RHBA-2009:0105 - Bug Fix Advisory

Topic

Updated kexec-tools packages that fix several bugs and add various
enhancements are now available.

Description

kexec is a set of system calls which allow a kernel to be loaded from the
currently-executing Linux kernel. In addition to the kexec command, these
packages contain ancillary utilities which comprise the user-space component
of the kernel's kexec feature.

The following bugs have been fixed:

• the mkdumprd command incorrectly parsed comment lines in the

/etc/mdadm.conf configuration file.

• kdump's initial ramdisk, dumprd, was not being rebuilt after its

components were updated.

• kdump became unresponsive when a device entry listed in /etc/fstab did

not exist (such as a nonexistent swap directory).

A bug in the IDE/ATA driver stack that could prevent a system using
kernel-xen from booting into the kdump environment has been now fixed. In
previous releases, this occurred if the system encountered a kernel panic
while an IDE device was performing I/O and a device driver other than
libata controlled the device.

• kdump became unresponsive when parsing custom device nodes, such as

custom RAID software-created nodes.

• the Graphics Address Remapping Table (GART) region is now excluded from

/proc/vmcore, resolving potential problems such as unwanted reboots.

• the kdump initrd now waits for critical block devices.
• unnecessary warnings concerning the ld.conf file have been removed.
• VMCore files could not be saved to remote hosts when running the xend

daemon and network-bridge on a dom0 hypervisor.

• on the Itanium(r) architecture, when the kdump kernel was booted, this

message appeared in the boot log:

mknod: /tmp/initrd.[numbers omitted]/dev/efirtc: No such file or directory"

This message, which resulted from a malformed reuest to create the efirtc
in an incorrect path, is redundant, because this device path is created
statically in the initramfs when kdump is started, and as also harmless, in
that it should not affect kdump's performance.

• in certain circumstances, makedumpfile failed on the 64-bit PowerPC

architecture due to a memory region size miscalculation.

• the kexec-kdump-howto.txt file now contains information about the correct

usage with scp.

• makedumpfile failed on the Itanium(r) architecture due to an issue with

overlapping segments.

• some invalid statements in the mkdumprd man page have been clarified.
• kdump failed to locate LVM volumes and save the crash dump due to a file

name parsing error.

• special "md" and "cciss" devices are now excluded from

/etc/critical_disks when running mkdumprd.

• mkdumprd is now able to handle more complicated bonded interface

configurations.

• the kdump init script contained some now-corrected typos.
• mkdumprd was unable to drop the user to a shell prompt as a last resort

after failing to capture a VMCore.

• previously, RAID 5 systems resulted in mkdumprd issuing the following

(abbreviated) error message: "No module raid5 found...Failed to run mkdumprd".

• makedumpfile now handles a corner case when building filter maps.

These packages also feature these enhancements:

• the makedumpfile command can now be used on dom0 hypervisors without

requiring the use of the kernel debugging packages.

• dumprd now displays VMCore capture progress.
• kdump.conf can now override the /etc/modprobe.conf module options.
• a manual page for the kdump.conf configuration file has been added, which

include information about the kdump_pre and kdump_post directives.

• makedumpfile now supports hypervisor kernels.
• new udev rules have been added to restart the kdump service on CPU and

memory hot plug devices.

• kexec is now statically linked to the zlib library to enable faster

reboot times.

• "makedumpfile --help" now provides help information.
• kdump now benefits from improved support for certain HP systems when used

with the 5.3 kernel.

Users are advised to upgrade to these updated kexec-tools packages, which
resolve these issues and add these enhancements.

Solution

Before applying this update, make sure that all previously-released errata
relevant to your system have been applied.

This update is available via Red Hat Network. Details on how to use the Red
Hat Network to apply this update are available at
http://kbase.redhat.com/faq/docs/DOC-11259

Affected Products

• Red Hat Enterprise Linux Server 5 x86_64
• Red Hat Enterprise Linux Server 5 ia64
• Red Hat Enterprise Linux Server 5 i386
• Red Hat Enterprise Linux Workstation 5 x86_64
• Red Hat Enterprise Linux Workstation 5 i386
• Red Hat Enterprise Linux Desktop 5 x86_64
• Red Hat Enterprise Linux Desktop 5 i386
• Red Hat Enterprise Linux for IBM z Systems 5 s390x
• Red Hat Enterprise Linux for Power, big endian 5 ppc
• Red Hat Enterprise Linux Server from RHUI 5 x86_64
• Red Hat Enterprise Linux Server from RHUI 5 i386

Fixes

• BZ - 254163 - kdump under kernel-xen fails with I/O load on aic7xxx or megaraid_mbox
• BZ - 360111 - kdump not working with non-default LVM names
• BZ - 438056 - makedumpfile -c on an existing vmcore fails on ppc
• BZ - 438459 - [5.2][kdump] "ls: /etc/ld.so.conf.d/*: No such file or directory" warning
• BZ - 438755 - mkdumprd creates /dev/efirtc in wrong location in initrd
• BZ - 440983 - Description about ssh/scp in kexec-kdump-howto.txt needs to be updated
• BZ - 442678 - [5.2][kdump] unable to mount rootfs. Dropping to shell
• BZ - 444824 - not all mkdumprd options work
• BZ - 446279 - kdump initrd does not handle lun scanning race condition
• BZ - 449111 - [RHEL5.2] makedumpfile corrupts vmcore on ia64: crash's bt fails to unwind
• BZ - 450549 - makedumpfile core_collector comment in kdump.conf incorrect
• BZ - 456154 - mdadm usage bug in /sbin/mkdumprd may cause dumps to be lost
• BZ - 457758 - /sbin/mkdumprd parses comment lines of /etc/mdadm.conf
• BZ - 457904 - Confusion about Default Action
• BZ - 458702 - kdump initrd hangs waiting on cciss or md devices
• BZ - 459614 - kdump's dumprd not rebuilding when appropriate.
• BZ - 459622 - pre- and post- scripts silently ignored
• BZ - 459879 - kdump via bond device doesn't work for non-basic config.
• BZ - 461551 - initscripts unmounts /usr before calling kexec which needs it...
• BZ - 461616 - Typos in kdump inistscript
• BZ - 462829 - [5.3] Bad Message - not space enough
• BZ - 463878 - [5.3] mkdumprd Not Robust Enough
• BZ - 463915 - [5.3] SCP - dd: /dev/mem: Bad address
• BZ - 465361 - mkdumprd doesn't handle raid5 properly
• BZ - 469147 - [5.3] TPS Verify Test Failure
• BZ - 472575 - kdump is failed by nonexistent device entry in /etc/fstab
• BZ - 475505 - need to add code to exclude gart region from vmcore

CVEs

(none)

References

(none)