Description of problem: kdump service should show an appropriate message when it fails to start up due to unspecified crashkernel. Version-Release number of selected component (if applicable): kexec-tools-1.101-151.el5 How reproducible: Always Steps to Reproduce: 1. Load 1st kernel that crashkernel parameter is not specified. 2. Run "service kdump start" 3. Actual results: No appropriate message that lets users know what happens to the service is shown. Expected results: The following message should be shown. No crashkernel parameter is specified for the running kernel (<kernel version>) Additional info:
Created attachment 144014 [details] This patch is based on -151.el5 and, can fix this problem.
No, that patch won't work. For what its worth, I don't think there is anything particularly misleading or bad about the current error that kexec provides, but if you insist on checking for the presence of crashkernel on the commandline the regex needs to handle malformed crashkernel specifications that are missing a parameter, or those which use an alternate unit of measure to specify memory sizes (K or G rather than M, or the use of lower case letters, which the kernel allows). I'm checking a more accurate test into -152.el5
According to Comment #2, the following statement in load_kdump() should be modified, shouldn't it? KDUMP_COMMANDLINE=`echo $KDUMP_COMMANDLINE | sed -e 's/crashkernel=[0-9]\+M@[0-9]\+M//g'` I guess $MEM_RESERVED can be used there.
Fix confirmed with kexec-tools-1.101-164.el5.