=Comment: #0================================================= SRIPATHI KODI <sripathik.com> - Crash shipped with RHEL5.4 (crash 4.0-8.9.1.el5) cannot open MRG1.2 vmcores. I see this error while trying to open a kdump produced on 2.6.31.4-rt14.27.el5rt kernel: crash: invalid structure size: x8664_pda FILE: x86_64.c LINE: 584 FUNCTION: x86_64_cpu_pda_init() [/usr/bin/crash] error trace: 44a0ff => 4ced4d => 4d0607 => 5098a5 5098a5: SIZE_verify+168 4d0607: (undetermined) 4ced4d: x86_64_init+3205 44a0ff: main_loop+152 =Comment: #1================================================= SRIPATHI KODI <sripathik.com> - I tried the latest version of crash from here: http://people.redhat.com/anderson/crash-4.1.0.tar.gz and it is able to open the vmcore. Hence we need a newer version of crash to be shipped with MRG1.2.
crash package needs to be updated to deal with MRG 1.2 (2.6.31+) kernels
crash version 4.1.2-1.el5 (ON_QA for RHEL5.5 crash utility errata) is available: /mnt/redhat/brewroot/packages/crash/4.1.2/1.el5
------- Comment From sripathik.com 2009-12-17 00:57 EDT------- Since MRG1.2 shipped with 2.6.24 based kernels, the crash shipped with RHEL5.4 is perfectly capable of handling the core files. However, this will be needed when 2.6.31 or later real-time kernel ships with MRG1.3. We can verify this once MRG1.3 kernels are available.
------- Comment From dvhltc.com 2010-03-29 16:50 EDT------- Suggest we close as a dup of Bug 57159?
(In reply to comment #4) > ------- Comment From dvhltc.com 2010-03-29 16:50 EDT------- > Suggest we close as a dup of Bug 57159? Huh?
That's his IBM BZ instance number. (and yeah, I went there first and was confused too. What the heck is 'relust' anyway?) the RH bz he's referring to is: https://bugzilla.redhat.com/show_bug.cgi?id=530325
(In reply to comment #6) > That's his IBM BZ instance number. > > (and yeah, I went there first and was confused too. What the heck is 'relust' > anyway?) > > the RH bz he's referring to is: > > https://bugzilla.redhat.com/show_bug.cgi?id=530325 Ok, then yes, by all means this BZ can be closed as a dup of 530325.
------- Comment From dvhltc.com 2010-03-30 11:48 EDT------- Re-opening to mark as a duplicate of LTC BZ 57159 (RH 530325). The solution to each is a new version of crash. ------- Comment From dvhltc.com 2010-03-30 11:49 EDT------- *** This bug has been marked as a duplicate of bug 57159 ***
------- Comment From amitarora.com 2010-03-30 12:57 EDT------- (In reply to comment #14) > Re-opening to mark as a duplicate of LTC BZ 57159 (RH 530325). The solution to > each is a new version of crash. Found following two bugs which are related here - LTC Bug#55402 (Redhat Bug#517952) : This bug is for requesting an upgrade to latest version of crash tool to be able to open vmcore from newer kernels. LTC Bug#57159 (Redhat Bug#530325) : This bug talks about not able to dump the vmcore. This is now fixed on .33-rt kernel, since we are able to get the vmcore now. Thus this bug should be marked as fixed. Looking at the above two bugs, this one (bug#57442/Redhat Bug#532024) should be marked duplicate of LTC Bug#55402 (Redhat Bug#517952) instead, since thats the one which requests an upgrade to latest crash in MRG1.3. Doing the same. Thanks! ------- Comment From amitarora.com 2010-03-30 12:57 EDT------- *** This bug has been marked as a duplicate of bug 55402 ***
*** This bug has been marked as a duplicate of bug 517952 ***