Name

noderange - syntax for compactly expressing a list of node names


Synopsis

Examples:

 node1,node2,node8,node20,group1
 node14-node56,node70-node203,group1-group10
 node1,node2,node8,node20,node14-node56,node70-node203
 node[14-56]
 f[1-3]n[1-20]
 all,-node129-node256,-frame01-frame03
 /node.*
 ^/tmp/nodes
 node10+5
 10-15,-13
 group1@group2
 table.attribute<operator>value


Description

noderange is a syntax that can be used in most xCAT commands to conveniently specify a list of nodes. The result is that the command will be applied to a range of nodes, often in parallel.

noderange is a comma-separated list. Each token (text between commas) in the list can be any of the forms listed below:

Individual node or group:

 node01
 group1

A range of nodes or groups:

 node01-node10  (equivalent to: node01,node02,node03,...node10)
 node[01-10]    (same as above)
 node01:node10  (same as above)
 node[01:10]    (same as above)
 f[1-2]n[1-3]   (equivalent to: f1n1,f1n2,f1n3,f2n1,f2n2,f2n3)
 group1-group3  (equivalent to: group1,group2,group3)
 (all the permutations supported above for nodes are also supported for groups)

nodeRange tries to be intelligent about detecting padding, so you can specify "node001-node200" and it will add the proper number of zeroes to make all numbers 3 digits.

An incremented range of nodes:

 node10+3  (equivalent to: node10,node11,node12,node13)

A node shorthand range of nodes:

 10-20   (equivalent to: node10,node11,node12,...node20)
 10+3    (equivalent to: node10,node11,node12,node13)

Currently, the prefix that will be prepended for the above syntax is always "node". Eventually, the prefix and optional suffix will be settable via the environment variables XCAT_NODE_PREFIX and XCAT_NODE_SUFFIX, but currently this only works in bypass mode.

A regular expression match of nodes or groups:

 /node[345].*   (will match any nodes that start with node3, node4, or node5)
 /group[12].*   (will match any groups that start with group1 or group2)

The path of a file containing noderanges of nodes or groups:

 ^/tmp/nodelist

where /tmp/nodelist can contain entries like:

 #my node list (this line ignored)
 ^/tmp/foo #ignored
 node01    #node comment
 node02
 node03
 node10-node20
 /group[456].*
 -node50

Node ranges can contain any combination:

 node01-node30,node40,^/tmp/nodes,/node[13].*,2-10,node50+5

Any individual noderange may be prefixed with an exclusion operator (default -) with the exception of the file operator (default ^). This will cause that individual noderange to be subtracted from the total resulting list of nodes.

The intersection operator @ calculates the intersection of the left and right sides:

 group1@group2   (will result in the list of nodes that group1 and group2 have in common)

Any combination or multiple combinations of inclusive and exclusive ranges of nodes and groups is legal. There is no precedence implied in the order of the arguments. Exclusive ranges have precedence over inclusive. Parentheses can be used to explicitly specify precendence of any operators.

Nodes have precedence over groups. If a node range match is made then no group range match will be attempted.

All node and group names are validated against the nodelist table. Invalid names are ignored and return nothing.

xCAT Node Name Format

Throughout this man page the term xCAT Node Name Format is used. xCAT Node Name Format is defined by the following regex:

 ^([A-Za-z-]+)([0-9]+)(([A-Za-z-]+[A-Za-z0-9-]*)*)

In plain English, a node or group name is in xCAT Node Name Format if starting from the begining there are:

noderange supports node/group names in any format. xCAT Node Name Format is not required, however some node range methods used to determine range will not be used for non-conformant names.

Example of xCAT Node Name Format node/group names:

 NODENAME           PREFIX      NUMBER   SUFFIX
 node1              node        1
 node001            node        001
 node-001           node-       001
 node-foo-001-bar   node-foo-   001      -bar
 node-foo-1bar      node-foo-   1        bar
 foo1bar2           foo         1        bar2
 rack01unit34       rack        01       unit34
 unit34rack01       unit        34       rack01
 pos0134            pos         0134


Examples

  1. Generates a list of all nodes (assuming all is a group) listed in the nodelist table less node5 through node10:

     all,-node5-node10
  2. Generates a list of nodes 1 through 10 less nodes 3,4,5. Note that node4 is listed twice, first in the range and then at the end. Because exclusion has precedence node4 will be excluded.

     node1-node10,-node3-node5,node4
  3. Generates a list of nodes 1 through 10 less nodes 3 and 5.

     node1-node10,-node3,-node5
  4. Generates a list of all (assuming `all' is a group) nodes in the nodelist table less 17 through 32.

     -node17-node32,all
  5. Generates a list of nodes 1 through 128, and user nodes 1 through 4.

     node1-node128,user1-user4
  6. Generates a list of all nodes (assuming `all' is a group), less nodes in groups rack1 through rack3 (assuming groups rack1, rack2, and rack3 are defined), less nodes 100 through 200, less nodes in the storage group. Note that node150 is listed but is excluded.

     all,-rack1-rack3,-node100-node200,node150,-storage
  7. Generates a list of nodes matching the regex node[23].*. That is all nodes that start with node2 or node3 and end in anything or nothing. E.g. node2, node3, node20, node30, node21234 all match.

     /node[23].*
  8. Generates a list of nodes which have the value hmc in the nodehm.cons attribute.

     nodehm.cons==hmc
     nodehm.cons=~hmc
  9. Generate a list of nodes in the 1st two frames:

     f[1-2]n[1-42]


SEE ALSO

nodels(1)