Identification and verification of common cluster files residing on nodes in a cluster
First Claim
1. A method for identification and verification of common cluster files and/or directories residing on nodes in a cluster, the method comprising:
- compiling a first list of all files stored on all nodes of the cluster, wherein a file name concatenated to a complete directory path concatenated to a node number for each file constitutes a primary key for the first list;
compiling a second list of primary keys from the first list by removing the node number from each primary key;
sorting the second list by primary key;
compiling a third list of unique primary keys from the second list together with a primary key count representing the number of items in the first list for each unique primary key;
sorting the third list by primary key count;
compiling a fourth list of unique primary keys from the third list of unique primary keys by discarding any entry from the third list in which the primary key count is equal to the number of nodes in the cluster;
compiling a fifth list of unique primary keys from the fourth list of unique primary keys by discarding any entry from the fourth list in which the primary key count is less than or equal to a predetermined threshold indicative of the population of unique files; and
storing the fifth list on a computer readable medium.
1 Assignment
0 Petitions
Accused Products
Abstract
In accordance with a particular embodiment of the present disclosure, common cluster files residing on nodes in a cluster may be managed by compiling a first list of all files stored on all nodes of the cluster, compiling a second list indicative of unique files and the number of nodes on which each unique file is stored from the first list, determining, from the second list, unique files which are not stored on all nodes, determining, from the second list, which files are required by all nodes, and determining, from the first list and the second list, which files must be added to each node.
122 Citations
20 Claims
-
1. A method for identification and verification of common cluster files and/or directories residing on nodes in a cluster, the method comprising:
-
compiling a first list of all files stored on all nodes of the cluster, wherein a file name concatenated to a complete directory path concatenated to a node number for each file constitutes a primary key for the first list; compiling a second list of primary keys from the first list by removing the node number from each primary key; sorting the second list by primary key; compiling a third list of unique primary keys from the second list together with a primary key count representing the number of items in the first list for each unique primary key; sorting the third list by primary key count; compiling a fourth list of unique primary keys from the third list of unique primary keys by discarding any entry from the third list in which the primary key count is equal to the number of nodes in the cluster; compiling a fifth list of unique primary keys from the fourth list of unique primary keys by discarding any entry from the fourth list in which the primary key count is less than or equal to a predetermined threshold indicative of the population of unique files; and storing the fifth list on a computer readable medium. - View Dependent Claims (2, 3, 4, 5, 6, 7)
-
-
8. A method for managing common cluster files residing on nodes in a cluster, the method comprising:
-
compiling a first list of all files stored on all nodes of the cluster; compiling a second list indicative of unique files and the number of nodes on which each unique file is stored from the first list; determining, from the second list, unique files which are not stored on all nodes; determining, from the second list, which files are required by all nodes; and determining, from the first list and the second list, which files must be added to and/or replaced on each node. - View Dependent Claims (9, 10, 11, 12, 13, 14)
-
-
15. Logic for managing common cluster files residing on nodes in a cluster, the logic embodied in a computer-readable medium and operable to:
-
compile a first list of all files stored on all nodes of the cluster; compile a second list indicative of unique files and the number of nodes on which each unique file is stored from the first list; determine, from the second list, unique files which are not stored on all nodes; determine, from the second list, which files are required by all nodes; and determine, from the first list and the second list, which files must be added to and/or replaced on each node. - View Dependent Claims (16, 17, 18, 19, 20)
-
Specification