clffdc command

Purpose

Collects snap data from every node in the cluster, and stores the snap data in a single convenient cluster snapshot (csnap) as a compressed .tar file on the node that initiated this command. The snap data contains the configuration information that might be necessary to identify and resolve system problems.

Syntax

clffdc -c component [-l localCorrelator] [-p priority] [-v verbosity] [-f file] 
[-n lineNumber] [-g correlator] [-s]

Description

The clffdc command captures snap data from all the nodes in a Cluster Aware AIX® (CAA) cluster. A cluster-wide snap operation gets triggered automatically by the operating system when a severe problem is detected. You can use the clffdc command to simplify snap data collection across the cluster.

The cluster-wide snap file is created in a default directory. For a Virtual I/O Server (VIOS) environment, the cluster-wide snap files are located in the /home/ios/logs/ssp_ffdc directory. For a non-VIOS environment, the cluster-wide snap files are located in the /var/adm/ras/cl_ffdc directory.

Each node in the cluster creates a snap file. The snap files are collected from each node and merged into a single convenient csnap file on the node that initiated the cluster-wide snap operation. The csnap file name uses the following format:

csnap_date_time_by_component_priority_ccorrelator.tar.gz

The snap file name uses the following format:

snap_date_time_by_component_priority_ccorrelator.tar.gz

Only a single cluster-wide snap operation can occur at a time. If a previous cluster-wide snap operation is in progress, a new cluster-wide snap operation cannot be initiated until the previous operation times out. Each cluster-wide snap operation is associated with a correlator value on the CAA repository disk. This correlator value increments when a new cluster-wide snap operation occurs. If the repository disk is inaccessible when a cluster-wide snap operation is initiated, a csnap file is not generated. In this scenario, each node generates a snap file with a timestamp, but a correlator value is not specified.

If the node that initiated the cluster-wide snap operation goes offline while the cluster-wide snap operation is in progress, each node creates a snap file but a csnap file is not created. If a noninitiator node goes offline while the cluster-wide snap operation is in progress, the initiator node waits for a timeout period before it captures the csnap file from the available nodes.

A new initiator node can collect the snap files by running the clffdc -g command.

The -c, -f, and -n flags are used to identify the location in the code that requested the snap data if the snap file was created automatically by the AIX operating system. If you manually collect the snap data, you must specify the -c flag to identify the component that is responsible for calling any other associated peer components during a snap collection.

Each new cluster-wide snap operation deletes the old csnap files and old snap files that are located in the default directory.

Flags

Flag name Description
-c component Specifies the component that requested the cluster-wide snap operation. The component attribute can have the following values:
  • CAA (Cluster Aware AIX)
  • RSCT (Reliable Scalable Cluster Technology)
  • VIOS (Virtual I/O Server)
  • POOL (Shared storage pool)
  • PHA (PowerHA® SystemMirror®)
  • FULL
Note: The FULL value indicates that full snap data is collected on each node by using the snap -a command. Any other value indicates that a miniature snap data is collected on each node. The miniature snap data starts with the specified component and includes all peer components that are associated with that component.
-f file Specifies the source file name within the component that initiated the cluster-wide snap operation. If the file name is not specified, the clffdc.c file name is used by default.
-g correlator Gathers the cluster-wide snap files. The gathering operation collects a series of snap files that have the specified correlator value on each node, and groups the snap files to create a single csnap file on the initiator node. The correlator value is specified as a decimal value. This flag is useful when used with the -s flag, or when a previous cluster-wide snap operation was interrupted before a csnap file might be generated.

Each node generates a snap file that has the specified correlator value. You can use this flag to collect the individual snap files and create a csnap file that represents the snap data from the entire cluster.
-l localCorrelator Requests snap operation on a local node. The localCorrelator value is the correlator value in decimal format that is used to name the resulting snap file.
-p priority Specifies the priority for the cluster-wide snap operation. The priority attribute can have the following values:
  • 1 (high priority)
  • 2 (medium priority)
  • 3 (low priority)
The priority is used as part of the name in the resulting snap file and the csnap file.
-n lineNumber Specifies the line number of the caller who requested the cluster-wide snap operation.
-s Initiates a staged cluster-wide snap collection. A staged collection indicates that the snap files are created on each node, but not gathered into a csnap file on the initiator node. This flag is useful when used with the -g flag, which gathers the snap files into a single csnap file on the initiator node.
-v verbosity Specifies the verbosity for the cluster-wide snap operation. Possible values that can be specified with the -v flag are 0 or 1. You can specify 1 to collect more information for certain components during the cluster-wide snap operation.

Exit Status

Return code Description
0 The command completed successfully.
>0 A problem occurred.

Examples

  1. To collect a cluster-wide snap data that is associated with the CAA component with medium priority, enter the following command:
    clffdc -c CAA -p 2
    Note: In a VIOS environment, the associated components are CAA, RSCT, POOL, and VIOS. In a (PowerHA SystemMirror) environment, the associated components are CAA, RSCT, and PHA. The specified component and each associated peer component collect snap data for the cluster-wide snap operation.
  2. To collect a cluster-wide snap data that contains the full snap data (collected by the snap -a command) with low priority, enter the following command:
    clffdc -c FULL -p 3
  3. To initiate a staged cluster-wide snap operation that is associated with the PHA component (PowerHA SystemMirror) with high priority, enter the following command:
    clffdc -c PHA -p 1 -s
  4. To gather snap files on each node with the correlator value 77 into a single convenient csnap file on the initiator node, enter the following command:
    clffdc -g 77

Files

Item Description
/usr/sbin/clffdc Contains the clffdc command.
/var/adm/ras/cl_ffdc Contains the clffdc command output in a non-VIOS environment.
/home/ios/logs/ssp_ffdc Contains the clffdc command output in a VIOS environment.