Presentation is loading. Please wait.

Presentation is loading. Please wait.

Ceph Distributed File System: Simulating a Site Failure s: {bazli.abkarim, mt.wong, Mohd Bazli Ab Karim, Ming-Tat Wong, Jing-Yuan.

Similar presentations


Presentation on theme: "Ceph Distributed File System: Simulating a Site Failure s: {bazli.abkarim, mt.wong, Mohd Bazli Ab Karim, Ming-Tat Wong, Jing-Yuan."— Presentation transcript:

1 Ceph Distributed File System: Simulating a Site Failure s: {bazli.abkarim, mt.wong, Mohd Bazli Ab Karim, Ming-Tat Wong, Jing-Yuan Luke Advanced Computing Lab MIMOS Berhad, Malaysia In PRAGMA 26, Tainan, Taiwan 9-11 April 2014

2 Outline Motivation Problems Solution Demo Moving forward 2

3 Motivations Explosion of both structured and unstructured data in cloud computing as well as in traditional datacenters presents a challenge for existing storage solution from cost, redundancy, availability, scalability, performance, policy, etc. Our motivation thus focus leveraging on commodity hardware/storage and networking to create a highly available storage infrastructure to support future cloud computing deployment in a Wide Area Network, multi-sites/multi- datacenters environment. 3

4 Problems Data Center Disaster Recovery Site(s) SAN/NAS R/W Replication/De- duplication R/W 4 Performance  Performance Redundancy  Redundancy Availability/Reliability  Availability/Reliability

5 Data Center SAN/NAS Disaster Recovery Site(s) SAN/NAS Solution Data Center 1 Local/DAS Data Center 2 Local/DAS Data Center n Local/DAS... One/Multiple Virtual Volume(s) R/W Replications Data Striping and Parallel R/W 5

6 CRUSH – Controlled, Scalable, Decentralized Placement of Replicated Data – It is an algorithm to determine how to store and retrieve data by computing data storage locations. Why? – To use the algorithm to organize and distribute the data to different datacenters. 6 Challenging the CRUSH algorithm

7 7 CRUSH Map osd.0 Bucket osd.0 Bucket host Bucket host Bucket osd.1 Bucket osd.1 Bucket host Bucket host Bucket osd.2 Bucket osd.2 Bucket host Bucket host Bucket osd.3 Bucket osd.3 Bucket host Bucket host Bucket osd.4 Bucket osd.4 Bucket host Bucket host Bucket osd.5 Bucket osd.5 Bucket host Bucket host Bucket osd.6 Bucket osd.6 Bucket host Bucket host Bucket osd.7 Bucket osd.7 Bucket host Bucket host Bucket osd.8 Bucket osd.8 Bucket host Bucket host Bucket osd.9 Bucket osd.9 Bucket host Bucket host Bucket osd.10 Bucket osd.10 Bucket host Bucket host Bucket osd.11 Bucket osd.11 Bucket host Bucket host Bucket datacenter Bucket datacenter Bucket datacenter Bucket datacenter Bucket datacenter Bucket datacenter Bucket rootBucketrootBucket OBJECT A Replica A OBJECT B Replica B OBJECT C Replica C DEFAULT SANDBOX ENVIRONMENT IF LARGE SCALE, WE NEED A CUSTOM CRUSH MAP OBJECT D Replica D Replications ENSURE DATA SAFETY

8 ceph osd tree # id weight type name up/down reweight root default host poc-tpm1-osd osd.0 up host poc-tpm1-osd osd.1 up host poc-tpm1-osd osd.2 up host poc-tpm1-osd osd.3 up host poc-tpm2-osd osd.4 up host poc-tpm2-osd osd.5 up host poc-tpm2-osd osd.6 up host poc-tpm2-osd osd.7 up host poc-khtp-osd osd.8 up host poc-khtp-osd osd.9 up host poc-khtp-osd osd.10 up host poc-khtp-osd osd.11 up 1 8 CRUSH Map - default

9 # rules rule data { ruleset 0 type replicated min_size 1 max_size 10 step take default step chooseleaf firstn 0 type host step emit } rule metadata { ruleset 1 type replicated min_size 1 max_size 10 step take default step chooseleaf firstn 0 type host step emit } 9 CRUSH Map Rules - default Pick one leaf node of type host Pick one leaf node of type host

10 ceph osd tree # id weight type name up/down reweight root default datacenter tpm host poc-tpm1-osd osd.0 up host poc-tpm1-osd osd.1 up host poc-tpm1-osd osd.2 up host poc-tpm1-osd osd.3 up datacenter tpm host poc-tpm2-osd osd.4 up host poc-tpm2-osd osd.5up host poc-tpm2-osd osd.6 up host poc-tpm2-osd osd.7 up datacenter khtp host poc-khtp-osd osd.8 up host poc-khtp-osd osd.9 up host poc-khtp-osd osd.10 up host poc-khtp-osd osd.11 up 1 10 CRUSH Map - New

11 # rules rule data { ruleset 0 type replicated min_size 2 max_size 10 step take default step chooseleaf firstn 0 type datacenter step emit } rule metadata { ruleset 1 type replicated min_size 2 max_size 10 step take default step chooseleaf firstn 0 type datacenter step emit } 11 CRUSH Map Rules – New Pick one leaf node of type datacenter Pick one leaf node of type datacenter

12 DEMO 12

13 13 WAN DC1 DC2 Mimos Berhad Kulim Hi-Tech Park DC3 Mimos Berhad Technology Park Malaysia, Kuala Lumpur 350 KM It was first started as a proof of concept for Ceph as a DFS over wide area network. It was first started as a proof of concept for Ceph as a DFS over wide area network. Two sites had been identified to host the storage servers – MIMOS HQ and MIMOS Kulim Two sites had been identified to host the storage servers – MIMOS HQ and MIMOS Kulim Collaboration work between MIMOS and SGI. Collaboration work between MIMOS and SGI. In PRAGMA 26, we will use this Ceph POC setup to demonstrate a site failure of a geo-replication distributed file system over wide area network. In PRAGMA 26, we will use this Ceph POC setup to demonstrate a site failure of a geo-replication distributed file system over wide area network. Demo Background

14 This Demo… 14 WAN DC1 DC2 Mimos Berhad Kulim Hi-Tech Park Demo: Simulate node/site failure while doing read write ops. Test Plan: (a)From DC1, continuously ping servers in Kulim. (b)Upload 500Mb file to the file system. (c)While uploading, take down nodes in Kulim. From (a), check if nodes are down. (d)Upload completed, download the same file. (e)While downloading, bring up the nodes in Kulim. (f)Checksum both files. Both should be same. DC3 Mimos Berhad Technology Park Malaysia, Kuala Lumpur 350 KM

15 15 Demo in progress… mon-01 osd01-1 osd01-2 osd01-3 osd01-4 mon-02 osd02-1 osd02-2 osd02-3 osd02-4 mon-03 osd03-1 osd03-2 osd03-3 osd03-4 Edge switch Core switch Edge switch WAN Client Client client Client Client We will go HERE to disconnect the ports Datacenter MIMOS KULIM Datacenter MIMOS HQ Datacenter MIMOS HQ We will ping Kulim hosts HERE! Owncloud sits HERE!

16 Challenges during POC which running on top of our production network infrastructure. Next, can we set up the distributed storage system with virtual machines plus SDN? – Simulate DFS performance over WAN in a virtualized environment. – Fine-tuning and run experiments: Client’s file-layout, TCP parameters for the network, routing, bandwidth size/throughput, multiple VLANs etc. 16 Moving forward…

17 17


Download ppt "Ceph Distributed File System: Simulating a Site Failure s: {bazli.abkarim, mt.wong, Mohd Bazli Ab Karim, Ming-Tat Wong, Jing-Yuan."

Similar presentations


Ads by Google