cmeval(1m)																cmeval(1m)

NAME

       cmeval - evaluate the impact of Serviceguard configuration and state changes on the packages

SYNOPSIS

       [-v] output_file] input_file

DESCRIPTION

       is  a "what if" analysis tool that evaluates the effect of configuration and state changes on Serviceguard packages.  Combinations of pack-
       ages usually have numerous constraints, such as co-existence or mutual exclusion, amongst others; and allows one to evaluate  the  possible
       impact under various "what if" scenarios of Serviceguard specific events.

       accepts input from a file containing the machine parsable output of the command specifying the whole cluster configuration that needs to be
       evaluated.  If the root user invokes the command, the input file must be owned by root and be writable only by root.

       The -v option displays verbose output. Without the -v option, only the final states of the packages are displayed. With the -v option,  the
       intermediate states of the packages are also displayed.

       The  command  performs  limited	validation  on	the input file before feeding the data for evaluation. Since this command  is intended for
       expert users, it is user's responsibility to provide the data input in a syntactically correct  order.  The  behavior  of  the  program	is
       unspecified  (including	the exit codes) when the syntax is unexpected. The best way to determine the syntax is to look at the output of or
       to read the man page.

       can safely be used in a production environment without affecting the current state of the cluster or packages. The command doesn't  execute
       the package run/halt scripts so the failures encountered by those scripts are not taken into account while generating the output file.  One
       need not be logged in to the cluster being evaluated in order to run the evaluations, but is  only  supported  when  running  on  the  same
       release/patch  version as the cluster whose data is being evaluated.  Reports can be generated offline.	doesn't support package in mainte-
       nance mode. If the information of package in maintenance mode in the cmviewcl line output is modified, the prediction might  not  be  accu-
       rate.

   Options
       supports the following options:

       The output file contains the list of affected packages and is designed
			 for  simple  machine parsing using tools such as grep(1), cut(1), or awk(1).  By default, the output of this command will
			 go to the stdout.  The output file will be empty if there are no affected packages for the simulated events.

			 The information contains the following parameters:

			 Name of the affected package

			 Name of the node where effect takes place

			 New state the package will transition to

			 Each line of output is in the form:

			 The possible action values are :

			 failwait, relocate_wait, starting, down, and failing.

			 Please refer cmviewcl(5) manpage for the meaning of the states.

       input_file
	      The input file contains the output of command for the configuration/state changes that are to be evaluated.

       -v     Verbose output will be displayed. This option provides intermediate actions associated with the state changes on Serviceguard  pack-
	      ages.

       The following table illustrates the mappings for the supported configuration events with a two node/two package cluster example in mind.  A
       combination of these events can be evaluated simultaneously.  The events include the actual commands for comparision.  Although the  actual
       commands  would	check that the input is valid before executing, and might not execute, or might fail, under some conditions, does not per-
       form these validations and will always execute.

------------------------------------------------------------------------------
Event		   How to Simulate the	 Example Segment From an Output
		   Event Using		 File
		   cmviewcl Line
		   Output?
------------------------------------------------------------------------------
Cluster   member   Change     specific	 node:node1|status=up
node1  start  up   node status	to  up	 package:pkg1|node:node1|avail-
cmrunnode     -t   and node availabil-	 able=yes
node1)		   ity to yes for each	 package:pkg2|node:node1|avail-
		   package  configured	 able=yes
		   to run on the node.
------------------------------------------------------------------------------
Cluster   member   Change     specific	 node:node1|status=down
node1  shutdown/   node status to down	 package:pkg1|node:node1|available=no
failure 	   and node availabil-	 package:pkg2|node:node1|available=no
(cmhaltnode   -t   ity	to no for each
node1)		   package  configured
		   to run on the node.
------------------------------------------------------------------------------
Cluster bring up   Change node	status	 node:node1|status=up
(cmruncl -t)	   to	up   and  node	 node:node2|status=up
		   availability to yes	 package:pkg1|autorun=enable
		   for	 each  package	 package:pkg1|node:node1|avail-
		   configured  to  run	 able=yes
		   on the cluster mem-	 package:pkg1|node:node2|avail-
		   ber	 nodes.   Make	 able=yes
		   sure that all pack-	 package:pkg2|autorun=enable
		   ages  autorun  flag	 package:pkg2|node:node1|avail-
		   is enabled.		 able=yes
					 package:pkg2|node:node2|avail-
					 able=yes
------------------------------------------------------------------------------
Enable	 package   Change      package	 package:pkg1|autorun=enabled
pkg1   switching   autorun   flag   to	 package:pkg1|node:node1|avail-
(cmmodpkg  -t -e   enabled   and   the	 able=yes
pkg1)		   desired    (usually	 package:pkg1|node:node2|avail-
		   all)  node	avail-	 able=yes
		   ability to yes.
------------------------------------------------------------------------------
Enable	 package   Change package node	 package:pkg1|node:node1|avail-
pkg1	    node   availability     to	 able=yes
switching	   yes.
(cmmodpkg -t  -n
n1 -e pkg1)
------------------------------------------------------------------------------
Package     pkg1   Change      package	 package:pkg1|autorun=enabled
startup   on   a   autorun   flag   to	 package:pkg1|node:node1|avail-
viable	    node   enabled and all the	 able=yes
{cmrunkpg  -t -a   node   availability	 package:pkg1|node:node2|avail-
pkg1)		   to yes.		 able=yes
------------------------------------------------------------------------------
Package     pkg1   Change      package	 package:pkg1|autorun=enabled
startup on node1   autorun   flag   to	 package:pkg1|node:node1|avail-
(cmrunpkg -t  -n   enabled,  the  spe-	 able=yes
node1 pkg1)	   cific  node	avail-	 package:pkg1|node:node2|available=no
		   ability  to yes and
		   all	 other	  node
		   availability to no.
------------------------------------------------------------------------------

Package     pkg1   Change      package	 package:pkg1|autorun=enabled
startup on local   autorun   flag   to	 package:pkg1|node:node1|avail-
node   (cmrunpkg   enabled,  the local	 able=yes
-t pkg1)	   node   availability	 package:pkg1|node:node2|available=no
		   to	yes   and  all
		   other  node	avail-
		   ability to no.
------------------------------------------------------------------------------
Package     pkg1   Change      package	 package:pkg1|autorun=disabled
shutdown/  fail-   autorun   flag   to	 package:pkg1|node:node1|available=no
ure   (cmhaltpkg   disabled   and  the
-t    pkg1    or   package owner  node
cmhaltpkg -t  -n   availability to no
node1 pkg1)
------------------------------------------------------------------------------
Package     pkg1   Change node	avail-	 package:pkg1|autorun=enabled
migration   from   ability for all the	 package:pkg1|node:node1|available=no
one   node1   to   configured nodes to	 package:pkg1|node:node2|avail-
node2		   no  except  for the	 able=yes
		   node where  package
		   needs  to  move  to
		   yes. Make sure that
		   the	  autorun   is
		   enabled.
------------------------------------------------------------------------------
Package pkg1 and   Set the priority to	 package:pkg1|priority=10
pkg2	priority   the desired number	 package:pkg2|priority=20
change
------------------------------------------------------------------------------
Package     pkg1   Set the node  order	 package:pkg1|node:node1|order=2
node	   order   to the desired num-	 package:pkg1|node:node2|order=1
change		   ber
------------------------------------------------------------------------------
Subnet,  EMS  or   Change   the   node	 package:pkg1|node:node1|available=no
Storage unavail-   availability  to no	 package:pkg2|node:node2|available=no
ability on node1   for	each   package
		   configured  to  run
		   on that node
------------------------------------------------------------------------------
Add/change	   Add/change capacity	 node:node1|capacity:memory|name=mem-
capacity   named   name and limit	 ory	     node:node1|capacity:mem-
"memory"     for			 ory|limit=10000
node node1
------------------------------------------------------------------------------
Add/change pack-   Add/change	weight	 weight_default:memory|name=memory
age	  weight   default   name  and	 weight_default:memory|value=10
default    named   value
"memory" for all
packages
------------------------------------------------------------------------------
Add/change	   Add/change	weight	 package:pkg1|weight:memory|name=mem-
weight	   named   name and value	 ory
"memory"     for			 package:pkg1|weight:memory|value=20
package pkg1
------------------------------------------------------------------------------

RETURN VALUE

       returns the following value:

	      0  Successful completion.

	      1  Command failed with a syntax error or inaccuracies in the input file.

EXAMPLES

       To evaluate the impact of a node bring-up on two packages with a configured dependency and failback policy set to automatic:(1)  Obtain an existing cluster configuration using command.

	    NOTE:  In  the  existing configuration, is running on an alternate node, as the primary node is not available.  The failback policy is
	    set to automatic.  This means that when is eligible to run it, will automatically fail back to is running on its primary  node.   How-
	    ever, it depends on pkg1 (which has the higher priority setting of 2719) with a same node dependency.  This means when pkg1 fails back
	    to its primary node, will be dragged along with pkg1 to run on the same node as

	    node:node1|name=node1
	    node:node1|status=down
	    :  :
	    node:node2|name=node2
	    node:node2|status=up
	    :  :
	    package:pkg1|name=pkg1
	    package:pkg1|type=failover
	    package:pkg1|status=up
	    package:pkg1|state=running
	    package:pkg1|autorun=enabled
	    package:pkg1|owner=node2
	    package:pkg1|id=3200
	    package:pkg1|initial_autorun=enabled
	    package:pkg1|failover_policy=configured_node
	    package:pkg1|failback_policy=automatic
	    package:pkg1|priority=2719
	    :	:
	    package:pkg1|node:node1|name=node1
	    package:pkg1|node:node1|status=down
	    package:pkg1|node:node1|switching=enabled
	    package:pkg1|node:node1|available=no
	    package:pkg1|node:node1|type=Primary
	    package:pkg1|node:node1|order=1
	    :	:
	    package:pkg1|node:node2|name=node2
	    package:pkg1|node:node2|status=up
	    package:pkg1|node:node2|switching=enabled
	    package:pkg1|node:node2|available=yes
	    package:pkg1|node:node2|type=Alternate
	    package:pkg1|node:node2|order=2
	    :  :
	    package:pkg2|name=pkg1
	    package:pkg2|type=failover
	    package:pkg2|status=up
	    package:pkg2|state=running
	    package:pkg2|autorun=enabled
	    package:pkg2|owner=node2
	    package:pkg2|id=3201
	    package:pkg2|initial_autorun=enabled
	    package:pkg2|failover_policy=configured_node
	    package:pkg2|failback_policy=manual
	    package:pkg2|node:node1|order=1
	    package:pkg2|priority=no_priority
	    :	:
	    package:pkg2|node:node2|name=node2
	    package:pkg2|node:node2|status=up
	    package:pkg2|node:node2|switching=enabled
	    package:pkg2|node:node2|available=yes
	    package:pkg2|node:node2|type=Primary
	    package:pkg2|node:node2|order=1
	    :	:
	    package:pkg2|node:node1|name=node1
	    package:pkg2|node:node1|status=down
	    package:pkg2|node:node1|switching=enabled
	    package:pkg2|node:node1|available=no
	    package:pkg2|node:node1|type=Alternate
	    package:pkg2|dependency:pkg1|node:spanish|satisfied=yes(2)  Modify the configuration file to simulate the event
	    node:node1|status=up
	    package:pkg1|node:node1|available=yes
	    package:pkg2|node:node1|available=yes
	    package:pkg2|node:node1|available=yes(3)  Invoke to simulate the event.  The output will show the effect of the node1 bring-up .
	    package:pkg1|node:node2|action=failing
	    package:pkg2|node:node2|action=failing
	    package:pkg1|node:node1|action=starting
	    package:pkg2|node:node1|action=starting(4)  Invoke to simulate the event.  The output will show the end effect of the node1 bring-up.
	    package:pkg1|node:node1|action=starting
	    package:pkg2|node:node1|action=starting

AUTHOR

       was developed by HP.

SEE ALSO

       cmhaltnode(1M), cmhaltpkg(1M), cmmodpkg(1M), cmrunnode(1M), cmrunpkg(1M), cmruncl(1M), cmviewcl(1M).

						      Requires Optional Serviceguard Software						cmeval(1m)