Hide Forgot
If an x86_64 compressed kdump generated with "makedumpfile -c" contains more than 454 cpus, a crash session using RHEL6 crash-5.1.1-2.el6 will fails immediately with these two error messages: crash: compressed kdump: invalid nr_cpus value: <cpus> crash: vmcore: not a supported file format This is a vestige of the old diskdump dumpfile format upon which the compressed kdump format is based, where the dumpfile header in diskdump dumpfiles imposed such a limit, but which does not exist in compressed kdump dumpfile headers. The upstream crash utility version 5.1.5 has this fix, which restricts the error to only diskdump dumpfiles: http://people.redhat.com/anderson/crash.changelog.html - Fix for the handling of x86_64 compressed kdump dumpfiles where the crashing system contained more than 454 cpus. Without the patch, the crash session fails during initialization with the error message "crash: compressed kdump: invalid nr_cpus value: <cpus>" followed by "crash: vmcore: not a supported file format". (tindoh, tachibana.nec.co.jp) This problem is only applicable to compressed kdump dumpfiles, and is not a problem with ELF-style kdump dumpfiles. So the only way to work around this problem with a crash utility version earler than crash-5.1.5 is to avoid compressing the kernel. This can be done by *not* specifying the "makedumpfile -c" argument to the "core_collector" directive in /etc/kdump.conf. However, it is still acceptable to use "makedumpfile -d<option>" to filter out pages, because the output will still be an ELF kdump dumpfile.
FYI -- this issue was originally discovered and reported by Cliff Wickman of SGI (cpw) in this crash-utility mailing list thread: [Crash-utility] x86_64 limit of 454 cpu's? https://www.redhat.com/archives/crash-utility/2011-April/msg00020.html
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Previously, compressed kdump dump files were handled incorrectly on AMD64 and Intel 64 architectures if a system contained more than 454 CPUs. In such a case, the crash session terminated during initialization with the "crash: compressed kdump: invalid nr_cpus value: [cpus]" error message. A patch has been provided to address this issue, and the compressed dump files are now handled properly, thus fixing this bug.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2011-1648.html