[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#418823: marked as done (linux-image-2.6.18-3-k7: fails to shutdown (some) software raid arrays on userspace suspend)

Your message dated Thu, 3 Jul 2008 16:10:46 +0200
with message-id <20080703141046.GE13910@stro.at>
and subject line Re: linux-image-2.6.18-3-k7: fails to shutdown (some) software raid arrays on userspace suspend
has caused the Debian Bug report #418823,
regarding linux-image-2.6.18-3-k7: fails to shutdown (some) software raid arrays on userspace suspend
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)


-- 
418823: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=418823
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message --- 
Package: linux-image-2.6.18-3-k7
Version: 2.6.18-7
Severity: normal

[Initial report, to uswsusp, with input from uswsusp maintainer, mdadm
maintainer, and some others on #415441. I'm not sure whether to merge
this report with that, or keep them as separate issues with one
blocking the other, as whatever is resolved here might have an impact
on uswsusp's initrd scripts. Advice would be appreciated.]

When using software suspend on this kernel, (some) raid arrays are not
shutdown before the machine is. Thus, on the next startup, when the
initrd detects them (which happens before attempting to resume from
suspend) it schedules a reconstruction. This, in turn, prevents resume
from working, as the reconstructing process cannot be frozen. (Also,
it was said on the original report that this reconstruction can cause
data loss).

Now, I have worked around this in two ways:

1. By having the resume script in the initrd wait for the
reconstruction. Workable if I first set the array to have a write
intent bitmap (otherwise the reconstruction, and thus the resume,
takes ~40 minutes). But it sounds wrong to do this, as I assume the
resume image to expect the array to be unsync'd, since it was saved in
such state.

2. By moving the resume to happen *before* raid discovery in the
initrd's scripts. Which takes care of above's issue, but is not a
workable solution for the uswsusp or mdadm maintainers, since the
resume image itself might be in an array. Or in an LVM volume which is
itself in an array, ... FWIW, in this system this workaround has been
in use for about 15 days, without apparent ill effects.

Why do I think the kernel ought to shutdown the array before final
system shutdown? Because a home compiled vanilla 2.6.17, with kernel
space suspend, and built-in (not modules) sw raid, does so. In that
kernel, without an initrd, suspending by
 /bin/echo shutdown > /sys/power/disk; /bin/echo disk > /sys/power/state
then restarting finds and assembles raid arrays before resuming
without triggering resyncs.

Just to be complete, my set up is as follows:

Two identical ide disks, hda, hde.
Four partitions on each. 
hd{a,e}3 combined into a raid1 array used as / (md1).
hd{a,e}4 combined into a raid1 array used as pv for lvm with all other
 filesystems (md2).
hda1 as /boot, hde1 unused.
hde2 as swap.
hda2 as low-priority swap, and used to hold suspend image.

On this setup, the without my workarounds behaviour is that md1
(appears to be) propperly shutdown. Next bootup simply finds it
and assembles it. But md2 is not shutdown and triggers a resync on
assembly, thus propting all this bug report.

I'm using uswsusp 0.6~cvs20070202-1 with compressed, encrypted images,
altough that seems to be irrelevant, since the resume image is always
properly found and (if the array rebuild is avoided) loaded.

-- System Information:
Debian Release: lenny/sid
  APT prefers testing
  APT policy: (500, 'testing'), (400, 'unstable'), (50, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.18-3-k7 (SMP w/1 CPU core)
Locale: LANG=es_MX, LC_CTYPE=es_MX (charmap=ISO-8859-1)
Shell: /bin/sh linked to /bin/bash

Versions of packages linux-image-2.6.18-3-k7 depends on:
ii  coreutils                     5.97-5.3   The GNU core utilities
ii  debconf [debconf-2.0]         1.5.13     Debian configuration management sy
ii  initramfs-tools [linux-initra 0.85g      tools for generating an initramfs
ii  module-init-tools             3.3-pre4-2 tools for managing Linux kernel mo

Versions of packages linux-image-2.6.18-3-k7 recommends:
ii  libc6-i686                  2.3.6.ds1-13 GNU C Library: Shared libraries [i

-- debconf information:
  linux-image-2.6.18-3-k7/postinst/old-initrd-link-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/postinst/kimage-is-a-directory:
  linux-image-2.6.18-3-k7/postinst/bootloader-error-2.6.18-3-k7:
  linux-image-2.6.18-3-k7/preinst/elilo-initrd-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/postinst/old-dir-initrd-link-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/preinst/lilo-has-ramdisk:
  linux-image-2.6.18-3-k7/preinst/abort-install-2.6.18-3-k7:
* linux-image-2.6.18-3-k7/preinst/lilo-initrd-2.6.18-3-k7: false
  linux-image-2.6.18-3-k7/preinst/abort-overwrite-2.6.18-3-k7:
  linux-image-2.6.18-3-k7/preinst/failed-to-move-modules-2.6.18-3-k7:
  linux-image-2.6.18-3-k7/postinst/create-kimage-link-2.6.18-3-k7: true
  shared/kernel-image/really-run-bootloader: true
  linux-image-2.6.18-3-k7/postinst/depmod-error-initrd-2.6.18-3-k7: false
  linux-image-2.6.18-3-k7/postinst/old-system-map-link-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/postinst/depmod-error-2.6.18-3-k7: false
  linux-image-2.6.18-3-k7/prerm/removing-running-kernel-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/preinst/overwriting-modules-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/preinst/bootloader-initrd-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/preinst/already-running-this-2.6.18-3-k7:
  linux-image-2.6.18-3-k7/prerm/would-invalidate-boot-loader-2.6.18-3-k7: true
  linux-image-2.6.18-3-k7/preinst/initrd-2.6.18-3-k7:
  linux-image-2.6.18-3-k7/postinst/bootloader-test-error-2.6.18-3-k7:

Attachment: signature.asc
Description: Digital signature


--- End Message ---
--- Begin Message --- 
Version: 2.6.25-1

closing as no follow up to initial report, no second report.
old kernel version.

thanks for the report.

-- 
maks

--- End Message ---
Reply to: