RBD Cache Types explanation persistent write log cache and immutable object cache and it's use cases

RBD Cache Types explanation persistent write log cache and immutable object cache and it's use cases

1. RBD Cache Types and It’s use cases IBM/RH Presented By: Sunil Angadi

2. RBD Cache types PWL cache IOC cache Persistent Write Log cache (PWL) Immutable object cache (IOC)

3. Persistent Write Log cache (PWL) Persistence Modes: Persistent-on-Write Mode Persistent-on-Flush Mode Log-Ordered Write-Back: Checkpoints Crash Consistency Fault-Tolerant Data Staleness PMEM or SSDs Disabled by default Write Path : • Application → Write Request → PWL Cache → Flush to Main Storage. Read Path : • Application → Read Request → Check PWL Cache → (If Cached) Serve from PWL Cache → (If Not Cached) Serve from Main Storage.

4. Implementation of PWL 1. Write Request Initiation 2. Acquiring Exclusive Lock 3. Writing to the Cache 4. Logging and Checkpointing 5. Write Confirmation (Persistent-on-Write Mode) 6. Write Completion (Persistent-on-Flush Mode) 7. Flushing Data to Main Storage 8. Data Persistence Verification 9. Releasing Exclusive Lock

5. Configuration of the PWL cache https://privatebin.corp.redhat.com/?cd2feffddd5523f1#97RX1wJb5UViXWoNRFXJZnPFv2m9KmoHbXxqm16oLdRM [ceph: root@tala014 /]# rbd status pool3/test_image1 Watchers: watcher=10.8.131.20:0/2745637112 client.73277 cookie=140236654977664 Persistent cache state: host: tala020 path: /mnt/IrybnSzqgc/rbd-pwl.pool3.f6e8fe6dfbd2.pool size: 1 GiB mode: ssd stats_timestamp: Tue Nov 28 09:37:54 2023 present: true empty: false clean: false allocated: 410 MiB cached: 386 MiB dirty: 43 MiB free: 614 MiB hits_full: 23757 / 24% hits_partial: 0 / 0% misses: 75109 hit_bytes: 93 MiB / 24% miss_bytes: 293 MiB

6. PWL Use Cases Enhanced Performance for Write-Intensive Workloads Reduces Latency for Virtual Machine (VM) Operations Data Integrity and Crash Consistency for Critical Applications Efficient Use of Persistent Memory (PMEM) or High-Speed SSDs Cost-Effective Performance Scaling

7. Immutable object cache (IOC) Local Disk Caching Local Caching Directory Optimized Future Reads

8. Components of the IOC daemon Domain socket based inter-process communication (IPC) Least recently used (LRU) based promotion or demotion policy File-based caching store

9. Configure the ceph-immutable-object-cache http://magna002.ceph.redhat.com/cephci-jenkins/cephci-run-EB52Z8/ [root@plena007 ceph-immutable-object-cache]# ls 0 12 16 2 23 27 30 34 38 41 45 49 52 56 6 63 67 70 74 78 81 85 89 92 96 1 13 17 20 24 28 31 35 39 42 46 5 53 57 60 64 68 71 75 79 82 86 9 93 97 10 14 18 21 25 29 32 36 4 43 47 50 54 58 61 65 69 72 76 8 83 87 90 94 98 11 15 19 22 26 3 33 37 40 44 48 51 55 59 62 66 7 73 77 80 84 88 91 95 99 [[root@plena007 0]# ls :51:204:rbd_data.180684f5e5ab8a.0000000000000112 :51:204:rbd_data.180684f5e5ab8a.0000000000000248 :51:204:rbd_data.180684f5e5ab8a.000000000000018a :51:204:rbd_data.180684f5e5ab8a.00000000000002f0 :51:204:rbd_data.180684f5e5ab8a.0000000000000197 :51:204:rbd_data.180684f5e5ab8a.00000000000003af :51:204:rbd_data.180684f5e5ab8a.00000000000001f1 :51:204:rbd_data.180684f5e5ab8a.00000000000003da :51:204:rbd_data.180684f5e5ab8a.00000000000001f7 :51:204:rbd_data.180684f5e5ab8a.00000000000003dd

10. Immutable object cache use cases Virtual Desktop Infrastructure (VDI) Deployments Cloud Service Providers Offering Virtual Machines DevOps and Continuous Integration/Continuous Deployment (CI/CD) Pipelines High-Performance Computing (HPC) Environments Database Cloning for Testing and Development

11. References PWL: https://docs.redhat.com/en/documentation/red_hat_ceph_storage/7/html-single/block_device_guide/index#configuring-the-rbdmap-service_block IOC: https://docs.redhat.com/en/documentation/red_hat_ceph_storage/7/html-single/block_device_guide/index#management-of-ceph-immutable-object-cache-daemons https://docs.ceph.com/en/latest/rbd/rbd-persistent-read-only-cache/

12. Q&A Thank you!