RHBA-2010:0179 - Bug Fix Advisory

Topic

An updated kexec-tools package that fixes numerous bugs is now
available.

Description

kexec-tools provides kexec binary that facilitates a new kernel to
boot using the kernel's kexec feature either on a normal or a panic
reboot, together with ancillary utilities that form the userspace
component of the kernel's kexec feature

This update addresses the following issues:

• previously, the kdump kernel command line supported a "mem=" parameter

that limited the memory that was dumped. When this parameter was set, the
dump would result in an I/O error. The "mem=" parameter has been removed
from kexec to ensure that core dumps succeed. (BZ#239791)

• previously kdump waited indefinitely for all devices in its

critical_disks list to be available before it performed a dump.
Kexec-tools now has a disk_timeout parameter that limits how long kdump
will wait for storage to respond. (BZ#500741)

• a logical flaw meant that the presence of files with certain names in

current directory of mkdumprd would prevent a dump. The code used to
evaluate the remote server name has been corrected. (BZ#509404)

• host names specified in the kdump.conf script needs to be entered

as fully-qualified domain names to allow for DNS changes. The
documentation for kdump has been revised to make this requirement clear.
(BZ#510816)

• previously, the vmcore file created by starting kdump with an initscript

used the "--sparse=always" option when copying, resulting in a smaller
file. The same option has now been added as a default value in the
kdump.conf configuration file, ensuring that the default behavior is
consistent. (BZ#511003)

• previously, faulty logic in the code that cleans up files used by kdump

in /tmp meant that files were sometimes left behind in /tmp. This has been
corrected to ensure that files in /tmp are cleaned up. (BZ#512098)

• previously, the order in which kdump loaded storage drivers meant that

USB-attached storage was sometimes not correctly detected. The USB driver
is now loaded later in the boot sequence, so that device enumeration is
correct and that dumps takes place successfully. (BZ#513608)

• Makedumpfiles can now accept kdump compressed dumps, therefore

allowing users to transfer smaller files. (BZ#516877)

• previously, the code used by kdump to find network interfaces could not

correctly identify slaves in a Xen environment when the dom0 was
configured for bonding. The code has now been updated so that it
recognizes this type of interface. (BZ#516907)

• Previously, the sample grub.conf file provided in the

kexec-kdump-howto.txt omitted the "crashkernel" parameter. The
sample file now describes a correctly configured grub.conf. (BZ#531244)

• Kexec now supports Enhanced Disk Drive Services (EDD) and up to 128

memory ranges from BIOS, so dumps on recent Intel 64-bit platforms now
complete successfully. (BZ#531340)

• previously, kdump did not test to see whether an NFS location was

writeable before commencing a dump. If the location was unwritable, kdump
would therefore start the dump anyway, which would inevitably fail. Kdump
now checks that NFS locations are writeable. (BZ#533565)

• previously, mkdumprd resolved hostnames for NFS locations specified in

kdump.conf and stored their IP addresses. Mkdumprd now stores the
hostnames instead and can therefore find the hosts successfully even if
their IP addresses change. (BZ#545980)

• kexec-tools now pulls in nsslibs and using the settings in

/etc/resolv.conf, can therefore perform DNS lookups and find NFS locations
specified in kdump.conf. (BZ#549946)

• a reference to libc.so.6 has now been removed for Itanium systems. This

avoids a potentially confusing warning message. (BZ#559126)

• a recent reorganization of the library directories on PowerPC systems

placed glibc in /lib/power6x, although libnss files expected it to be
in /lib. The correct path is now explicitly provided in mkdumprd, so DNS
lookups work on PowerPC systems. (BZ#569119)

All kexec-tools users should upgrade to this updated package, which
resolves these issues.

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 - 239791 - kdump incomplete if you set "mem=" parameter on x86_64
• BZ - 500741 - [Stratus 5.5 bug] "critical_disks" makes kdump unreliable
• BZ - 510816 - kdump fails to send vmcore via ssh/scp when configured using short hostname
• BZ - 511003 - mkdumprd copy vmcore without --sparse=always
• BZ - 513608 - RHEL5 Update4 i386: kdump got stuck on Dell2950
• BZ - 516877 - FEAT RHEL5.5: [makedumpfile] Support kdump-compressed format for input file(*)
• BZ - 516907 - kdump fails on dom0 when using bonding
• BZ - 531244 - fix for grub example in kexec-kdump-howto.txt
• BZ - 533565 - Read-only nfs mount point is configured, kdump service can still be started successfully.
• BZ - 549946 - cannot dump vmcore through NFS: invalid argument
• BZ - 559126 - [mkdumprd] cp: cannot stat `//lib/libc.so.6': No such file or directory
• BZ - 569119 - cannot dump vmcore through NFS on ppc system

CVEs

(none)

References

(none)