[Pkg-sysvinit-devel] Bug#431966: Shutdown & Reboot scripts try umount CIFS but CIFSD is killed first

[Renato Yamane]

Petter Reinholdtsen wrote:
> The best theory I find in any of them is that the network is pulled
> before the umounting is happening.  What kind of network setup are you
> using? Wireless or cable? Is there some VPN involved?

Tested in wireless *and* cable network.
The problem happening in *both* situations on AMD64 arch.

The problem *don't* happening on i386 arch (even on wireless or cable network).


> scripts linked in rc0.d and rc7.d are all called with 'stop' argument - i
> didn't know that.

I don't see rc7.d in my machine.

$ cd /etc/rc[TAB][TAB]
rc0.d/
rc1.d/
rc2.d/
rc3.d/
rc4.d/
rc5.d/
rc6.d/
rc.local
rcS.d/


> Now i'm thinking the problem was that wpa-ifupdown was being called
> before umountnfs. Renaming S15wpa-ifupdown to S33wpa-ifupdown also
> fixes the problem.

No. Renaming don't fix this problem.


> The update-rc.d call should be changed like this if this ordering
> make more sense:
>    update-rc.d wpa-ifupdown start 32 0 6 . >/dev/null || exit $?

I did this command above, but don't fix the problem.

Can someone send us the .config diff between linux-kernel i686 and amd64?

Maybe a net option is doing this problem, because I have 2 laptops (one is
i686 and other is amd64) and I see this problem *only* in amd64 arch
(wireless and cable network):

CIFS VFS: No response for cmd 50 mid 9487

Regards,
Renato S. Yamane
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: untitled-1
URL: <http://lists.alioth.debian.org/pipermail/pkg-sysvinit-devel/attachments/20090824/3cfdaeab/attachment-0001.asc>