"Couldn't find UUID" error only when schedule is "when drive get connected"

"when drive get connected" uses udev to start a new snapshot as soon as a physical drive (via USB or eSata) with the correct UUID get connected. A remote share doesn't have a UUID (and also will not be processed by udev), so you can't use this schedule here.
To get (nearly) the same result you could create a script which mounts your share and call 'backintime --profile-id 1 --backup-job' to take a new snapshot. And always use that script to mount.

I have a problem, though, since I'm getting this error message even if my target folder is on a USB drive.
The strange thing is that this worked when I first set up the profile (some months ago). Then, I had to make some adjustments to the profile configuration, but couldn't save it anymore because of this error message. I then changed the backup schedule in order to save the backup profile, but now I can't set it back to "when drive gets connected" anymore :-(
How can I fix this?

I solved in a somewhat weird way: I had to unmount the USB device (which was mounted as my current user) and re-mount it as root. Then, BIT did not complain any more (just asked for the root password and allowed to save changes). I don't know how it could work the first time, maybe I was just lucky that I had mounted it as root that time too... although it's strange.

If you have an idea on what was wrong and why it does not work if the device is mounted as a normal user, please let me know.
Just an additional info: this is a SATA hard drive connected through an USB docking station. It has a BTRFS filesystem on it, which cleanly mounts as my current user (it's done automatically as soon as I connect the device) but I must always unmount it as root, otherwise the system says it's busy... I don't know why, I haven't tried to fix this yet.

Please try this patch from http://pastebin.ubuntu.com/8402410/ with 'sudo patch /usr/share/backintime/common/tools.py < diff'

I'm not sure anymore why I used os.stat to compare UUID with device. But it maybe was a bad idea because stat also includes atime which will change during the drive is used. The patch will now compare the realpath (after resolving all symlinks).

If this doesn't help please run these commands in python and post the output:

$ python
Python 2.7.6 (default, Mar 22 2014, 22:59:56)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path.append('/usr/share/backintime/common')
>>> import tools
>>> tools.get_mountpoint('/your/path/to/backintime/') ##insert your path here
('/dev/sdb1', '/media/germar/usb')
>>> tools.get_uuid('/dev/sdb1') ##insert the device from above
'1234-5678'

Hi Germar,
unfortunately applying the patch did not help. The output of tools.get_uuid('/dev/sdb2') is empty.

Please post the output of 'ls -la /dev/disk/by-uuid/' and 'sudo blkid'

# ls -la /dev/disk/by-uuid
drwxr-xr-x 2 root root 160 set 23 09:03 .
drwxr-xr-x 7 root root 140 set 23 09:03 ..
lrwxrwxrwx 1 root root 10 set 21 21:58 5726-CEAE -> ../../sda1
lrwxrwxrwx 1 root root 10 set 21 21:58 d85382b5-a2d1-492f-bd8b-e76aef2b08ab -> ../../sda3
lrwxrwxrwx 1 root root 10 set 21 21:58 e2bba17d-6159-483a-93ad-e1284be04eb1 -> ../../dm-2
lrwxrwxrwx 1 root root 10 set 21 21:58 e8f9db74-d243-49b6-a8cc-ea2c80615054 -> ../../sda2
lrwxrwxrwx 1 root root 10 set 23 09:03 ef539b3f-e2f4-4f10-bc64-28a736fa59f7 -> ../../sdb1
lrwxrwxrwx 1 root root 10 set 21 21:58 f6041671-d4b0-45ab-abfe-29b10cf9bb06 -> ../../dm-1

# sudo blkid
/dev/sda1: UUID="5726-CEAE" TYPE="vfat"
/dev/sda2: UUID="e8f9db74-d243-49b6-a8cc-ea2c80615054" TYPE="ext2"
/dev/sda3: UUID="d85382b5-a2d1-492f-bd8b-e76aef2b08ab" TYPE="crypto_LUKS"
/dev/mapper/sda3_crypt: UUID="tjWDvi-QwCI-ODth-pmgL-Zh5P-dhBo-BMvwwl" TYPE="LVM2_member"
/dev/mapper/mint--vg-root: UUID="f6041671-d4b0-45ab-abfe-29b10cf9bb06" TYPE="ext4"
/dev/mapper/mint--vg-swap_1: UUID="e2bba17d-6159-483a-93ad-e1284be04eb1" TYPE="swap"
/dev/sdb1: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="f1b03d9a-95d7-4948-a7fe-d6534980bdea" TYPE="btrfs"
/dev/sdb2: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="91ed6ded-5bec-4528-ab4a-34f2153b29fa" TYPE="btrfs"

The backup folder is in /media/mauro/WDC-Green, which is a mount point for /dev/sdb2:
# mount
[...]
/dev/sdb2 on /media/mauro/WDC-Green type btrfs (rw,nosuid,nodev,uhelper=udisks2)

The mount point creation and device mounting is performed automatically by KDE (I think), as soon as I connect my device.

# ll /media/mauro/
totale 4
drwxrwxrwx 1 root root 34 ago 2 09:55 WDC-Green/

If you're wondering about the output of blkid, please note that /dev/sdb contains two partitions (/dev/sdb1 and /dev/sdb2, 50% of the disk each) with a btrfs volume on them configured as mirror for both data and metadata. Well, there's a reason for this apparently nonsense configuration ;-)

Another weird thing that happens with this device and BIT (which might be related) is the following: once I got to save the "when device is connected" preference, when I then connect my device BIT correctly starts the backup process. However, most of the times (not always) I get a KDE dialog saying BIT has crashed. However the backup process is actually still running: if I open the BIT UI immediately after confirming the crash dialog, BIT opens and correctly reports that a backup is in progress.

I suspect the problem here is that /dev/disk/by-uuid does not report a UUID for /dev/sdb2... this might be because of the internals of btrfs which can use more than one partition for a single filesystem volum e. The UUID_SUB might be of help in this case, so blkid seems to be more reliable than /dev/disk/by-uuid... I'm not an expert, though...

You're right. The missing symlink for ../../sdb2 is the problem in here. Normally udev should create that. Maybe this is a bug in udev in combination with btrfs. I'll make my self a btrfs drive to play around a bit and report back soon. Which distribution and kernel do you use?

Sure blkid would be more reliable. But you need to run blkid with sudo (at least the first time after system reboot). I tried to avoid that and rather used /dev/disk/by-uuid/

Are you sure the snapshot is still running after the crash report? I mean, BIT might show a status bar message false indicating there is a snapshot running. This is because ~/.local/share/backintime/worker.message wasn't removed correctly. This will be fixed in next major version 1.1.0

Can you please post the corresponding syslog part during such a crash?

Just for my own interest, what's this reason for your wired mirror config?

Since btrfs is still considered "experimental", the kernel version may actually make a difference here. I'm using Linux Mint 16 KDE, based on Ubuntu 13.10, and kernel version is 3.11.0-24-generic #41-Ubuntu.

You say that for blkid you need to ask for sudo password, but the current way of setting the "when drive gets connected" still prompts for the sudo password, doesn't it? At least this is what the UI says (and what I've experienced so far).

I'm sure the snapshot is running because the status bar of BIT shows the paths of files being processed, I hear the target hard disk seeking noise and the new snapshot gets created on the target file system. Next time it happens I will check the syslog. Meanwhile I tried to use the KDE bult-in bug report system to send a notification about the crash, but that procedure failed because (I think) debug information is missing.

Regarding my btrfs configuration: I know it is not exactly the most orthodox approach and that this can be risky to be done on the backup target drive, but it's something like an experiment used on a secondary backup store. The case is this: I have a 1 TB drive that has some bad clusters on it. It might fail one day, but until then it's a pity I have to waste around 1 TB of space just for a bunch of bad blocks. So, what I thought to do is this: let's sacrifice half the space, while writing all the data in two copies on that drive. Btrfs is perfect for this, also because it has sophisticated algorithms to detect and fix media errors. Now, the drive as a whole may still fail, but the chances to have a piece of information copied into two blocks that are both bad is (at the moment) quite remote. So now I have at least a 500 GB drive in "good" state. For non-mission critical things it's perfectly fine for me. My backups are also stored in a (slower) perfectly functioning and mirrored NAS device.

> You say that for blkid you need to ask for sudo password, but the current way of setting the
> "when drive gets connected" still prompts for the sudo password, doesn't it?

Yes and no. When you leave Settingsdialog BIT will always need the UUID for all profiles which have udev schedule. But it will only set up a new udev rule (and ask for your sudo password to do so) if the rule need to be changed. So with blkid you'd need to give sudo access even if you just added a new exclude.

Here is what gets printed to /var/log/syslog when I plug in my USB drive with btrfs on it and BIT crashes:

Sep 25 23:34:55 hppb anacron[6934]: Anacron 2.3 started on 2014-09-25
Sep 25 23:34:55 hppb anacron[6935]: Anacron 2.3 started on 2014-09-25
Sep 25 23:34:55 hppb anacron[6934]: Will run job `1_Profilo_principale' in 0 min.
Sep 25 23:34:55 hppb anacron[6934]: Jobs will be executed sequentially
Sep 25 23:34:55 hppb anacron[6934]: Job `1_Profilo_principale' started
Sep 25 23:34:55 hppb anacron[6935]: Job `1_Profilo_principale' locked by another anacron - skipping
Sep 25 23:34:55 hppb anacron[6935]: Normal exit (0 jobs run)
Sep 25 23:34:56 hppb anacron[6970]: Anacron 2.3 started on 2014-09-25
Sep 25 23:34:56 hppb anacron[6971]: Anacron 2.3 started on 2014-09-25
Sep 25 23:34:56 hppb anacron[6971]: Will run job `1_Profilo_principale' in 0 min.
Sep 25 23:34:56 hppb anacron[6971]: Jobs will be executed sequentially
Sep 25 23:34:56 hppb anacron[6971]: Job `1_Profilo_principale' started
Sep 25 23:34:56 hppb anacron[6970]: Job `1_Profilo_principale' locked by another anacron - skipping
Sep 25 23:34:56 hppb anacron[6970]: Normal exit (0 jobs run)
Sep 25 23:34:56 hppb kernel: [ 7945.813684] device label WDC-Green devid 2 transid 366 /dev/sdb2
Sep 25 23:34:56 hppb kernel: [ 7945.815682] btrfs: disk space caching is enabled
Sep 25 23:34:56 hppb backintime (root): INFO: Lock
Sep 25 23:34:56 hppb backintime (mauro): INFO: Lock
Sep 25 23:34:56 hppb backintime (root): INFO: on process begins
Sep 25 23:34:56 hppb backintime (root): INFO: Profile_id: 1
Sep 25 23:34:56 hppb backintime (mauro): INFO: on process begins
Sep 25 23:34:56 hppb udisksd[2419]: Mounted /dev/sdb2 at /media/mauro/WDC-Green on behalf of uid 1000
Sep 25 23:34:56 hppb backintime (mauro): INFO: Profile_id: 1
Sep 25 23:34:56 hppb backintime (mauro): INFO: Call rsync to take the snapshot
Sep 25 23:34:57 hppb backintime (root): INFO: Call rsync to take the snapshot
Sep 25 23:35:02 hppb backintime (root): INFO: Command "rsync -rtDH --links -pEgo --delete --delete-excluded -v --chmod=Du+wx --exclude="/media/mauro/WDC-Green/Backup" --exclude="/root/.lo$
Sep 25 23:35:02 hppb backintime (root): INFO: Save config file
Sep 25 23:35:02 hppb backintime (root): WARNING: Command "cp /root/.config/backintime/config /media/mauro/WDC-Green/Backup/backintime/hppb/root/Configurazione sistema/new_snapshot/backup/.." $
Sep 25 23:35:02 hppb backintime (root): INFO: Create info file
Sep 25 23:35:02 hppb backintime (root): INFO: Remove backups older than: 20120901-000000
Sep 25 23:35:02 hppb backintime (root): INFO: Keep min free disk space: 10240 Mb
Sep 25 23:35:02 hppb backintime (root): INFO: Keep min 2% free inodes
Sep 25 23:35:04 hppb backintime (root): INFO: Unlock
Sep 25 23:35:04 hppb anacron[6971]: Job `1_Profilo_principale' terminated
Sep 25 23:35:04 hppb anacron[6971]: Normal exit (1 job run)

Please note I have two jobs (one for root and one for mauro) that should start on device plug. I don't see anything evident here.

The backtrace of the crash is this:
Application: Back In Time (kde4systrayicon.py), signal: Aborted
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[KCrash Handler]
#6 0x00007f5f4e0d4f77 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#7 0x00007f5f4e0d85e8 in __GI_abort () at abort.c:90
#8 0x00007f5f49b61912 in qt_message_output (msgType=<optimized out>, buf=0x322fca8 "backintime(7046)/kdeui (kdelibs): Session bus not found \nTo circumvent this problem try the following command (with Linux and bash) \nexport $(dbus-launch) ") at global/qglobal.cpp:2347
#9 0x00007f5f4a822a89 in QDebug::~QDebug (this=0x7fff6c6fb2d0, __in_chrg=<optimized out>) at /usr/include/qt4/QtCore/qdebug.h:85
#10 0x00007f5f4a8fab48 in KApplicationPrivate::init (this=0x311aac0, GUIenabled=GUIenabled@entry=true) at ../../kdeui/kernel/kapplication.cpp:516
#11 0x00007f5f4a8fb532 in KApplication::KApplication (this=0x2f04680, GUIenabled=true) at ../../kdeui/kernel/kapplication.cpp:352
#12 0x00007f5f4b65343d in sipKApplication::sipKApplication (this=0x2f04680, a0=<optimized out>) at sip/kdeui/sipkdeuipart7.cpp:29314
#13 0x00007f5f4b653932 in init_KApplication (sipSelf=0x2f3fe90, sipArgs=<optimized out>, sipKwds=0x0, sipUnused=0x7fff6c6fb500, sipParseErr=0x7fff6c6fb518) at sip/kdeui/sipkdeuipart7.cpp:30239
#14 0x00007f5f446bf87a in sipSimpleWrapper_init (self=self@entry=0x2f3fe90, args=args@entry=0x7f5f4e83e050, kwds=kwds@entry=0x0) at /build/buildd/sip4-4.15.2/siplib/siplib.c:9481
#15 0x00000000005336db in type_call.25488 (type=<optimized out>, args=0x7f5f4e83e050, kwds=0x0) at ../Objects/typeobject.c:745
#16 0x0000000000561669 in PyObject_Call (kw=0x0, arg=0x7f5f4e83e050, func=0x2cc8100) at ../Objects/abstract.c:2529
#17 do_call (nk=<optimized out>, na=<optimized out>, pp_stack=0x7fff6c6fb6c0, func=0x2cc8100) at ../Python/ceval.c:4239
#18 call_function (oparg=<optimized out>, pp_stack=0x7fff6c6fb6c0) at ../Python/ceval.c:4044
#19 PyEval_EvalFrameEx (f=f@entry=0x29f6d90, throwflag=throwflag@entry=0) at ../Python/ceval.c:2666
#20 0x0000000000556cea in PyEval_EvalCodeEx (closure=<optimized out>, defcount=<optimized out>, defs=0x0, kwcount=<optimized out>, kws=<optimized out>, argcount=44002704, args=<optimized out>, locals=0x0, globals=<optimized out>, co=<optimized out>) at ../Python/ceval.c:3253
#21 function_call (func=func@entry=0x2f9d050, arg=arg@entry=0x7f5f4e771610, kw=kw@entry=0x0) at ../Objects/funcobject.c:526
#22 0x00000000004b5164 in PyObject_Call (kw=0x0, arg=0x7f5f4e771610, func=0x2f9d050) at ../Objects/abstract.c:2529
#23 instancemethod_call.8798 (func=0x2f9d050, func@entry=0x7f5f4c7801e0, arg=0x7f5f4e771610, arg@entry=0x7f5f4e83e050, kw=kw@entry=0x0) at ../Objects/classobject.c:2602
#24 0x00000000004d997b in PyObject_Call (kw=0x0, arg=0x7f5f4e83e050, func=0x7f5f4c7801e0) at ../Objects/abstract.c:2529
#25 PyEval_CallObjectWithKeywords (func=func@entry=0x7f5f4c7801e0, arg=arg@entry=0x7f5f4e83e050, kw=kw@entry=0x0) at ../Python/ceval.c:3890
#26 0x00000000004d485d in PyInstance_New (klass=<optimized out>, arg=0x7f5f4e83e050, kw=0x0) at ../Objects/classobject.c:581
#27 0x0000000000561669 in PyObject_Call (kw=0x0, arg=0x7f5f4e83e050, func=0x2f9a120) at ../Objects/abstract.c:2529
#28 do_call (nk=<optimized out>, na=<optimized out>, pp_stack=0x7fff6c6fbc30, func=0x2f9a120) at ../Python/ceval.c:4239
#29 call_function (oparg=<optimized out>, pp_stack=0x7fff6c6fbc30) at ../Python/ceval.c:4044
#30 PyEval_EvalFrameEx (f=f@entry=0x28b5cb0, throwflag=throwflag@entry=0) at ../Python/ceval.c:2666
#31 0x000000000054b7d4 in PyEval_EvalCodeEx (co=0x7f5f4e765db0, globals=<optimized out>, locals=<optimized out>, args=<optimized out>, argcount=<optimized out>, kws=<optimized out>, kwcount=0, defs=0x0, defcount=0, closure=0x0) at ../Python/ceval.c:3253
#32 0x000000000055830c in PyEval_EvalCode (locals=0x2867690, globals=0x2867690, co=0x7f5f4e765db0) at ../Python/ceval.c:667
#33 run_mod.42568 (mod=mod@entry=0x2908340, filename=filename@entry=0x7fff6c6fcd7f "/usr/share/backintime/kde4/kde4systrayicon.py", globals=globals@entry=0x2867690, locals=locals@entry=0x2867690, flags=flags@entry=0x7fff6c6fbf20, arena=arena@entry=0x28935c0) at ../Python/pythonrun.c:1365
#34 0x0000000000468567 in PyRun_FileExFlags (fp=fp@entry=0x28e5fa0, filename=filename@entry=0x7fff6c6fcd7f "/usr/share/backintime/kde4/kde4systrayicon.py", start=start@entry=257, globals=globals@entry=0x2867690, locals=locals@entry=0x2867690, closeit=closeit@entry=1, flags=flags@entry=0x7fff6c6fbf20) at ../Python/pythonrun.c:1351
#35 0x0000000000468aa0 in PyRun_SimpleFileExFlags (fp=fp@entry=0x28e5fa0, filename=<optimized out>, filename@entry=0x7fff6c6fcd7f "/usr/share/backintime/kde4/kde4systrayicon.py", closeit=closeit@entry=1, flags=flags@entry=0x7fff6c6fbf20) at ../Python/pythonrun.c:943
#36 0x0000000000468b9c in PyRun_AnyFileExFlags (fp=fp@entry=0x28e5fa0, filename=filename@entry=0x7fff6c6fcd7f "/usr/share/backintime/kde4/kde4systrayicon.py", closeit=closeit@entry=1, flags=flags@entry=0x7fff6c6fbf20) at ../Python/pythonrun.c:747
#37 0x000000000046a1e3 in Py_Main (argc=<optimized out>, argv=0x7fff6c6fc0d8) at ../Modules/main.c:640
#38 0x00007f5f4e0bfde5 in __libc_start_main (main=0x46a2a1 <main>, argc=3, ubp_av=0x7fff6c6fc0d8, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fff6c6fc0c8) at libc-start.c:260
#39 0x00000000005735fe in _start ()

It seems like a crash in the tray icon component, rather than in the engine, doesn't it?

This sounds like bug #1332126. If you're on BIT version < 1.0.36 please try the patch from that bug report (or even better, update to 1.0.36).

I'm wondering what this 'WARNING: Command "cp ...config ...' is about. Looks like the line is cut by 'less' at the most interesting point. Could you please post the full line.

Unfortunately I'm still on Linux Mint 16 (based on Saucy) and the PPA for this does not have 1.0.36. I'll try to patch and let you know.

Regarding the cut row:

WARNING: Command "cp /root/.config/backintime/config /media/mauro/WDC-Green/Backup/backintime/hppb/root/Configurazione sistema/new_snapshot/backup/.." returns 256

Ubuntu 14.04 will create UUID symlinks just fine for btrfs. So this is a bug in 13.10 which is outdated by now. Please upgrade to Mint 17

Hi Germar,
I upgraded to Linux Mint 17 some weeks ago, but unfortunately the behaviour is the same. Look at this:

uname -a:
Linux hppb 3.13.0-36-generic #63-Ubuntu SMP Wed Sep 3 21:30:07 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

mount | grep WDC
/dev/sdb2 on /media/mauro/WDC-Green type btrfs (rw,nosuid,nodev,uhelper=udisks2)

sudo blkid | grep WDC
/dev/sdb1: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="f1b03d9a-95d7-4948-a7fe-d6534980bdea" TYPE="btrfs"
/dev/sdb2: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="91ed6ded-5bec-4528-ab4a-34f2153b29fa" TYPE="btrfs"

ll /dev/disk/by-uuid/
[...]
lrwxrwxrwx 1 root root 10 ott 13 22:27 ef539b3f-e2f4-4f10-bc64-28a736fa59f7 -> ../../sdb1

As you can see, UUID for /dev/sdb2 is missing. With some experiments, I concluded that it depends on which of the two partitions the system decides to try to mount first (I have automatic mounting enabled) when the disk is plugged: if it decides to mount /dev/sdb1 all works fine (because there's the UUID in /dev/disk/by-uuid), if it decides to mount /dev/sdb2 the problem happens. After all, there can't be more than one UUID with the same name in /dev/disk/by-uuid, so it's natural that this can happen and I don't think that it will ever change.
I think the discriminant here is that I'm using a BTRFS file system that spans multiple volumes. I don't know if the fact that both volumes are on the same physical disk is important or not in this scenario (but if the UUID is unique per file system, it shouldn't matter where volumes are located).

Also please note that I'm still experiencing the crash after plugging in the disk (I'm using BIT 1.0.38 now). Here is the backtrace:
Application: Back In Time (kde4systrayicon.py), signal: Aborted
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[KCrash Handler]
#5 0x00007fde6ad85bb9 in raise () from /lib/x86_64-linux-gnu/libc.so.6
#6 0x00007fde6ad88fc8 in abort () from /lib/x86_64-linux-gnu/libc.so.6
#7 0x00007fde64ae0c92 in qt_message_output(QtMsgType, char const*) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#8 0x00007fde657a5be9 in ?? () from /usr/lib/libkdeui.so.5
#9 0x00007fde6587ddf8 in ?? () from /usr/lib/libkdeui.so.5
#10 0x00007fde6587e7d2 in KApplication::KApplication(bool) () from /usr/lib/libkdeui.so.5
#11 0x00007fde665d8f6d in sipKApplication::sipKApplication(bool) () from /usr/lib/python2.7/dist-packages/PyKDE4/kdeui.so
#12 0x00007fde665d9442 in ?? () from /usr/lib/python2.7/dist-packages/PyKDE4/kdeui.so
#13 0x00007fde5ffd1b4c in ?? () from /usr/lib/python2.7/dist-packages/sip.so
#14 0x00000000004f5d0b in ?? ()
#15 0x000000000052cc20 in PyEval_EvalFrameEx ()
#16 0x000000000056d0aa in ?? ()
#17 0x00000000004d9854 in ?? ()
#18 0x00000000004da20b in PyEval_CallObjectWithKeywords ()
#19 0x0000000000497c7d in PyInstance_New ()
#20 0x000000000052cc20 in PyEval_EvalFrameEx ()
#21 0x000000000055c594 in PyEval_EvalCodeEx ()
#22 0x00000000005b7392 in PyEval_EvalCode ()
#23 0x0000000000469663 in ?? ()
#24 0x00000000004699e3 in PyRun_FileExFlags ()
#25 0x0000000000469f1c in PyRun_SimpleFileExFlags ()
#26 0x000000000046ab81 in Py_Main ()
#27 0x00007fde6ad70ec5 in __libc_start_main () from /lib/x86_64-linux-gnu/libc.so.6
#28 0x000000000057497e in _start ()

Okay, I made my self a Mint17 KDE VM and a USB Thumbdrive with btrfs raid1 to play around.

I found out that 'df -P <path>' will always return /dev/sdb1 for device of <path>. But /dev/disk/by-uuid/ sometimes links /dev/sdb1 and sometimes /dev/sdb2.

I'm not sure how to deal with this. But I'll try to find a way.

I wasn't able to reproduce the kdesystrayicon crash yet. You could remove /usr/share/backintime/plugins/kdeplugin.py so it won't crash anymore. But you'd loose the systray icon, too.

Could you please run these command once in 'python' and in 'sudo -H python' to see if they crash:

import os, sys, subprocess
cmd = [ sys.executable, '/usr/share/backintime/kde4/kde4systrayicon.py', '1' ]
p = subprocess.Popen( cmd, env = {'HOME': os.environ['HOME']} )
p.terminate()

I found an easy solution for the uuid prob (should have seen this before...)
Please use this patch with 'sudo patch /usr/share/backintime/common/tools.py < tools.diff'

http://pastebin.ubuntu.com/8567374/

Here are the result of executing those python commands:
- as normal user:
>>> import os, sys, subprocess
>>> cmd = [ sys.executable, '/usr/share/backintime/kde4/kde4systrayicon.py', '1' ]
>>> p = subprocess.Popen( cmd, env = {'HOME': os.environ['HOME']} )
>>> Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 14: reading configurations from ~/.fonts.conf is deprecated. please move it to /home/mauro/.config/fontconfig/fonts.conf manually
INFO: [kde4systrayicon] begin loop
p.terminate()

- with sudo:
>>> import os, sys, subprocess
>>> cmd = [ sys.executable, '/usr/share/backintime/kde4/kde4systrayicon.py', '1' ]
>>> p = subprocess.Popen( cmd, env = {'HOME': os.environ['HOME']} )
>>> No protocol specified
kde4systrayicon.py: cannot connect to X server :0.0
>>> p.terminate()

However, couldn't the crash problem be related to the UUID problem? Something like this: although BIT starts when I plug in the device (whatever is the partition that gets mounted), maybe the tray icon crashes for some circumstances when the wrong partition gets mounted?

I have now installed the debug symbols that were available... maybe now this backtrace is more meaningful?
Here it is: http://pastebin.ubuntu.com/8575852/

Regarding the UUID thing: I do not verify what you say about df -P; here is my current situation:

mount | grep WDC
/dev/sdb2 on /media/mauro/WDC-Green type btrfs (rw,nosuid,nodev,uhelper=udisks2)

sudo blkid | grep WDC
/dev/sdb1: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="f1b03d9a-95d7-4948-a7fe-d6534980bdea" TYPE="btrfs"
/dev/sdb2: LABEL="WDC-Green" UUID="ef539b3f-e2f4-4f10-bc64-28a736fa59f7" UUID_SUB="91ed6ded-5bec-4528-ab4a-34f2153b29fa" TYPE="btrfs"

ll /dev/disk/by-uuid/
[...]
lrwxrwxrwx 1 root root 10 ott 13 22:27 ef539b3f-e2f4-4f10-bc64-28a736fa59f7 -> ../../sdb1

df -P /media/mauro/WDC-Green
File system 1024-blocchi Usati Disponib. Capacità Montato su
/dev/sdb2 976760832 337725728 638289880 35% /media/mauro/WDC-Green

So, df -P shows /dev/sdb2, coherent with mount.

This said, after applying your patch, I can finally change the scheduling settings successfully, even when the mounted partition is /dev/sdb2 instead of /dev/sdb1! Well done! :-)

Do you prefer me to open a new question/bug for the tray icon issue?

Yes. Please create a new bug report for this. I'm already lost in this thread.

systrayicon crash has nothing to do with uuid. Did it crash during those tests in #18?

I opened bug #1382384, thank you for your help!