clubak - format output from clush/pdsh-like output and more

       Stephane Thiell <sthiell@stanford.edu>

UsethefollowingURLtosubmitabugreportorfeedback:
               <https://github.com/cea-hpc/clustershell/issues>

       GNU Lesser General Public License version 2.1 or later (LGPLv2.1+)

1.9.3                                              2025-01-23                                          CLUBAK(1)

clubak formats text from standard input containing lines of the form "node:output".  It is fully backward
       compatible  with  dshbak(1)  but  provides  additional features. For instance, clubak always displays its
       results sorted by node/nodeset.

       You do not need to use clubak when using clush(1) as all output formatting features are already  included
       in. It is provided for other usages, like post-processing results of the form "node:output".

       Like  clush(1),  clubak  uses  the  ClusterShell.MsgTree  module  of  the ClusterShell library (see pydocClusterShell.MsgTree).

       1. clubak can be used to gather some recorded clush(1) results:

       Recordclush(1)resultsinafile:
              # clush -w node[1-7] uname -r >/tmp/clush_output
              # clush -w node[32-159] uname -r >>/tmp/clush_output

       Displayfilegatheredresults(inline-mode):
              # clubak -bL </tmp/clush_output

       2. Another example, iterate over node* text files in current directory and gather  characters  count  for
          all of them:
             # find -name "node*" -exec wc -c {} ; | awk '{ gsub("./","",$2); print $2": "$1 }' | clubak -bL
             node[1,3]: 7
             node2: 9

       An exit status of zero indicates success of the clubak command.

clubak should be started with connected standard input.

       clubak - format output from clush/pdsh-like output and more

--version
              show clubak version number and exit

       -b, -c gather nodes with same output (-c is provided for dshbak(1) compatibility)

       -d, --debug
              output more messages for debugging purpose

       -L     disable header block and order output by nodes

       -r, --regroup
              fold nodeset using node groups

       -sGROUPSOURCE, --groupsource=GROUPSOURCE
              optional groups.conf(5) group source to use

       --groupsconf=FILE
              use alternate config file for groups.conf(5)

       -G, --groupbase
              do not display group source prefix (always @groupname)

       -SSEPARATOR, --separator=SEPARATOR
              node / line content separator string (default: :)

       -F, --fast
              faster but memory hungry mode (preload all messages per node)

       -T, --tree
              message tree trace mode; switch to enable ClusterShell.MsgTree trace mode,  all  keys/nodes  being
              kept for each message element of the tree, thus allowing special output gathering

       --color=WHENCOLORclush  can  use  NO_COLOR, CLICOLOR and CLICOLOR_FORCE environment variables. --color command line
              option always  takes  precedence  over  environment  variables.  NO_COLOR  takes  precedence  over
              CLICOLOR_FORCE  which  takes precedence over CLICOLOR. --color tells whether to use ANSI colors to
              surround node or nodeset prefix/header with escape sequences to  display  them  in  color  on  the
              terminal.  WHENCOLOR  is  never,  always  or  auto (which use color if standard output refers to a
              terminal). Color is set to [34m (blue foreground text) and cannot be modified.

       --diff show diff between gathered outputs

cluset(1), clush(1), nodeset(1), groups.conf(5).

        <http://clustershell.readthedocs.org/>

clubak [ OPTIONS ]