Robust recovery of a page parallel RIP system
What is disclosed is a novel fault tolerant page parallel RIP system having a plurality of distributed RIP nodes and a method for robust recovery in the event of a fault having occurred on the system. In one embodiment, the present fault tolerant page parallel RIP system comprises at least a plurality of RIP nodes, a supervisor node, and a splitter node. The splitter receives a location of the job file and splits the job file into at least one original chunk. In formation sufficient to regenerate the original chunk in the event of a fault condition having occurred on the page parallel RIP system is stored. The splitter then sends the original chunk to a destination RIP node wherein the original chunk is RIP'ed into at least one page in print-ready form. The one page is the provided in print-ready form to an output subsystem. Various embodiments are disclosed.
1. A method for robust recovery in a page parallel RIP system that includes a splitter node and a plurality of distributed RIP nodes, the method comprising:
receiving a job file to be processed in a page parallel RIP system;
informing, by a supervisor node, a splitter node of a location of said job file;
splitting, by said splitter node, said job file into a plurality of chunks;
storing, by said splitter node, a chunk data structure for each of said plurality of chunks, each chunk data structure containing information required to regenerate a corresponding chunk in an event of a fault condition on said page parallel RIP system while performing a RIP process by a RIP node on a chunk that corresponds to a chunk data structure stored by said splitter node, said information comprising a list of resources required by said corresponding chunk, said list of resources comprising page content and a list of dependencies comprising a set of resource identifiers, said splitter node maintaining said set of resource identifiers for each RIP node;
assigning, by said supervisor node, one of said chunks to a first RIP node;
in response to a signal indicating that said first RIP node is no longer active or is unresponsive, said supervisor node selecting a second RIP node which is still online and active in said page parallel RIP system, said second RIP node being different than said first RIP node, and communicating an identity of said selected second RIP node to said splitter node, said splitter node receiving a list of chunks pending on said first RIP node, said list of chunks pending including said one chunk;
retrieving, by said splitter node, said chunk data structure corresponding to said one chunk;
using said retrieved chuck data structure to regenerate said one chunk;
sending said regenerated chunk to said selected second RIP node;
processing, by said second RIP node, said regenerated chunk into at least one print-ready page;
transferring said at least one print-ready page from said second RIP node to an output subsystem; and
in response to said output subsystem having received all print-ready pages for this chunk, said splitter node deleting said chunk data structure corresponding to said regenerated chunk.
2. The method of claim 1 , further comprising, in response to said at least one print-ready page having been transferred to said output subsystem, suppressing print-ready pages of said regenerated chunk that were previously transferred.
3. A fault tolerant page parallel RIP system having a plurality of distributed nodes, the fault tolerant page parallel RIP system comprising:
a plurality of RIP nodes; and
at least one splitter node, performing:
receiving a job file to be processed;
informing, by a supervisor node, a splitter node of a location of said job file;
splitting, by a splitter node, said job file into a plurality of chunks;
storing, by said splitter node, a chunk data structure for each of said plurality of chunks, each chunk data structure containing information required to regenerates a corresponding chunk in an event of a fault condition having occurred on said page parallel RIP system while performing a RIP process by a RIP node on a chunk that corresponds to a chunk data structure stored by said splitter node, said information comprising a list of resources required by said corresponding chunks, said list of resources comprising page content and a list of dependencies comprising a set of resource identifiers, said splitter node maintaining said set of resource identifiers for each RIP node;
assigning, by said supervisor node, one of said chunks to a first RIP node;
after sending said one chunk to said first RIP node and in response to a signal indicating that said first RIP node is no longer active or unresponsive, said supervisor node selecting a second RIP node which is still online and active in said page parallel RIP system, said second RIP node being different than said first RIP node, and communicating an identity of said selected second RIP node to said splitter node, said splitter node receiving a list of chunks pending on said first RIP node, said list of chunks pending including said one chunk;
retrieving, by said splitter node, said chunk data structure corresponding to said one chunk;
using said retrieved chunk data structure to regenerate said one chunk;
sending said regenerated chunk to said second RIP node, wherein said second RIP node RIPs said regenerated chunk into at least one print-ready page, and transferring said at least one print-ready page to an output subsystem; and
in response to said output subsystem having received all print-ready pages for this chunk, said splitter node deleting said chunk data structure corresponding to said regenerated chunk.
4. The fault tolerant page parallel RIP system of claim 3 , wherein, in response to said at least one print-ready page having been transferred to said output subsystem, further comprising suppressing print-ready pages of said regenerated chunk that were previously transferred.
5. The method of claim 1 , wherein said list of resources further includes content of a preamble, a file offset, an identity of at least one destination RIP node in said page parallel RIP system, and RIP node addresses to which chunks of each job being RIP'ed have been sent.
6. The method of claim 1 , wherein, in response to said first RIP node no longer being active, said splitter node re-tailoring said chunk to a new destination RIP node, and re-sending said chunk to said new destination RIP node.
7. The fault tolerant page parallel RIP system of claim 3 , wherein said list of resources further includes content of a preamble, a file offset, an identity of at least one destination RIP node in said page parallel RIP system, and RIP node addresses to which chunks of each job being RIP'ed have been sent.
8. The fault tolerant page parallel RIP system of claim 3 , wherein, in response to said first RIP node no longer being active, said splitter node re-tailoring said chunk to a new destination RIP node, and re-sending said chunk to said new destination RIP node.