mdadm segfaults in lxd container

Bug #1691763 reported by Brian Candler on 2017-05-18
This bug affects 3 people
Affects Status Importance Assigned to Milestone
mdadm (Ubuntu)

Bug Description

Inside a 16.04 lxd container (running in default unprivileged mode), inside a 16.04 host:

# /etc/cron.daily/mdadm
Segmentation fault (core dumped)

# strace -f /sbin/mdadm --monitor --scan --oneshot
brk(0x1ba1000) = 0x1ba1000
mknod("/dev/.tmp.md1", S_IFBLK|0600, makedev(9, 1)) = -1 EPERM (Operation not permitted)
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0} ---
+++ killed by SIGSEGV (core dumped) +++
Segmentation fault (core dumped)

This is troublesome because (a) mdadm is installed by default, (b) mdadm runs under a daily cron job, (c) cron sends out errors via E-mail. So you get spammed from each container.

As others have observed[^1], the workaround is to remove mdadm from inside lxd containers.


ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: mdadm 3.3-2ubuntu7.2
ProcVersionSignature: Ubuntu 4.8.0-52.55~16.04.1-generic 4.8.17
Uname: Linux 4.8.0-52-generic x86_64
ApportVersion: 2.20.1-0ubuntu2.5
Architecture: amd64
Date: Thu May 18 13:48:22 2017
 Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
 Bus 001 Device 002: ID 413c:a001 Dell Computer Corp. Hub
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Dell Inc. PowerEdge R230
 PATH=(custom, no user)
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-4.8.0-52-generic root=/dev/mapper/ix--mon2-root ro console=tty1 console=ttyS0,115200
SourcePackage: mdadm
UpgradeStatus: No upgrade log present (probably fresh install) 01/12/2017
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 2.0.8 0DWX9P
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 23
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr2.0.8:bd01/12/2017:svnDellInc.:pnPowerEdgeR230:pvr:rvnDellInc.:rn0DWX9P:rvrA00:cvnDellInc.:ct23:cvr: PowerEdge R230
dmi.sys.vendor: Dell Inc. Error: [Errno 2] No such file or directory: '/etc/'
initrd.files: Error: [Errno 2] No such file or directory: '/boot/initrd.img-4.8.0-52-generic'

Brian Candler (b-candler) wrote :
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in mdadm (Ubuntu):
status: New → Confirmed
lordaro (charlespigott) wrote :

I just encountered this myself. Was a bit bored, so did some testing.

mdadm master HEAD (4.0) does not segfault, nor does the HEAD of the branch mdadm-3.3.4

However, if you acquire the actual version of source used by ubuntu (3.3, from 2013(!)) and compile, and run through gdb, you get:

Program received signal SIGSEGV, Segmentation fault.
strlen () at ../sysdeps/x86_64/strlen.S:106
106 ../sysdeps/x86_64/strlen.S: No such file or directory.
(gdb) bt
#0 strlen () at ../sysdeps/x86_64/strlen.S:106
#1 0x00007ffff7a9847e in __GI___strdup (s=0x0) at strdup.c:41
#2 0x0000000000477ddf in xstrdup (str=0x0) at xmalloc.c:66
#3 0x0000000000435009 in add_new_arrays (mdstat=0x6b4480, statelist=0x7fffffffe020, test=0, info=0x7fffffffe070)
    at Monitor.c:670
#4 0x00000000004336e7 in Monitor (devlist=0x0, mailaddr=0x6b4510 "root", alert_cmd=0x0, c=0x7fffffffe210,
    daemonise=0, oneshot=1, dosyslog=0, pidfile=0x0, increments=20, share=0) at Monitor.c:223
#5 0x0000000000406cd1 in main (argc=4, argv=0x7fffffffe6d8) at mdadm.c:1448

Monitor.c:670 is

st->devname = xstrdup(get_md_name(mse->devnm));

with the obvious unchecked return value of get_md_name

Looking through the git history, this was fixed with;a=commitdiff;h=1e08717f0b7856b389e9d5eb2dc330d146636183

I'd recommend backporting this patch :)

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers