snap Command for HACMP
snap Command for HACMP
This document describes a new snap command flag that collects HACMP-related data from all nodes in a cluster. The information in this document pertains to AIX 4.3.3 and above.
It is now possible to gather HACMP testcase data through the AIX snap command on AIX 4.3.3 and AIX 5.1. You must first install the required APARs for AIX, HACMP, and RSCT. This function will also be available in the base of AIX 5.2 and above.
This capability is added through mainly a new flag, -e, although a -m flag was also added to allow for the gathering of information from only those nodes specified rather than from all nodes of the cluster. The snap -e command gathers all HACMP related data from all nodes in the cluster configuration unless the -m flag is included and specifies which nodes to collect data from. The -e flag may not be used with any other flags except -d (data collection directory), and -m, because, unlike all other snap command flags, the -e flag gathers data from multiple nodes.
Previously, HACMP testcase data was gathered through getinfo.sc script provided by L1 HACMP support in Dallas (mainly used by US customers though did get used by some World Trade Centers) or files were gathered manually.
Japan is one example of a location that created its own data gathering script, which stored the data in a much different format from getinfo.sc.
In HAES, a phoenix.snap script exists to gather RSCT data. Some duplicate data was gathered by the getinfo.sc and phoenix.snap scripts, since both scripts were modified to gahter data from both HAES and RSCT.
All HA related data for either HAS (classic) or HAES is now gathered through the one command snap -e issued from the command line. This feature provides the following benefits:
All data gathered through one command, snap -e.
The command is readily available on the system. (previous getinfo.sc script had to be downloaded and installed before running).
All data provided in same format from all customers.
All data gathered at same time. Previously, especially in HAES, HA related data may have been gathered but not RSCT data and customers would then need to gather the RSCT data later, which was sometimes too late.
Duplication between HA and RSCT data has been eliminated, although there is still some duplication of AIX data.
HA data is gathered more quickly through processing nodes in parallel.
Additional data has been added which has been found to be required for many types of problems.
Capability has been added to specify the collection from a list of nodes rather than from all nodes of the cluster, although this function is recommended only when support personnel that you use it.
Pre-calculation of space required for data is implemented through the existing first-pass capability of the snap command.