LMDE UP 6 regression: /etc/fstab cifs no longer works!

[on4aa]

Ok, I got this regression confirmed on two wired desktops and two wifi-connected laptops that I upgraded this week from LMDE 64 UP5 to UP6.

Previously, on all four machines my server drive mounted automatically over CIFS at boot with the following line in /etc/fstab :

Code: Select all

//servername/home               /media/mountname   cifs   defaults,credentials=/home/username/.smbcredentials,iocharset=utf8,file_mode=0640,dir_mode=0640,uid=username,gid=users   0   0

After upgrading from UP5 to UP6, the drive no longer mounts at boot. I can mount it manually though with:

Code: Select all

sudo mount -a

This feels like the network coming up too slowly at boot. However, I am not sure how to debug this one, nor which logs to check.
Thank you in advance for your help! Merry Christmas, peace & love and the whole thing...

[Monsta]

The closest bug report I could find is this one: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=674039
This looks quite complicated though. Note that the bug's title refers to a different behavior. It seems to have eventually changed to something like what you've experienced.

[zerozero]

The closest bug report I could find

and here http://bugs.debian.org/cgi-bin/bugrepor ... g=674039#6 they mention bug #673936 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=673936 also mentioned here >> viewtopic.php?f=198&t=112981 (in the last post, ginjabunny reports that the fix is working fine)

[altair4]

I don't use LMDE but if you're looking for some workarounds:

[1] Back in the olden tymes you were able to add another option to the list that would prevent the boot process from executing that specific line in fstab until the network was up: _netdev

It doesn't seem to work like it used to but it might for you.

[2] Another option is to formalize the " 'mount -a' after booting" process by adding a file to /etc/network/if-up.d:

Code: Select all

#!/bin/sh
mount -a


Save the file to ... /etc/network/if-up.d/remount.
Make it executable: sudo chmod +x /etc/network/if-up.d/remount

It will do a "mount -a" only after the network is up.

Side question: How on earth did that mount statement work with a dir_mode argument of 640 and not 750? Are you connecting to another Linux box?

[on4aa]

@altair4:

[1] I first tried following the documentation by adding _netdev as an option to my cifs line in /etc/fstab
As you correctly predicted, it no longer works.

[2] Then I added your remount script to /etc/network/if-up.d/
SUCCESS! Many thanks!

Anyhow, it is a bit sad that Debian Linux got -at least temporarily- complicated to that level.
Imagine being a complete n00b and trying to mount a cifs drive for the first time...

[Side matter] You do have sharp eyes! I indeed never noticed that dir_mode=0640 does not make a lot of sense for newly created directories on my remote server.
I changed it to dir_mode=0750

[Monsta]

Anyhow, it is a bit sad that Debian Linux got -at least temporarily- complicated to that level.

Well, this is Debian Testing, after all.

[iLobster]

In my system mount cifs (smbfs) resources at boot with fstab doesn't work for a quite long time. Same commands could be done manually when system finally boot without any question.
So there is a definite problem with fstab, yes.

[on4aa]

@Montsa
Thanks for reminding me about Debian Testing. I tend to forget this.

[nspboarderdude]

I had the same issue as on4aa after UP6. The second suggestion from altair4 worked for me as well. Feels like a dirty hack, but hey it works.

[on4aa]

If you are logging in and out different users during the same session, an occasional hiccup may still happen.

A more reliable location to place

Code: Select all

mount -a

is in:
/etc/mdm/PostLogin/Default
/etc/gdm3/PostLogin/Default
/etc/kdm/PostLogin/Default
/etc/lightdm/PostLogin/Default
/etc/xdm/PostLogin/Default
depending on the display manager in use.

Be sure to rename the Default.sample file to Default for this to work.

[kcpoole]

I am having the same issue as mentioned and Rather than a workaround, is it going to be fixed at any time?

Ken