Description of problem: The kdump service needs to initially check on whether available size is greater than $memtotal. The service should warn admins in the case that there is not space enough to save a vmcore. According to mkdumprd, this issue remains TODO. Version-Release number of selected component (if applicable): kexec-tools-1.101-158.el5 How reproducible: Always Steps to Reproduce: 1. Configure kdump.conf so that there is not space enough to save a vmcore. 2. Run "service kdump restart" 3. Actual results: The kdump service shows no warning as below. Stopping kdump: [ OK ] Detected /etc/kdump.conf or /boot/efi/efi/redhat/vmlinuz-2.6.18-1.2910.el5 change Rebuilding /boot/efi/efi/redhat/initrd-2.6.18-1.2910.el5kdump.img Starting kdump: [ OK ] Expected results: The kdump service shows warning as below. Stopping kdump: [ OK ] Detected /etc/kdump.conf or /boot/efi/efi/redhat/vmlinuz-2.6.18-1.2910.el5 change Rebuilding /boot/efi/efi/redhat/initrd-2.6.18-1.2910.el5kdump.img Warning: There is not space enough to save a vmcore. The size of my.server.com:/export/tmp should be much greater than 15883760 bytes. Starting kdump: [ OK ] Additional info:
Created attachment 144924 [details] This patch is based on -158.el5 and, can fix this problem.
looks good, but you skipped the ssh case.
Created attachment 144931 [details] This patch is based on -158.el5 and, can fix this problem. I modified the previous patch.
modified patch checked into -160.el5. Thanks.
Note: you're modified patch doesn't work for the ssh case, the cut has to be made at field 11, not 4. The patch I checked in corrects that.
On my server, -160.el5 deosn't work for the ssh case. The mkdumprd shows the following error message. /sbin/mkdumprd: line 1599: [: 64%: integer expression expected Once I change feild to 10, it works for the ssh case. At that time, I found the following matter in the ssh case. It is inappropriate that ':' is used for the tr because a colon is used to separate user name and hostname. So, I think '|' is better than ':' for the tr. The following patch can fix this problem. Index: mkdumprd =================================================================== RCS file: /cvs/dist/rpms/kexec-tools/RHEL-5/mkdumprd,v retrieving revision 1.50 diff -p -u -r1.50 mkdumprd --- mkdumprd 5 Jan 2007 21:11:18 -0000 1.50 +++ mkdumprd 5 Jan 2007 23:47:39 -0000 @@ -1593,7 +1593,7 @@ if [ -n "$KDUMP_CONFIG_FILE" ]; then exit 1 fi remote_df=`ssh -q $s_opts $rlocation df $tdir` - available_size=$(echo $remote_df | tail -1 | tr -s ' '':' | cut -d: -f11) + available_size=$(echo $remote_df | tr -s ' ' '|' | cut-d\| -f10) #check for available size is greater than $memtotal if [ $available_size -lt $memtotal ]; then
Fix confirmed with kexec-tools-1.101-164.el5.