SHRUB Diagnostics - UG1586

Determining Which Controllers Are Running

To determine which controllers are running, you can list which processes are running:

$ pgrep -af shrub_controller
98940 /sbin/shrub_controller -c 1 -D -K

Alternatively you can check which controllers have created domain sockets:

$ ls /run/onload/
controller-1

Using onload_shrubdump

You can use onload_shrubdump to output diagnostic information.

The output below is an example of the output from onload_shrubdump for Onload-9.1. The format and details provided are likely to change in future releases. In the event of problems with SHRUB, open a technical support case via support-nic@amd.com and include this output.

At present, the dump must be run for a specified controller number.

$ onload_shrubdump -i 1
---------------------------------------------------------
shrub controller
  name: controller-1
  dir: /run/onload/controller-1/
  config socket: /run/onload/controller-1/shrub_config
---------------------------------------------------------
shrub server
  ifindex: 23 hwports: 5
  buffer count: 4 client count: 6 token: 2
  vi: 512 active_qs: 0
---------------------------------------------------------
shrub server
  ifindex: 22 hwports: 4
  buffer count: 4 client count: 6 token: 1
  vi: 512 active_qs: 1
  rxq[0]: hw: 0
    sbseq: 2297215 free_head: -1 fifo_head: 1
    tail_hw: -1 tail_sw: -1 count_hw: 0 count_sw: 0
---------------------------------------------------------
controller statistics:
  client negotiation failures: 0
  accept failures: 0
  response send failures: 0
  incompatible clients detected: 0