Previously, when a kernel dump is configured with a local path in Red Hat Enterprise Virtualization Hypervisor 6.4, the dump file failed to be generated. This occured due to the increase in memory requirements of the kdump crash kernel causing out of memory errors when the crash kernel is started. This patch increases the assigned memory size to prevent the out of memory errors. The kernel dump file should correctly generate.
Created attachment 659276[details]
log file
Description of problem:
there is no dump file in /data/core when configured saving dump file to local through TUI
Version-Release number of selected component (if applicable):
rhev-hypervisor6-6.4-20121126.0.el6
How reproducible:
90%
Steps to Reproduce:
1. Install rhev-hypervisor6-6.4-20121126.0.el6.
2. Configure the Kernel Dump with Restore(Local) in the TUI.
3. echo c>/proc/sysrq-trigger to make the system coredump.
4. check the /data/core folder
Actual results:
There is no dump file in the /data/core folder since I have tried 5 times.
Expected results:
There is dump file in /data/core folder
Additional info:
1. there is no problem when configure with ssh/nfs.
2. I have tested rhev-hypervisor6-6.3-20121121.0.el6_3 build, and it can generate the local dump file, so it's a regression from 6.3 to 6.4
http://gerrit.ovirt.org/#/c/10151/
This patch increases the memory for crashkernels (now at 128M)
Seems as if the kdump crashkernel requires more mem now ...
Tested as follows:
1. Boot RHEV-H and ensure that the cmdline contains crashkernel=128M
2. Configure kdump with "local"
3. Drop to shell and run echo c > /proc/sysr-trigger
4. Crashkernel boots fine and crashdump appears in /data/core
Test version:
rhev-hypervisor6-6.4-20121212.1.auto1332.el6.
The cmdline contains "crashkernel=128M"
kdump with local can successful, crashkernel boots fine and crashdump appears in /data/core.
Test version:
rhev-hypervisor6-6.4-20130109.1.el6
The cmdline contains "crashkernel=128M"
kdump with local can successful, crashkernel boots fine and crashdump appears in /data/core.
So the bug is fixed, change bug status to VERIFIED.
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-2013-0556.html
Created attachment 659276 [details] log file Description of problem: there is no dump file in /data/core when configured saving dump file to local through TUI Version-Release number of selected component (if applicable): rhev-hypervisor6-6.4-20121126.0.el6 How reproducible: 90% Steps to Reproduce: 1. Install rhev-hypervisor6-6.4-20121126.0.el6. 2. Configure the Kernel Dump with Restore(Local) in the TUI. 3. echo c>/proc/sysrq-trigger to make the system coredump. 4. check the /data/core folder Actual results: There is no dump file in the /data/core folder since I have tried 5 times. Expected results: There is dump file in /data/core folder Additional info: 1. there is no problem when configure with ssh/nfs. 2. I have tested rhev-hypervisor6-6.3-20121121.0.el6_3 build, and it can generate the local dump file, so it's a regression from 6.3 to 6.4