Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #230829 > unrolled thread

Too many levels of symbolic links

Back to article view | Back to linux.debian.user

Contents

  Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-19 10:10 +0100
    Re: Too many levels of symbolic links Greg Wooledge <wooledg@eeg.ccf.org> - 2021-01-19 13:50 +0100
      Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-19 14:50 +0100
        Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-19 15:50 +0100
          Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-19 16:00 +0100
            Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-19 16:10 +0100
              Re: Too many levels of symbolic links Greg Wooledge <wooledg@eeg.ccf.org> - 2021-01-19 17:20 +0100
                Re: Too many levels of symbolic links David Wright <deblis@lionunicorn.co.uk> - 2021-01-19 18:10 +0100
                  Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 08:50 +0100
                  [SOLVED] Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 08:50 +0100
                  Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 09:50 +0100
                    Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-20 10:20 +0100
                      Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 10:50 +0100
                        Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-20 11:00 +0100
                          Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 11:20 +0100
                            Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 11:30 +0100
                            Bluewin bouncing my mails [was: Too many levels of symbolic links] tomas@tuxteam.de - 2021-01-20 11:30 +0100
                              Re: Bluewin bouncing my mails tomas@tuxteam.de - 2021-01-20 12:50 +0100
                            Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-20 11:30 +0100
                        Re: Too many levels of symbolic links Erwan David <erwan@rail.eu.org> - 2021-01-20 11:40 +0100
                          Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-20 11:40 +0100
                          Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 12:10 +0100
                    Re: Too many levels of symbolic links Stefan Monnier <monnier@iro.umontreal.ca> - 2021-01-20 16:20 +0100
                      Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-20 16:40 +0100
                        Re: Too many levels of symbolic links Stefan Monnier <monnier@iro.umontreal.ca> - 2021-01-20 22:10 +0100
                          Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-21 08:10 +0100
                          Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-21 08:40 +0100
                            Re: Too many levels of symbolic links Andrei POPESCU <andreimpopescu@gmail.com> - 2021-01-21 08:50 +0100
                              Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-22 08:10 +0100
                                Re: Too many levels of symbolic links Vincent Lefevre <vincent@vinc17.net> - 2021-01-22 13:50 +0100
                                  Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-23 09:40 +0100
                                    Re: Too many levels of symbolic links Vincent Lefevre <vincent@vinc17.net> - 2021-01-23 16:20 +0100
                                Re: Too many levels of symbolic links Andrei POPESCU <andreimpopescu@gmail.com> - 2021-01-22 14:00 +0100
                                  remove me from this list J B Martin <bryant@ab-domain.com> - 2021-01-22 14:10 +0100
                                    Re: remove me from this list Charles Curley <charlescurley@charlescurley.com> - 2021-01-22 18:50 +0100
                                    Re: remove me from this list "Andrew M.A. Cater" <amacater@einval.com> - 2021-01-23 11:00 +0100
                                  Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-23 09:40 +0100
                            Re: Too many levels of symbolic links David Wright <deblis@lionunicorn.co.uk> - 2021-01-22 03:50 +0100
                              Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-22 08:10 +0100
                                Re: Too many levels of symbolic links David Wright <deblis@lionunicorn.co.uk> - 2021-01-22 20:30 +0100
                                  Re: Too many levels of symbolic links Dan Ritter <dsr@randomstring.org> - 2021-01-22 21:10 +0100
                                  Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-23 15:40 +0100
              Re: Too many levels of symbolic links <tomas@tuxteam.de> - 2021-01-19 17:20 +0100
                Re: Too many levels of symbolic links steve <dlist@bluewin.ch> - 2021-01-19 18:00 +0100

Page 1 of 3  [1] 2 3  Next page →

#230829 — Too many levels of symbolic links

Hi,

When I boot the machine (5.10.0-1-amd64), I get:

Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[579]: sdd5: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[618]: sdg1: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[608]: sdd1: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[576]: sdc6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[626]: sdd6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[623]: sdc1: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[609]: sdc5: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[608]: sdd1: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[576]: sdc6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[579]: sdd5: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[626]: sdd6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:47:52 box systemd-udevd[623]: sdc1: Failed to update device symlinks: Too many levels of symbolic links

These partitions are part of Raid1 array:

cat /proc/mdstat

Personalities : [raid1] [linear] [multipath] 
md2 : active raid1 sdc6[5] sdg6[3] sdd6[4]
       97589120 blocks super 1.2 [3/3] [UUU]
       
md0 : active raid1 sdc1[5] sdg1[3] sdd1[4]
       19514240 blocks super 1.2 [3/3] [UUU]
       
md1 : active raid1 sdc5[5] sdg5[3] sdd5[4]
       117120896 blocks super 1.2 [3/3] [UUU]

The system seems to work alright but these messages are pretty scary. I'm seing
them since months and with different kernel versions.

I have spent a lot of time trying to troubleshoot this but without any success yet.

Help would be very much appreciated.

Thanks

Steve

[toc] | [next] | [standalone]

#230836

On Tue, Jan 19, 2021 at 10:03:12AM +0100, steve wrote:
> Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
> Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links

Look for a symlink that points to itself, presumably in either /dev or
/etc/udev.  If you can't find one, then look for a symlink that points
to an earlier point in the path to itself, creating a loop.

[toc] | [prev] | [next] | [standalone]

#230840

[Multipart message — attachments visible in raw view] — view raw

On Tue, Jan 19, 2021 at 07:41:13AM -0500, Greg Wooledge wrote:
> On Tue, Jan 19, 2021 at 10:03:12AM +0100, steve wrote:
> > Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
> > Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
> 
> Look for a symlink that points to itself, presumably in either /dev or
> /etc/udev.  If you can't find one, then look for a symlink that points
> to an earlier point in the path to itself, creating a loop.

FWIW GNU find is said to detect circular links. Try

  find $YOUR_SUSPECT_ROOT -follow -printf ""

(the printf is there to quench the noise). It should complain
when it detects a cycle.

Warning: untested.

Cheers
 - t

[toc] | [prev] | [next] | [standalone]

#230843

Le 19-01-2021, à 14:49:36 +0100, tomas@tuxteam.de a écrit :

>On Tue, Jan 19, 2021 at 07:41:13AM -0500, Greg Wooledge wrote:
>> On Tue, Jan 19, 2021 at 10:03:12AM +0100, steve wrote:
>> > Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
>> > Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
>>
>> Look for a symlink that points to itself, presumably in either /dev or
>> /etc/udev.  If you can't find one, then look for a symlink that points
>> to an earlier point in the path to itself, creating a loop.
>
>FWIW GNU find is said to detect circular links. Try
>
>  find $YOUR_SUSPECT_ROOT -follow -printf ""

Ok, I found some circular links:

find /usr -follow -printf ""
find: Boucle détectée dans le système de fichiers ; « ‘/usr/bin/X11’ » est dans la même boucle que ‘/usr/bin’.

ls -l /usr/bin/X11
lrwxrwxrwx 1 root root 1  6 mai  2013 /usr/bin/X11 -> .


find /sys -follow -printf ""
find: Boucle détectée dans le système de fichiers ; « ‘/sys/kernel/iommu_groups/55/devices/0000:64:0a.5/subsystem/devices/0000:16:0e.2/iommu_group/devices/0000:16:0e.0/firmware_node/wakeup/wakeup44/subsystem/wakeup58/device/physical_node/i2c-3/subsystem/devices/i2c-0/firmware_node/physical_node/iTCO_wdt/subsystem/devices/PNP0C14:00/firmware_node/subsystem/devices/LNXCPU:13/thermal_cooling/subsystem/cooling_device13/device/physical_node/node0/cpu21/driver/cpu26/subsystem/devices/cpu15/firmware_node/thermal_cooling/subsystem’ » est dans la même boucle que ‘/sys/kernel/iommu_groups/55/devices/0000:64:0a.5/subsystem/devices/0000:16:0e.2/iommu_group/devices/0000:16:0e.0/firmware_node/wakeup/wakeup44/subsystem/wakeup58/device/physical_node/i2c-3/subsystem/devices/i2c-0/firmware_node/physical_node/iTCO_wdt/subsystem/devices/PNP0C14:00/firmware_node/subsystem/devices/LNXCPU:13/thermal_cooling/subsystem’.

And a lot of others very similar.


find /dev -follow -printf ""
find: Boucle détectée dans le système de fichiers ; « ‘/dev/fd/3/proc/2669/root’ » est dans la même boucle que ‘/dev/fd/3’.
find: Boucle détectée dans le système de fichiers ; « ‘/dev/fd/3/proc/2670/task/2670/cwd’ » est dans la même boucle que ‘/dev/fd/3’.
find: Boucle détectée dans le système de fichiers ; « ‘/dev/fd/3/proc/2670/task/2670/root’ » est dans la même boucle que ‘/dev/fd/3’.

and a lot more like that.

find /proc -follow -printf ""
ind: Boucle détectée dans le système de fichiers ; « ‘/proc/self/task/65823/fd/3/sys/kernel/iommu_groups/55/devices/0000:64:0a.5/subsystem/devices/0000:16:0e.2/iommu_group/devices/0000:16:0e.0/firmware_node/wakeup/wakeup44/subsystem/wakeup58/device/physical_node/i2c-3/subsystem/devices/i2c-0/firmware_node/physical_node/iTCO_wdt/subsystem/devices/PNP0C14:00/firmware_node/subsystem/devices/LNXCPU:13/thermal_cooling/subsystem/cooling_device13/device/physical_node/node0/cpu21/driver/cpu5/subsystem/devices/cpu19/firmware_node’ » est dans la même boucle que ‘/proc/self/task/65823/fd/3/sys/kernel/iommu_groups/55/devices/0000:64:0a.5/subsystem/devices/0000:16:0e.2/iommu_group/devices/0000:16:0e.0/firmware_node/wakeup/wakeup44/subsystem/wakeup58/device/physical_node/i2c-3/subsystem/devices/i2c-0/firmware_node/physical_node/iTCO_wdt/subsystem/devices/PNP0C14:00/firmware_node/subsystem/devices/LNXCPU:13’.

idem


Now, what should I do with this information ?

[toc] | [prev] | [next] | [standalone]

#230845

[Multipart message — attachments visible in raw view] — view raw

On Tue, Jan 19, 2021 at 03:45:41PM +0100, steve wrote:
> Le 19-01-2021, à 14:49:36 +0100, tomas@tuxteam.de a écrit :
> 
> >On Tue, Jan 19, 2021 at 07:41:13AM -0500, Greg Wooledge wrote:
> >>On Tue, Jan 19, 2021 at 10:03:12AM +0100, steve wrote:
> >>> Jan 19 09:47:52 box systemd-udevd[611]: sdg5: Failed to update device symlinks: Too many levels of symbolic links
> >>> Jan 19 09:47:52 box systemd-udevd[581]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
> >>
> >>Look for a symlink that points to itself, presumably in either /dev or
> >>/etc/udev.  If you can't find one, then look for a symlink that points
> >>to an earlier point in the path to itself, creating a loop.
> >
> >FWIW GNU find is said to detect circular links. Try
> >
> > find $YOUR_SUSPECT_ROOT -follow -printf ""
> 
> Ok, I found some circular links:
> 
> find /usr -follow -printf ""
> find: Boucle détectée dans le système de fichiers ; « ‘/usr/bin/X11’ » est dans la même boucle que ‘/usr/bin’.
> 
> ls -l /usr/bin/X11
> lrwxrwxrwx 1 root root 1  6 mai  2013 /usr/bin/X11 -> .
> 
> 
> find /sys -follow -printf ""
[...]

Not every one of those are "bad". The X11 one is normal. Many of those in
/sys and /dev, too.

It'd be interesting to know which one your startup is choking on.

Cheers
 - t

[toc] | [prev] | [next] | [standalone]

#230847

Le 19-01-2021, à 15:53:00 +0100, tomas@tuxteam.de a écrit :

>> find /usr -follow -printf ""
>> find: Boucle détectée dans le système de fichiers ; « ‘/usr/bin/X11’ » est dans la même boucle que ‘/usr/bin’.
>>
>> ls -l /usr/bin/X11
>> lrwxrwxrwx 1 root root 1  6 mai  2013 /usr/bin/X11 -> .
>>
>>
>> find /sys -follow -printf ""
>[...]
>
>Not every one of those are "bad". The X11 one is normal. Many of those in
>/sys and /dev, too.

How can I tell the difference ?

>
>It'd be interesting to know which one your startup is choking on.

What does it mean more precisely? The lines come from syslog and only
mention partitions:

Jan 19 09:09:33 box systemd-udevd[607]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:09:33 box systemd-udevd[611]: sdc1: Failed to update device symlinks: Too many levels of symbolic links
Jan 19 09:09:33 box systemd-udevd[572]: sdc6: Failed to update device symlinks: Too many levels of symbolic links

[toc] | [prev] | [next] | [standalone]

#230852

On Tue, Jan 19, 2021 at 04:03:14PM +0100, steve wrote:
> > It'd be interesting to know which one your startup is choking on.
> 
> What does it mean more precisely? The lines come from syslog and only
> mention partitions:
> 
> Jan 19 09:09:33 box systemd-udevd[607]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
> Jan 19 09:09:33 box systemd-udevd[611]: sdc1: Failed to update device symlinks: Too many levels of symbolic links
> Jan 19 09:09:33 box systemd-udevd[572]: sdc6: Failed to update device symlinks: Too many levels of symbolic links

Well, the key information is that the program doing the complaining
is systemd-udevd, so the problem is in one of the places that program
works on.  That's why I suggested looking in /etc/udev and /dev first.

If there are any internal pieces of systemd that it might be complaining
about, then I don't know what those would be.

[toc] | [prev] | [next] | [standalone]

#230858

On Tue 19 Jan 2021 at 11:15:05 (-0500), Greg Wooledge wrote:
> On Tue, Jan 19, 2021 at 04:03:14PM +0100, steve wrote:
> > > It'd be interesting to know which one your startup is choking on.
> > 
> > What does it mean more precisely? The lines come from syslog and only
> > mention partitions:
> > 
> > Jan 19 09:09:33 box systemd-udevd[607]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
> > Jan 19 09:09:33 box systemd-udevd[611]: sdc1: Failed to update device symlinks: Too many levels of symbolic links
> > Jan 19 09:09:33 box systemd-udevd[572]: sdc6: Failed to update device symlinks: Too many levels of symbolic links
> 
> Well, the key information is that the program doing the complaining
> is systemd-udevd, so the problem is in one of the places that program
> works on.  That's why I suggested looking in /etc/udev and /dev first.
> 
> If there are any internal pieces of systemd that it might be complaining
> about, then I don't know what those would be.

I think the OP is looking for a needle in a haystack. /proc and /sys
are full of perfectly correct loops of symlinks, and /dev/fd is a link
straight back into /proc. So I'd start searching at specific /dev/foo
where foo avoids fd.

I'd also check /dev/disk carefully for any partitions with duplicate
LABELs, UUIDs and suchlike. Have any changes been made in
/{etc,lib}/udev/rules.d/* ?

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#230870

Oups, spoke too quickly.

I decreased udev log level, rebooted and the messages came back in
syslog.

[toc] | [prev] | [next] | [standalone]

#230871 — [SOLVED] Re: Too many levels of symbolic links

Hi,

I deleted some symlinks found in my home, updated the system (bullseye)
this morning, increased udev log level to debug and rebooted the system. 

grep symlinks /var/log/syslog

didn't show any mention of previous problem.

Le 19-01-2021, à 11:01:24 -0600, David Wright a écrit :

>On Tue 19 Jan 2021 at 11:15:05 (-0500), Greg Wooledge wrote:
>> On Tue, Jan 19, 2021 at 04:03:14PM +0100, steve wrote:
>> > > It'd be interesting to know which one your startup is choking on.
>> >
>> > What does it mean more precisely? The lines come from syslog and only
>> > mention partitions:
>> >
>> > Jan 19 09:09:33 box systemd-udevd[607]: sdg6: Failed to update device symlinks: Too many levels of symbolic links
>> > Jan 19 09:09:33 box systemd-udevd[611]: sdc1: Failed to update device symlinks: Too many levels of symbolic links
>> > Jan 19 09:09:33 box systemd-udevd[572]: sdc6: Failed to update device symlinks: Too many levels of symbolic links
>>
>> Well, the key information is that the program doing the complaining
>> is systemd-udevd, so the problem is in one of the places that program
>> works on.  That's why I suggested looking in /etc/udev and /dev first.
>>
>> If there are any internal pieces of systemd that it might be complaining
>> about, then I don't know what those would be.
>
>I think the OP is looking for a needle in a haystack. /proc and /sys
>are full of perfectly correct loops of symlinks

Good info, thanks.

> and /dev/fd is a link
>straight back into /proc. So I'd start searching at specific /dev/foo
>where foo avoids fd.
>
>I'd also check /dev/disk carefully for any partitions with duplicate
>LABELs, UUIDs and suchlike. Have any changes been made in
>/{etc,lib}/udev/rules.d/* ?

Not in my memory…

Thanks for the help, I think the problem is now solved.

Have a nice day

Steve

[toc] | [prev] | [next] | [standalone]

#230873

Question. What does the following mean?

# find /dev -follow -printf ""
find: '/dev/fd/4': No such file or directory

[toc] | [prev] | [next] | [standalone]

#230874

[Multipart message — attachments visible in raw view] — view raw

On Wed, Jan 20, 2021 at 09:40:46AM +0100, steve wrote:
> Question. What does the following mean?
> 
> # find /dev -follow -printf ""
> find: '/dev/fd/4': No such file or directory

This is funny. At first I thought I could reproduce it by tricking `find'
into following a broken symlink, but with a "static" setup I couldn't.

Still, /dev/fd is bound to be highly dynamic: it's a link to /proc/self/fd,
i.e. a view on the current process's open file descriptors. It's quite
probable that `find' checks some directory entry, and at the time it tries
to do something with it, it has disappeared, leading to that error
message.

Cheers
 - t

[toc] | [prev] | [next] | [standalone]

#230875

Thanks Mike and Thomas for the answers.

Le 20-01-2021, à 10:15:09 +0100, tomas@tuxteam.de a écrit :

>On Wed, Jan 20, 2021 at 09:40:46AM +0100, steve wrote:
>> Question. What does the following mean?
>>
>> # find /dev -follow -printf ""
>> find: '/dev/fd/4': No such file or directory
>
>This is funny. At first I thought I could reproduce it by tricking `find'
>into following a broken symlink, but with a "static" setup I couldn't.
>
>Still, /dev/fd is bound to be highly dynamic: it's a link to /proc/self/fd,
>i.e. a view on the current process's open file descriptors. It's quite
>probable that `find' checks some directory entry, and at the time it tries
>to do something with it, it has disappeared, leading to that error
>message.

What is strange is that it is always /dev/fd/4 that is missing, even
after a reboot.

[toc] | [prev] | [next] | [standalone]

#230876

[Multipart message — attachments visible in raw view] — view raw

On Wed, Jan 20, 2021 at 10:49:08AM +0100, steve wrote:
> Thanks Mike and Thomas for the answers.
> 
> Le 20-01-2021, à 10:15:09 +0100, tomas@tuxteam.de a écrit :
> 
> >On Wed, Jan 20, 2021 at 09:40:46AM +0100, steve wrote:
> >>Question. What does the following mean?
> >>
> >># find /dev -follow -printf ""
> >>find: '/dev/fd/4': No such file or directory
> >
> >This is funny. At first I thought I could reproduce it by tricking `find'
> >into following a broken symlink, but with a "static" setup I couldn't.
> >
> >Still, /dev/fd is bound to be highly dynamic: it's a link to /proc/self/fd,
> >i.e. a view on the current process's open file descriptors. It's quite
> >probable that `find' checks some directory entry, and at the time it tries
> >to do something with it, it has disappeared, leading to that error
> >message.
> 
> What is strange is that it is always /dev/fd/4 that is missing, even
> after a reboot.

As I said -- this is the own process's set of open file descriptors.
This process in this case is the one running `find' itself. At the
point it is looking for that /proc/fd/4, it is bound to be in some
semi-deterministic state. It's dancing a strange ballet with itself.

I'm sure one could find out exactly what's going on given enough
patience :-)

Cheers
 - t

[toc] | [prev] | [next] | [standalone]

#230877

Le 20-01-2021, à 10:58:54 +0100, tomas@tuxteam.de a écrit :

>> >># find /dev -follow -printf ""
>> >>find: '/dev/fd/4': No such file or directory
>> >
>> >This is funny. At first I thought I could reproduce it by tricking `find'
>> >into following a broken symlink, but with a "static" setup I couldn't.
>> >
>> >Still, /dev/fd is bound to be highly dynamic: it's a link to /proc/self/fd,
>> >i.e. a view on the current process's open file descriptors. It's quite
>> >probable that `find' checks some directory entry, and at the time it tries
>> >to do something with it, it has disappeared, leading to that error
>> >message.
>>
>> What is strange is that it is always /dev/fd/4 that is missing, even
>> after a reboot.
>
>As I said -- this is the own process's set of open file descriptors.
>This process in this case is the one running `find' itself. At the
>point it is looking for that /proc/fd/4, it is bound to be in some
>semi-deterministic state. It's dancing a strange ballet with itself.

Ah ok, I misunderstood that. Very interesting indeed.

>I'm sure one could find out exactly what's going on given enough
>patience :-)

I' guess so.

My current is to find out why my system is spitting out these 'too many
symnolic links' messages and then fix it.

Thanks a lot for all your explanation !

Steve

[toc] | [prev] | [next] | [standalone]

#230879

Le 20-01-2021, à 11:20:25 +0100, tomas@tuxteam.de a écrit :


>The /proc file system is pretty interesting indeed. But I don't think
>that is at the root of your strange problem.

...

>> My current is to find out why my system is spitting out these 'too many
>> symnolic links' messages and then fix it.
>>
>> Thanks a lot for all your explanation !
>
>You're welcome -- although I fear I couldn't help you much with your
>primary issue.

Yes you did, you gave me the `find` command which helped me go a bit
further. 

[toc] | [prev] | [next] | [standalone]

#230880 — Bluewin bouncing my mails [was: Too many levels of symbolic links]

[Multipart message — attachments visible in raw view] — view raw

And oh, btw. Steve

your provider, bluewin.ch is bouncing my mails.

It seems it is using some "Cloudmark CSI Reputation Service" as anti-spam
measure. When I go to the form in their bounce message [1], they present
me with a form requesting my personal data.

To me, this looks like a scam by Cloudmark -- a way to suck up personal
data hidden behind an anti-spam "service".

Please, please: tell your provider to not support such a despicable action.

Thanks

[1] https://csi.cloudmark.com/en/reset/?ip=5.199.139.25

 - tomás

[toc] | [prev] | [next] | [standalone]

#230885 — Re: Bluewin bouncing my mails

[Multipart message — attachments visible in raw view] — view raw

On Wed, Jan 20, 2021 at 12:11:05PM +0100, steve wrote:
> Thomas,
> 
> I think we can keep this private since it is not Debian specific.

Sorry for CC'ing the list again, but for me, this is the only way
to reach you. Direct messages... bounce (I tried again).

> Thanks for letting me know of this problem.

Thanks for trying to look into it.

> Could you forward me the bounced message? Is there any clue in it
> explain why your message has been bounced?

I would do (actually, I did), but it bounces, as expected.

The reason is pretty clear (and I commend bluewin for making it
explicit -- other providers [hotmail, I'm looking at you!] just
mark the mail as spam and let their users cope with it). Quoting
from the bounce:

| dlist@bluewin.ch
|   SMTP error from remote mail server after end of data:
|   host mxbw.lb.bluewin.ch [195.186.227.50]: 554 5.2.0 sc962:
|   Blacklisting in effect - http://csi.cloudmark.com/reset-request/?ip=5.199.139.25

When I go to the URL, Cloudmark wants to collect some private
data from me to lift the blacklisting. I'm not amused ;-)

Cheers
 - tomás

[toc] | [prev] | [next] | [standalone]

#230881

[Multipart message — attachments visible in raw view] — view raw

On Wed, Jan 20, 2021 at 11:15:26AM +0100, steve wrote:
> Le 20-01-2021, à 10:58:54 +0100, tomas@tuxteam.de a écrit :
> 
> >>>># find /dev -follow -printf ""

[...]

> >This process in this case is the one running `find' itself. At the

[...]

> Ah ok, I misunderstood that. Very interesting indeed.
> 
> >I'm sure one could find out exactly what's going on given enough
> >patience :-)
> 
> I' guess so.

The /proc file system is pretty interesting indeed. But I don't think
that is at the root of your strange problem.

> My current is to find out why my system is spitting out these 'too many
> symnolic links' messages and then fix it.
> 
> Thanks a lot for all your explanation !

You're welcome -- although I fear I couldn't help you much with your
primary issue.

Cheers
 - t

[toc] | [prev] | [next] | [standalone]

#230882

Le 20/01/2021 à 10:49, steve a écrit :
> Thanks Mike and Thomas for the answers.
> 
> Le 20-01-2021, à 10:15:09 +0100, tomas@tuxteam.de a écrit :
> 
>> On Wed, Jan 20, 2021 at 09:40:46AM +0100, steve wrote:
>>> Question. What does the following mean?
>>>
>>> # find /dev -follow -printf ""
>>> find: '/dev/fd/4': No such file or directory
>>
>> This is funny. At first I thought I could reproduce it by tricking `find'
>> into following a broken symlink, but with a "static" setup I couldn't.
>>
>> Still, /dev/fd is bound to be highly dynamic: it's a link to 
>> /proc/self/fd,
>> i.e. a view on the current process's open file descriptors. It's quite
>> probable that `find' checks some directory entry, and at the time it 
>> tries
>> to do something with it, it has disappeared, leading to that error
>> message.
> 
> What is strange is that it is always /dev/fd/4 that is missing, even
> after a reboot.
> 
> 

it is not missing : it disapears. And /dev/fd/4 is a link to a file 
opend by find itself, it is sufficient that find closes it and the name 
does not exist anymore.

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | linux.debian.user

csiph-web