Bug#523763: grub-common: floating point exception in grub-probe

[Felix Koop]

Neither / nor /boot are on raid. I have the folowing layout:

sda1, sdb1, sdc1 are on md1, raid 1, not mounted
sda2, sdb2, sdc2 are on md0, raid 5, which form lvm vg jupiter in this
vg is /var, /tmp and other data partitions
sdd1 is /boot
sdd2 is lvm vg jupisdd, in this vg is /root, /usr

Here is the output of your command:

jupiter:~# mdadm -Q --detail /dev/md0
/dev/md0:
        Version : 0.90
  Creation Time : Mon Mar  9 20:18:00 2009
     Raid Level : raid5
     Array Size : 2929773824 (2794.05 GiB 3000.09 GB)
  Used Dev Size : 1464886912 (1397.03 GiB 1500.04 GB)
   Raid Devices : 3
  Total Devices : 3
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Sun Apr 12 15:50:13 2009
          State : active
 Active Devices : 3
Working Devices : 3
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 64K

           UUID : 6831619a:697a4dc0:e7d7836c:33eff9d4
         Events : 0.32803

    Number   Major   Minor   RaidDevice State
       0       8       18        0      active sync   /dev/sdb2
       1       8        2        1      active sync   /dev/sda2
       2       8       34        2      active sync   /dev/sdc2
jupiter:~# mdadm -Q --detail /dev/md1
/dev/md1:
        Version : 0.90
  Creation Time : Mon Mar  9 20:43:02 2009
     Raid Level : raid1
     Array Size : 248896 (243.10 MiB 254.87 MB)
  Used Dev Size : 248896 (243.10 MiB 254.87 MB)
   Raid Devices : 3
  Total Devices : 3
Preferred Minor : 1
    Persistence : Superblock is persistent

    Update Time : Sat Mar 21 16:23:35 2009
          State : clean
 Active Devices : 3
Working Devices : 3
 Failed Devices : 0
  Spare Devices : 0

           UUID : ba013d18:eda56b8c:ca0395a4:0b8d4cce
         Events : 0.204

    Number   Major   Minor   RaidDevice State
       0       8       17        0      active sync   /dev/sdb1
       1       8        1        1      active sync   /dev/sda1
       2       8       33        2      active sync   /dev/sdc1

jupiter:~# pvdisplay
  --- Physical volume ---
  PV Name               /dev/sdd2
  VG Name               jupisdd
  PV Size               74,29 GB / not usable 1,42 MB
  Allocatable           yes 
  PE Size (KByte)       4096
  Total PE              19018
  Free PE               16714
  Allocated PE          2304
  PV UUID               aX2k3W-7jlq-PomM-Prn5-3qIp-y1Z6-nGJx7I
   
  --- Physical volume ---
  PV Name               /dev/md0
  VG Name               jupiter
  PV Size               2,73 TB / not usable 3,25 MB
  Allocatable           yes 
  PE Size (KByte)       4096
  Total PE              715276
  Free PE               529586
  Allocated PE          185690
  PV UUID               6cmzij-UpfT-JmEt-IQqx-wIb5-kBoR-tYqFGf
   



Am Sonntag, den 12.04.2009, 14:50 +0200 schrieb Felix Zielcke:
> Am Sonntag, den 12.04.2009, 14:27 +0200 schrieb Felix Koop:
> > Here is the backtrace:
> > 
> > (gdb) file grub-probe
> > Reading symbols from /usr/src/grub2-1.96
> > +20090317/build/grub-common/grub-probe...done.
> > (gdb) run --target=device /
> > Starting program: /usr/src/grub2-1.96
> > +20090317/build/grub-common/grub-probe --target=device /
> > /dev/mapper/jupisdd-root
> > 
> > Program received signal SIGFPE, Arithmetic exception.
> > 0x0000000000406cf4 in grub_divmod64 (n=11784322, d=0, r=0x7fff240cb6bc)
> >     at /usr/src/grub2-1.96+20090317/kern/misc.c:609
> > 609		*r = ((grub_uint32_t) n) % d;
> > (gdb) backtrace
> > #0  0x0000000000406cf4 in grub_divmod64 (n=11784322, d=0,
> > r=0x7fff240cb6bc)
> >     at /usr/src/grub2-1.96+20090317/kern/misc.c:609
> > #1  0x0000000000420f05 in grub_raid_read (disk=0x241a0f0,
> > sector=1508393216, 
> >     size=8, buf=0x24d3eb0 "") at /usr/src/grub2-1.96
> > +20090317/disk/raid.c:228
> 
> Weird. That would mean array->total_devs is 0
> Beetween lenny and sid/squeeze version it hasn't changed that it's taken
> from the mdraid superblock. But the RAID code itself has.
> How does your RAID arrays look like? (`mdadm -Q --detail /dev/md0' etc.)
> Maybe there's somewhere a mdraid superblock lying around which you
> should `mdadm --zero-superblock'