Short story: /usr/bin/mlocate has changed permissions without reason.
Sending this here in case anyone has got an idea or so that if others
are seeing the same they could find this in google.
I've got a home server/desktop with an ITX board with Atom D510 that
ran perfectly stable since I bought the board a couple months ago. A
couple weeks ago I put a newly purchased SSD in it and reinstalled
Squeeze from scratch. I've barely installed any software yet (and I
deselected the desktop task, so it doesn't even have X on it), and
have just used it as a file server for some other attached disks
Today I've noticed that a system file has changed:
chris@tn:~$ locate Foo
locate: can not open `/var/lib/mlocate/mlocate.db': Permission denied
tn:~# find -uid 130 /
find: `/proc/1892/task/1892/fd/5': No such file or directory
find: `/proc/1892/task/1892/fdinfo/5': No such file or directory
find: `/proc/1892/fd/5': No such file or directory
find: `/proc/1892/fdinfo/5': No such file or directory
tn:~# find -gid 130 /
find: `/proc/2002/task/2002/fd/5': No such file or directory
find: `/proc/2002/task/2002/fdinfo/5': No such file or directory
find: `/proc/2002/fd/5': No such file or directory
find: `/proc/2002/fdinfo/5': No such file or directory
meaning it's the only file everywhere with that gid or even uid (of
course ignore the warnings, they are because of find finding its own
proc entries), and the group doesn't exist.
(Last change to the group file:
-rw-r--r-- 1 root root 605 2011-08-25 00:17 /etc/group
I think I've installed the system on 2011-08-18.)
I've rebooted the machine but it didn't change, so I've just run chgrp
(and chmod g+s) on the file. But I'm really wondering how this has
happened. The /usr filesystem is unencrypted and on the SSD, so it
could be some bit(s) flipped on the SSD, or in RAM and somehow written
out (seems unlikely).
104 is b01101000
130 is b10000010
So that's *5* changed bits in the gid. Maybe the way SSDs work it
could be the effect of a single bit change in the underlying flash?
But I would have expected that I would get a read error instead in
(The SSD is a Kingston SSDNow V100 Series 64GB 2.5IN SSD SATA. In
/proc/scsi/scsi it is shown as "Model: KINGSTON SV100S2 Rev: D110".)
I haven't found any other binary with an odd group. I can't think of
any way I would have caused it (I've also checked my bash history); in
case there are others who have had the same issue, please speak up
since then it's probably a bug somewhere.
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact email@example.com
Archive: CAEjYwfUKRiWrfLqCB=Cjc7fM=OvgcRuUPq=E8hmnFsokvb0CW Q@mail.gmail.com">http://lists.debian.org/CAEjYwfUKRiWrfLqCB=Cjc7fM=OvgcRuUPq=E8hmnFsokvb0CW Q@mail.gmail.com