How do I use the opacapture command?
The opacapture command is part of the tools installed with the Intel® Omni-Path Architecture (Intel® OPA) software. It captures critical system information into a zipped tar file. The resulting host capture file can require significant amounts of space on the host. The actual size varies, but sizes can be multiple megabytes. Therefore, Intel recommends that you ensure that adequate disk space is available on the host system.
The syntax for the capture command is: opacapture -d 3 <output_tgz_file>
<output_tgz_file>: Specifies the name of a file that opacapture creates. The file name specified is overwritten if it already exists. Intel recommends that you use the .tgz suffix in the file name supplied. If the file name given doesn't have a .tgz suffix, the .tgz suffix is added.
We also recommend including the node name and date (_yyyymmdd) as part of the capture file name.
Can you give me some examples of how to use opacapture?
opacapture -d 3 node_abc_001_20180102.tgz
The log file will be saved in the directory where it was initiated. A message at the end of the opacapture will indicate where the file is located.
Creating tar file /tmp/logs/node_abc_00120180102.tgz ...
Include /tmp/logs/Node_abc_00120180102.tgz with any problem reports to Customer Support.
What opacapture is needed?
An Intel OPA capture should be provided for the head node (node that is running the Fabric Manager (FM)) and one of the problem nodes. The Support team can provide you with additional logs later if needed.
Send the resulting tar file to Customer Support along with any Intel® Omni-Path Fabric problem description report regarding this system.
In the case of Intel® Omni-Path Edge Switches, the capture command depends on the type of switch:
|Reporting a Problem Related to Intel® Omni-Path Fabric or Intel® True Scale Fabric on a Linux* Node|