer/public automount[2518]: expiring path /virtual/jmoyer/public
automount[2518]: umount_multi: path=/virtual/jmoyer/public incl=1 automount[2518]: umount_multi: unmounting dir=/virtual/jmoyer/public automount[2518]: expired /virtual/jmoyer/public
automount[2495]: handle_child: got pid 2518, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2518: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=23
automount[2495]: handle_child: got pid 2515, sig 0 (0), stat 0 automount[2495]: sigchld: exp 2515 finished, switching from 2 to 1

What you are left with is a directory hierarchy that looks like this:

/virtual
/jmoyer
/private
/public

Clearly the private and public directories should be gone! As mentioned, an ls -l in the /virtual/jmoyer directory will look like so:

# ls -l /virtual/jmoyer/
total 0
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 private
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 public

Sorry I don't have more information to give at this time. Perhaps this will stir something in your memory, though. ;)

-Jeff

maintainer
#8 lore raw
Quoted 41 lines · Jeff Moyer

==> Regarding [autofs] Re: autofs timeout and large map; Ian Kent <[email protected]> adds:

Quoted 8 lines · Dannie Stanley
Quoted 8 lines

Ian Kent has pointed me to this list. I have included my original message thread posted to linux-kernel (sorry for the incorrect posting). Here are my answers to Ian's specific questions:

Quoted 1 line

Distribution? $ cat /etc/redhat-release Red Hat Enterprise Linux AS

release 3 (Taroon Update 6)

Quoted 1 line

Your version of autofs? $ rpm -q autofs autofs-4.1.3-154

raven> That's good.

Quoted 2 lines
Quoted 2 lines
Quoted 1 line

Kernel version? $ uname -a Linux ... 2.4.21-37.EL #1 Wed Sep 7

13:35:21 EDT 2005 i686 i686 i386 GNU/Linux

raven> The RHEL kernel should be fine also.

Quoted 2 lines
Quoted 2 lines
Quoted 1 line

Have any autofs4 patches been applied to your kernel? Standard

packaged RedHat kernel, no patches.

raven> Yep. Shouldn't need any with RHEL.

Quoted 11 lines
Quoted 11 lines
Quoted 1 line

What do you have in your master map? $ cat /etc/auto.master #/-

/etc/auto.virtual --ghost /- /etc/auto.virtual

Quoted 1 line

Then you have 1345 offset mount (multi-mount) entries similar to the >

one above. Each with two offsets. Correct? True.

Quoted 1 line

How many automount processes end up running after starting autofs?

Two... $ ps -fC automount |cut -c49-200 CMD /usr/sbin/automount --timeout=86400 /- file /etc/auto.virtual /usr/sbin/automount --submount --timeout=86400 /virtual file /etc/auto.virtual

raven> Now about the debug log we need to discover what's going on.

raven> Please add --debug to your master map entry and forward the raven> resulting log.

Quoted 6 lines

raven> You need to check that your syslog configuration includes priority raven> debug.

This is quite easy to reproduce. Basically, after the expire event, accesses to the /virtual/USERNAME directory do not trigger callouts to the daemon. Here is the expiry message from the logs:

OK. I'll have a go at reproducing it.

I don't have the exact versions but hopefully the problem will be evident in other versions as well.

Would it be possible for you the put the source rpm into your people area please.

Quoted 36 lines

automount[2495]: sigchld: exp 2513 finished, switching from 2 to 1 automount[2495]: get_pkt: state 2, next 1
automount[2495]: st_ready(): state = 2
automount[2495]: sig 14 switching from 1 to 2
automount[2495]: get_pkt: state 1, next 2
automount[2495]: st_expire(): state = 1
automount[2495]: expire_proc: exp_proc=2515
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 22, name jmoyer/private automount[2516]: expiring path /virtual/jmoyer/private
automount[2516]: umount_multi: path=/virtual/jmoyer/private incl=1 automount[2516]: umount_multi: unmounting dir=/virtual/jmoyer/private automount[2516]: expired /virtual/jmoyer/private
automount[2495]: handle_child: got pid 2516, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2516: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=22
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 23, name jmoyer/public automount[2518]: expiring path /virtual/jmoyer/public
automount[2518]: umount_multi: path=/virtual/jmoyer/public incl=1 automount[2518]: umount_multi: unmounting dir=/virtual/jmoyer/public automount[2518]: expired /virtual/jmoyer/public
automount[2495]: handle_child: got pid 2518, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2518: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=23
automount[2495]: handle_child: got pid 2515, sig 0 (0), stat 0 automount[2495]: sigchld: exp 2515 finished, switching from 2 to 1

What you are left with is a directory hierarchy that looks like this:

/virtual
/jmoyer
/private
/public

Quoted 7 lines

Clearly the private and public directories should be gone! As mentioned, an ls -l in the /virtual/jmoyer directory will look like so:

# ls -l /virtual/jmoyer/
total 0
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 private
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 public

If the --ghost option is used these directories should be creted when the map is first loaded and remain.

The fact that they are empty should be the mount trigger. This might be another example of the kernel problem we are currently working on.

A debug trace of the subsequent mount failure should confirm that one way or the other.

Ian

maintainer
#9 lore raw
Quoted 53 lines · Ian Kent
Quoted 3 lines · Jeff Moyer

This is quite easy to reproduce. Basically, after the expire event, accesses to the /virtual/USERNAME directory do not trigger callouts to the daemon. Here is the expiry message from the logs:

OK. I'll have a go at reproducing it.

I don't have the exact versions but hopefully the problem will be evident in other versions as well.

Would it be possible for you the put the source rpm into your people area please.

Quoted 39 lines

automount[2495]: sigchld: exp 2513 finished, switching from 2 to 1 automount[2495]: get_pkt: state 2, next 1
automount[2495]: st_ready(): state = 2
automount[2495]: sig 14 switching from 1 to 2
automount[2495]: get_pkt: state 1, next 2
automount[2495]: st_expire(): state = 1
automount[2495]: expire_proc: exp_proc=2515
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 22, name jmoyer/private automount[2516]: expiring path /virtual/jmoyer/private
automount[2516]: umount_multi: path=/virtual/jmoyer/private incl=1 automount[2516]: umount_multi: unmounting dir=/virtual/jmoyer/private automount[2516]: expired /virtual/jmoyer/private
automount[2495]: handle_child: got pid 2516, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2516: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=22
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 23, name jmoyer/public automount[2518]: expiring path /virtual/jmoyer/public
automount[2518]: umount_multi: path=/virtual/jmoyer/public incl=1 automount[2518]: umount_multi: unmounting dir=/virtual/jmoyer/public automount[2518]: expired /virtual/jmoyer/public
automount[2495]: handle_child: got pid 2518, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2518: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=23
automount[2495]: handle_child: got pid 2515, sig 0 (0), stat 0 automount[2495]: sigchld: exp 2515 finished, switching from 2 to 1

What you are left with is a directory hierarchy that looks like this:

/virtual
/jmoyer
/private
/public

Clearly the private and public directories should be gone! As mentioned, an ls -l in the /virtual/jmoyer directory will look like so:

Quoted 7 lines
Quoted 4 lines
# ls -l /virtual/jmoyer/
total 0
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 private
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 public

If the --ghost option is used these directories should be creted when the map is first loaded and remain.

Sorry, this is a multi-mount entry isn't it.

Then of course you are right Jeff.

Dannie can you confirm that your map looks like:

/virtual/<user> /private <mount> /public <mount>

Ian

#10 lore raw

Ian:

My auto.virtual file looks like this without the line breaks:

  /virtual/USERNAME \
     /public --bind 127.0.0.1:/usr/local/files/public \
     /private --bind 127.0.0.1:/usr/local/files/USERNAME

-- Dannie

Quoted 65 lines · Ian Kent
Quoted 53 lines · Ian Kent
Quoted 3 lines · Jeff Moyer

This is quite easy to reproduce. Basically, after the expire event, accesses to the /virtual/USERNAME directory do not trigger callouts to the daemon. Here is the expiry message from the logs:

OK. I'll have a go at reproducing it.

I don't have the exact versions but hopefully the problem will be evident in other versions as well.

Would it be possible for you the put the source rpm into your people area please.

Quoted 39 lines

automount[2495]: sigchld: exp 2513 finished, switching from 2 to 1 automount[2495]: get_pkt: state 2, next 1
automount[2495]: st_ready(): state = 2
automount[2495]: sig 14 switching from 1 to 2
automount[2495]: get_pkt: state 1, next 2
automount[2495]: st_expire(): state = 1
automount[2495]: expire_proc: exp_proc=2515
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 22, name jmoyer/private automount[2516]: expiring path /virtual/jmoyer/private
automount[2516]: umount_multi: path=/virtual/jmoyer/private incl=1 automount[2516]: umount_multi: unmounting dir=/virtual/jmoyer/private automount[2516]: expired /virtual/jmoyer/private
automount[2495]: handle_child: got pid 2516, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2516: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=22
automount[2495]: handle_packet: type = 2
automount[2495]: handle_packet_expire_multi: token 23, name jmoyer/public automount[2518]: expiring path /virtual/jmoyer/public
automount[2518]: umount_multi: path=/virtual/jmoyer/public incl=1 automount[2518]: umount_multi: unmounting dir=/virtual/jmoyer/public automount[2518]: expired /virtual/jmoyer/public
automount[2495]: handle_child: got pid 2518, sig 0 (0), stat 0 automount[2495]: sig_child: found pending iop pid 2518: signalled 0 (sig 0), exit status 0 automount[2495]: send_ready: token=23
automount[2495]: handle_child: got pid 2515, sig 0 (0), stat 0 automount[2495]: sigchld: exp 2515 finished, switching from 2 to 1

What you are left with is a directory hierarchy that looks like this:

/virtual
/jmoyer
/private
/public

Clearly the private and public directories should be gone! As mentioned, an ls -l in the /virtual/jmoyer directory will look like so:

Quoted 7 lines
Quoted 4 lines
# ls -l /virtual/jmoyer/
total 0
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 private
dr-xr-xr-x    2 root     root            0 Nov  1 07:16 public

If the --ghost option is used these directories should be creted when the map is first loaded and remain.

Sorry, this is a multi-mount entry isn't it.

Quoted 8 lines

Then of course you are right Jeff.

Dannie can you confirm that your map looks like:

/virtual/<user> /private <mount> /public <mount>

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

maintainer
#11 lore raw
In reply to Dannie Stanley
Quoted 3 lines · Dannie Stanley
Quoted 3 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Ummmm ....

Oops looks like there's a check that stop this from working correctly for direct mounts.

There was probably a reason for it so let me think about it for a bit and I'll put together a patch.

In the mean time you could work around this using an indirect map like:

In the master:

/virtual /etc/auto.virtual

In auto.virtual:

USERNAME /public localhost:/dir1 /private localhost:/dir2

for each of your USERNAME entries.

or you could use a direct map like:

/virtual/USERNAME/private localhost:/dir1
/virtual/USERNAME/public localhost:/dir2

for each of your USERNAME entries.

Ian

#12 lore raw

Ian:

I am glad you were able to find the bug. The work around I am currently using works ok for now so I probably won't change my configuration. Though I appreciate the effort to give me other options.

Thanks,
Dannie

Quoted 26 lines · Ian Kent
Quoted 3 lines · Dannie Stanley
Quoted 3 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Ummmm ....

Oops looks like there's a check that stop this from working correctly for direct mounts.

There was probably a reason for it so let me think about it for a bit and I'll put together a patch.

In the mean time you could work around this using an indirect map like:

In the master:

/virtual /etc/auto.virtual

In auto.virtual:

USERNAME /public localhost:/dir1 /private localhost:/dir2

for each of your USERNAME entries.

Quoted 10 lines

or you could use a direct map like:

/virtual/USERNAME/private localhost:/dir1
/virtual/USERNAME/public localhost:/dir2

for each of your USERNAME entries.

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

maintainer
#13 lore raw
Quoted 6 lines · Dannie Stanley

Ian:

I am glad you were able to find the bug. The work around I am currently using works ok for now so I probably won't change my configuration. Though I appreciate the effort to give me other options.

I'll need to fix it for 4.1.3 and 4.1.4.
This case is looking a little difficult so it may take a while.

Quoted 49 lines

Thanks,
Dannie

Quoted 26 lines · Ian Kent
Quoted 3 lines · Dannie Stanley
Quoted 3 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Ummmm ....

Oops looks like there's a check that stop this from working correctly for direct mounts.

There was probably a reason for it so let me think about it for a bit and I'll put together a patch.

In the mean time you could work around this using an indirect map like:

In the master:

/virtual /etc/auto.virtual

In auto.virtual:

USERNAME /public localhost:/dir1 /private localhost:/dir2

for each of your USERNAME entries.

Quoted 10 lines

or you could use a direct map like:

/virtual/USERNAME/private localhost:/dir1
/virtual/USERNAME/public localhost:/dir2

for each of your USERNAME entries.

Ian

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

maintainer
#14 lore raw

On Thu, 3 Nov 2005, Ian Kent wrote:

Some bad news I'm afraid Dannie.

Quoted 11 lines · Dannie Stanley
Quoted 7 lines

Ian:

I am glad you were able to find the bug. The work around I am currently using works ok for now so I probably won't change my configuration. Though I appreciate the effort to give me other options.

I'll need to fix it for 4.1.3 and 4.1.4.
This case is looking a little difficult so it may take a while.

Having looked more closely at this I found that it is actually another limitation of the pseudo direct mount implementation in autofs-4.1. The effort to change would be fairly significant both in the daemon and the kernel module.

Since this and other limitations of the 4.1 implementation will be fixed in autofs-5.0 I really don't think the time would be well spent as I need it for 5.0.

So there would have to be a very compelling case.

Sorry.

Perhaps you could test out the first alpha when I finally get it out?

Quoted 44 lines
Quoted 44 lines

Thanks,
Dannie

Quoted 36 lines · Ian Kent
Quoted 3 lines · Dannie Stanley
Quoted 3 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Ummmm ....

Oops looks like there's a check that stop this from working correctly for direct mounts.

There was probably a reason for it so let me think about it for a bit and I'll put together a patch.

In the mean time you could work around this using an indirect map like:

In the master:

/virtual /etc/auto.virtual

In auto.virtual:

USERNAME /public localhost:/dir1 /private localhost:/dir2

for each of your USERNAME entries.

or you could use a direct map like:

/virtual/USERNAME/private localhost:/dir1
/virtual/USERNAME/public localhost:/dir2

for each of your USERNAME entries.

Ian

Quoted 8 lines
Quoted 6 lines

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

--

   ,-._|\    Ian Kent
  /      \   Perth, Western Australia
  *_.--._/   E-mail: [email protected]
        v    Web: http://themaw.net/
#15 lore raw

Ian:

My work around is doing fine. Thanks for the update.

-- Dannie

Quoted 84 lines · Ian Kent

On Thu, 3 Nov 2005, Ian Kent wrote:

Some bad news I'm afraid Dannie.

Quoted 11 lines · Dannie Stanley
Quoted 7 lines

Ian:

I am glad you were able to find the bug. The work around I am currently using works ok for now so I probably won't change my configuration. Though I appreciate the effort to give me other options.

I'll need to fix it for 4.1.3 and 4.1.4.
This case is looking a little difficult so it may take a while.

Having looked more closely at this I found that it is actually another limitation of the pseudo direct mount implementation in autofs-4.1. The effort to change would be fairly significant both in the daemon and the kernel module.

Since this and other limitations of the 4.1 implementation will be fixed in autofs-5.0 I really don't think the time would be well spent as I need it for 5.0.

So there would have to be a very compelling case.

Sorry.

Perhaps you could test out the first alpha when I finally get it out?

Quoted 44 lines
Quoted 44 lines

Thanks,
Dannie

Quoted 36 lines · Ian Kent
Quoted 3 lines · Dannie Stanley
Quoted 3 lines

I have generated a DIRECT map file and defined it in /etc/auto.master.
My direct mappings look like this (located in /etc/auto.virtual):

Ummmm ....

Oops looks like there's a check that stop this from working correctly for direct mounts.

There was probably a reason for it so let me think about it for a bit and I'll put together a patch.

In the mean time you could work around this using an indirect map like:

In the master:

/virtual /etc/auto.virtual

In auto.virtual:

USERNAME /public localhost:/dir1 /private localhost:/dir2

for each of your USERNAME entries.

or you could use a direct map like:

/virtual/USERNAME/private localhost:/dir1
/virtual/USERNAME/public localhost:/dir2

for each of your USERNAME entries.

Ian

Quoted 8 lines
Quoted 6 lines

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann

Quoted 6 lines

--

   ,-._|\    Ian Kent
  /      \   Perth, Western Australia
  *_.--._/   E-mail: [email protected]
        v    Web: http://themaw.net/

--
Dannie M. Stanley

"The Feynman Problem-Solving Algorithm: (1) write down the problem;
(2) think very hard; (3) write down the answer." - Murray Gell-Mann