Additional Setting For Kdump Configuration With 1-Tb Main Memory - Hitachi Compute Blade 2000 User Manual

Additional setting for Kdump
configuration with 1-TB main memory
You need to execute the following procedure for Kdump configuration with 1-TB
main memory before getting Kdump. If not, getting Kdump may fail.
Target OS: Red Hat Enterprise Linux 5
1. You need to add a setting to the /etc/sysconfig/kdump file. Execute the following
command to output kernel boot option for obtaining information.
# cat /proc/cmdline
Example for output:
ro root=/dev/vgroot/ hashdist=1 mem= mem=1024G
2. Add the kernel boot option without mem= mem=1024G output in Step 1 to
KDUMP_COMMANDLINE parameter to edit /etc/sysconfig/kdump,
Example for edited KDUMP_COMMANDLINE parameter
KDUMP_COMMANDLINE="ro root=/dev/vgroot/ hashdist=1"
3. Restart the system for /etc/sysconfig/kdump to take effect.
For the system with many I/O adapters, getting Kdump

may fail even if this setting is added. If so, change
memory size for Kdump kernel from 128 MB by default
to 256 MB, 512 MB... and retry it.
If the kernel boot option has been changed after a

character string is added to KDUMP_COMMANDLINE,
make sure to execute the procedure above.
CPU planned degradation with 4-blade
SMP (X57A1/X57A2 models)
When a system is configured with 4-blade SMP using high-
performance server blades and many I/O boards are
installed in the system, using CPU planned degradation
may cause an OS to fail to boot.
1780