Core dumps are generated when a node panics; panics can be caused by hardware or software failures. When a core dump is generated, support and engineering personnel can use it to determine the problem that caused the panic. Use the system coredump commands to manage core dumps.
The system coredump config modify command enables you to modify the following core-dump configuration settings. The core-dump configuration is cluster wide; that is, every node in a cluster uses the same core-dump configuration.
Whether sparse cores (that is, cores that omit all buffers that contain user data) are enabled. The default is true (that is, sparse cores are enabled).
The minimum number of bytes that the core dump leaves available on the root-volume file system. If the minimum number of bytes cannot be guaranteed, core dumps are not generated. The default is 250 MB.
The maximum number of times the core-dump process attempts to generate a core dump when encountering repeated disk failures. The default is 2.
The maximum number of times the core-dump process attempts to save a core dump. The default is 2.
Whether core dumps are automatically saved when a node restarts. The default is true.
A URL to which core dumps are uploaded. The default is ftp://ftp.netapp.com/to-ntap/.
The following example disables sparse cores and sets the upload location to ftp://ftp.example.com/intsup/netapp_cores:
node::>system coredump config modify -sparsecore-enabled false -upload-location ftp://ftp.example.com/intsup/netapp_cores
To display a cluster's core-dump configuration, use the system coredump config show command, as shown in the following example:
node::>system coredump config showSparse Min Max Max Autosave Core Free Dump Save On Startup Node Enabled Bytes Attempts Attempts Enabled Coredump Location ------------------------------------------------------------------------------ node14 false 262144000 2 2 true ftp://ftp.example.com/intsup/ netapp_cores node15 false 262144000 2 2 true ftp://ftp.example.com/intsup/ netapp_cores
The system coredump show command displays the following information about core dumps:
Node name
Core dump name
Whether the core is saved
The time the panic that generated the core occurred
Whether the core is a partial core
Number of attempts to save the core
The following example displays information about the core dumps on all nodes in a cluster:
node::>system coredump showNode Core Name Saved Panic Time ---------------------------------------------------------------------------- node01 core.101170979.2006-01-09.16_26_29.nz true 1/9/2006 11:26:29 core.101170979.2006-01-09.17_00_44.nz true 1/9/2006 12:00:44 node02 core.101170953.2006-01-06.03_25_10.nz true 1/5/2006 22:25:10
A core dump must be saved before it can be uploaded. To save a specific core dump, use the system coredump save command and specify the name of the core dump and the node on which the core dump is located. The following example saves a core dump named core.101180243.2006-02-17.12_26_34.nz on a node named node14:
node::>system coredump save -node node14 -corename core.101180243.2006-02-17.12_26_34.nz
To upload a saved core dump to a remote site for support analysis, use the system coredump upload command. When you run the command, specify the node on which the core dump is located and the name of the core dump. You can optionally specify the URL to which the core dump is to be uploaded; if a URL is not specified, the command uploads the core to the location specified by the -upload-location parameter of the system coredump config modify command.
The following example uploads a core dump named core.101170979.2006-01-09.17_00_44.nz from a node named node01 to the URL ftp://ftp.netapp.com/to-ntap/:
node::>system coredump upload -node node01 -corename core.101170979.2006-01-09.17_00_44.nz -location ftp://ftp.netapp.com/to-ntap/
The system coredump status command displays the following information about core dumps:
Node name
State
Space available on the internal file system
Number of unsaved complete cores
Number of unsaved partial cores
Core-dump type
If a core is in the process of being saved, the command also displays the name of the core being saved, the total number of blocks that are to be saved, and the current number of blocks saved.
You can specify additional parameters to display other information. See the reference page for the command for details.
The following example displays core-dump status information for a two-node cluster:
node::>system coredump statusSpace Number of Unsaved Number of Unsaved Coredump Node State Available Complete Cores Partial Cores Type ------------------------------------------------------------------------------ node14 nocore - 0 0 spray node15 nocore - 0 0 spray
After you have uploaded a core dump for support analysis, you can delete it to reclaim space on its node's root-volume file system. Use the system coredump delete command to delete a specific core dump. The following example deletes a core dump named core.101170979.2006-01-09.17_00_44.nz from a node named node01:
node::>system coredump delete -node node01 -corename core.101170979.2006-01-09.17_00_44.nz