CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University CS 555: D ISTRIBUTED S YSTEMS [G OOGLE F ILE S YSTEM ] Shrideep Pallickara Computer Science Colorado State University CS555: Distributed Systems [Fall 2019] November 19, 2019 L25.1 Dept. Of Computer Science , Colorado State University Frequently asked questions from the previous class survey ¨ Which is better: GFS or Dynamo? L25. 2 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.1 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Topics covered in this lecture ¨ Google File System ¤ Metadata management ¤ Managing mutations L25. 3 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA A LL system metadata is managed by the Master and stored in Main Memory ① File and chunk namespaces ② Mapping from files to chunks Logs mutations into a permanent log ③ Location of chunks L25. 4 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.2 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Why have a single Master? ¨ Vastly simplifies design ¨ Easy to use global knowledge to reason about ¤ Chunk placements ¤ Replication decisions L25. 5 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA Communications with the chunk servers ¨ Periodic communications using heartbeats ¤ Instructions to the chunk server ¤ Collect/retrieve state from the chunk server L25. 6 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.3 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Chunk size ¨ This is fixed at 64 MB ¤ Much larger than typical filesystem block sizes (512 bytes) ¨ Lazy space allocation ¤ Stored as plain Linux file ¤ Extended only as needed L25. 7 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA But why this big? ¨ Reduces client interaction with the master ¤ Can cache info for a multi-TB working set ¨ Reduce network overhead ¤ With a large chunk, client performs more operations ¤ Persistent connections ¨ Reduce size of metadata stored in the master ¤ 64 bytes of metadata per 64 MB chunk L25. 8 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.4 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Why keep the entire metadata in memory? ¨ Speed ¨ Master can scan its state in the background ¤ Implement chunk garbage collection ¤ Re-replicate if there are failures ¤ Chunk migration to balance load and space ¨ Add extra memory to increase file system size L25. 9 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA Size of the file system with 1 TB of RAM: Assume file sizes are exact multiples of chunk sizes ¨ Number of entries = 2 40 /2 6 ¨ M AXIMUM S IZE of the file system = Number of entries x Chunk size = 2 40 x 2 6 x 2 20 2 6 = 2 60 = 1 EB L25. 10 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.5 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Tracking the chunk servers ¨ Master does not keep a persistent copy of the location of chunk servers ¨ List maintained via heart-beats ¤ Allows list to be in sync with reality despite failures ¤ Chunk server has final word on chunks it holds L25. 11 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA Caching at the client/chunk servers ¨ Clients do not cache file data ¤ At client the working set may be too large ¤ Simplify client; eliminate cache-coherence problems ¨ Chunk servers do not cache file data either ¤ Chunks are stored as local files ¤ Linux’s buffer cache already keeps frequently accessed data in memory L25. 12 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.6 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Handling writes and appends to a file M ANAGING M UTATIONS CS555: Distributed Systems [Fall 2019] November 19, 2019 L25.13 Dept. Of Computer Science , Colorado State University Mutations ¨ Mutation changes the content or metadata of a chunk ¤ Write ¤ Append ¨ Each mutation is performed at all chunk replicas L25. 14 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.7 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University GFS uses leases to maintain consistent mutation order across replicas ¨ Master grants lease to one of the replicas ¤ P RIMARY ¨ Primary picks serial-order ¤ For all mutations to the chunk ¤ Other replicas follow this order n When applying mutations L25. 15 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA Lease mechanism designed to minimize communications with the master ¨ Lease has initial timeout of 60 seconds ¨ As long as chunk is being mutated ¤ Primary can request and receive extensions ¨ Extension requests/grants piggybacked over heart-beat messages L25. 16 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.8 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Revocation and transfer of leases ¨ Master may revoke a lease before it expires ¨ If communications lost with primary ¤ Master can safely give lease to another replica n O NLY A FTER the lease period for old primary elapses L25. 17 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA How a write is actually performed Client MASTER Secondary Replica A Primary Replica Secondary Replica B L25. 18 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.9 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Client pushes data to all the replicas ( I ) ¨ Each chunk server stores data in an LRU buffer until ¤ Data is used ¤ Aged out L25. 19 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA Client pushes data to all the replicas ( II ) ¨ When chunk servers acknowledge receipt of data ¤ Client sends a write request to primary ¨ Primary assigns consecutive serial numbers to mutations ¤ Forwards to replicas L25. 20 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.10 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
CS555: Distributed Systems [Fall 2019] Dept. Of Computer Science , Colorado State University Data flow is decoupled from the control flow to utilize network efficiently ¨ Utilize each machine’s network bandwidth ¨ Avoid network bottlenecks ¨ Avoid high-latency links ¨ Leverage network topology ¤ Estimate distances from IP addresses L25. 21 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA What if the secondary replicas could not finish the write operation? ¨ Client request is considered failed ¨ Modified region is inconsistent ¤ No attempt to delete this from the chunk ¤ Client must handle this inconsistency ¨ Client retries the failed mutation L25. 22 CS555: Distributed Systems [Fall 2019] November 19, 2019 Dept. Of Computer Science , Colorado State University Professor: S HRIDEEP P ALLICKARA L28.11 S LIDES C REATED B Y : S HRIDEEP P ALLICKARA
Recommend
More recommend