|
| 1 | +# Multipeer Tests |
| 2 | + |
| 3 | +This document describes the multipeer replication tests for Couchbase Lite, testing multipeer replication functionality with multiple devices. |
| 4 | + |
| 5 | +## test_large_mesh_sanity |
| 6 | + |
| 7 | +Test basic multipeer mesh replication sanity with a small number of documents. |
| 8 | + |
| 9 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 10 | +2. Reset local database and load `empty` dataset on all devices. |
| 11 | +3. Add 20 documents to the database on device 1: |
| 12 | + * Documents named: `doc1`, `doc2`, ..., `doc20` |
| 13 | + * Each document contains a random integer value |
| 14 | +4. Start multipeer replicator on all devices: |
| 15 | + * peer_id: `"mesh-test"` |
| 16 | + * database: `db1` (on each device) |
| 17 | + * collections: `["_default._default"]` |
| 18 | +5. Wait for idle status on all devices: |
| 19 | + * timeout: 60 seconds |
| 20 | + * Verify no replicator errors |
| 21 | +6. Check that all device databases have the same content: |
| 22 | + * All databases should have identical documents |
| 23 | +7. Stop multipeer replicator on all devices. |
| 24 | + |
| 25 | +## test_large_mesh_consistency |
| 26 | + |
| 27 | +Test multipeer mesh replication consistency with documents added on all devices. |
| 28 | + |
| 29 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 30 | +2. Reset local database and load `empty` dataset on all devices. |
| 31 | +3. Add 50 documents to the database on all devices: |
| 32 | + * Each device adds documents with naming pattern: `device{device_idx}-doc{doc_num}` |
| 33 | + * Each document contains a random integer value |
| 34 | + * Total documents across all devices: 50 * number_of_devices |
| 35 | +4. Start multipeer replicator on all devices: |
| 36 | + * peer_id: `"mesh-test"` |
| 37 | + * database: `db1` (on each device) |
| 38 | + * collections: `["_default._default"]` |
| 39 | +5. Wait for idle status on all devices: |
| 40 | + * timeout: 300 seconds |
| 41 | + * Verify no replicator errors |
| 42 | +6. Check that all device databases have the same content: |
| 43 | + * All databases should have identical documents from all devices |
| 44 | +7. Stop multipeer replicator on all devices. |
| 45 | + |
| 46 | +## test_scalable_conflict_resolution |
| 47 | + |
| 48 | +Test scalable conflict resolution using merge conflict resolver in multipeer mesh replication. |
| 49 | + |
| 50 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 51 | +2. Reset local database and load `empty` dataset on all devices. |
| 52 | +3. Insert conflict1 on each device with its unique key in 'counter': |
| 53 | + * Document ID: `conflict1` |
| 54 | + * Format: `{"counter": {"deviceX": X}}` |
| 55 | + * Each device adds its own device key with device number as value |
| 56 | +4. Start multipeer replication with merge conflict resolver: |
| 57 | + * peer_id: `"couchtest"` |
| 58 | + * database: `db1` (on each device) |
| 59 | + * collections: `["_default._default"]` |
| 60 | + * conflict_resolver: `merge-dict` with property `counter` |
| 61 | +5. Wait for idle status on all devices: |
| 62 | + * timeout: 60 seconds |
| 63 | + * Verify no replicator errors |
| 64 | +6. Verify conflict1 is resolved identically on all devices with all device keys: |
| 65 | + * All devices should have the same document with all device keys merged |
| 66 | + * Each key's value must match the device_id |
| 67 | + * Retry up to 5 times if revision IDs don't match initially |
| 68 | +7. Stop multipeer replicator on all devices. |
| 69 | + |
| 70 | +## test_network_partition |
| 71 | + |
| 72 | +Test network partition scenarios with peer groups in multipeer mesh replication. |
| 73 | + |
| 74 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 75 | +2. Reset local database and load `empty` dataset on all devices. |
| 76 | +3. Verify we have 6-15 devices for this test. |
| 77 | +4. Dynamically split devices into 3 groups: |
| 78 | + * Group 1: first portion of devices |
| 79 | + * Group 2: middle portion of devices |
| 80 | + * Group 3: remaining devices |
| 81 | +5. Add unique documents to each peer group: |
| 82 | + * Group 1: documents named `group1-doc1`, `group1-doc2`, ... (100 docs total) |
| 83 | + * Group 2: documents named `group2-doc1`, `group2-doc2`, ... (100 docs total) |
| 84 | + * Group 3: documents named `group3-doc1`, `group3-doc2`, ... (100 docs total) |
| 85 | + * Documents distributed across devices in each group |
| 86 | +6. Start multipeer replicators with different peer groups: |
| 87 | + * Group 1 replicators: peer_id `"group1"` |
| 88 | + * Group 2 replicators: peer_id `"group2"` |
| 89 | + * Group 3 replicators: peer_id `"group3"` |
| 90 | + * database: `db1` (on each device) |
| 91 | + * collections: `["_default._default"]` |
| 92 | +7. Wait for initial replication within each group: |
| 93 | + * timeout: 300 seconds |
| 94 | + * Verify no replicator errors |
| 95 | +8. Verify groups are isolated from each other: |
| 96 | + * Each group should only see peers within its own group |
| 97 | + * Group 1 should only see Group 1 peers |
| 98 | + * Group 2 should only see Group 2 peers |
| 99 | + * Group 3 should only see Group 3 peers |
| 100 | +9. Verify each group can see its own documents: |
| 101 | + * Group 1 devices should have 100 documents (group1-doc*) |
| 102 | + * Group 2 devices should have 100 documents (group2-doc*) |
| 103 | + * Group 3 devices should have 100 documents (group3-doc*) |
| 104 | +10. Stop group 2 replicators and restart with group 1's peer ID: |
| 105 | + * Stop all group 2 replicators |
| 106 | + * Restart group 2 replicators with peer_id `"group1"` |
| 107 | +11. Wait for replication between group 1 and group 2: |
| 108 | + * timeout: 300 seconds |
| 109 | + * Verify no replicator errors |
| 110 | +12. Verify group 1 and group 2 devices have combined documents: |
| 111 | + * Combined group should have 200 documents (100 from group1 + 100 from group2) |
| 112 | +13. Stop group 3 replicators and restart with group 1's peer ID: |
| 113 | + * Stop all group 3 replicators |
| 114 | + * Restart group 3 replicators with peer_id `"group1"` |
| 115 | +14. Wait for replication across all groups: |
| 116 | + * timeout: 300 seconds |
| 117 | + * Verify no replicator errors |
| 118 | +15. Verify all devices have all documents: |
| 119 | + * All devices should have 300 documents total (100 from each group) |
| 120 | + * All databases should have identical content |
| 121 | +16. Stop all multipeer replicators. |
| 122 | + |
| 123 | +## test_dynamic_peer_addition_removal |
| 124 | + |
| 125 | +Test dynamic peer addition and removal during active replication in multipeer mesh. |
| 126 | + |
| 127 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 128 | +2. Reset local database and load `empty` dataset on all devices. |
| 129 | +3. Verify we have at least 6 devices for this test. |
| 130 | +4. Calculate device distribution: |
| 131 | + * Initial devices: at least 3, up to half of total devices |
| 132 | + * Additional devices: up to 3 additional devices |
| 133 | + * Devices to remove: 1-2 devices |
| 134 | +5. Add documents to initial devices: |
| 135 | + * Each device adds 20 documents |
| 136 | + * Documents named: `device{device_idx}-doc{doc_num}` |
| 137 | +6. Start multipeer replicator on initial devices: |
| 138 | + * peer_id: `"dynamic-mesh"` |
| 139 | + * database: `db1` (on each device) |
| 140 | + * collections: `["_default._default"]` |
| 141 | +7. Wait for some initial replication progress: |
| 142 | + * Wait for at least half of initial devices to reach idle |
| 143 | + * timeout: 300 seconds |
| 144 | + * Verify no replicator errors |
| 145 | +8. Add additional devices to the mesh: |
| 146 | + * Add 20 documents to each additional device |
| 147 | + * Documents named: `device{device_idx}-doc{doc_num}` |
| 148 | +9. Start replicators on additional devices: |
| 149 | + * peer_id: `"dynamic-mesh"` (same as initial devices) |
| 150 | + * database: `db1` (on each device) |
| 151 | + * collections: `["_default._default"]` |
| 152 | +10. Wait a short time for replication to start (2 seconds). |
| 153 | +11. Remove random devices from the mesh while replication is active: |
| 154 | + * Randomly select 1-2 devices to remove (mix of initial and additional) |
| 155 | + * Stop replicators on selected devices |
| 156 | +12. Wait for remaining devices to stabilize after removal: |
| 157 | + * timeout: 300 seconds |
| 158 | + * Verify no replicator errors |
| 159 | +13. Verify remaining devices achieve full data consistency: |
| 160 | + * All remaining devices should have the same document count |
| 161 | + * All databases should have identical content |
| 162 | +14. Stop all remaining multipeer replicators. |
| 163 | + |
| 164 | +## test_large_document_replication |
| 165 | + |
| 166 | +Test multipeer mesh replication with large documents containing blobs. |
| 167 | + |
| 168 | +1. Verify CBL version >= 3.3.0 on all test servers. |
| 169 | +2. Reset local database and load `empty` dataset on all devices. |
| 170 | +3. Add 10 large documents with xl1.jpg blob to the database on device 1: |
| 171 | + * Documents named: `large_doc1`, `large_doc2`, ..., `large_doc10` |
| 172 | + * Each document contains a blob: `{"image": "xl1.jpg"}` |
| 173 | +4. Start multipeer replicator on all devices: |
| 174 | + * peer_id: `"large-doc-mesh"` |
| 175 | + * database: `db1` (on each device) |
| 176 | + * collections: `["_default._default"]` |
| 177 | +5. Wait for idle status on all devices: |
| 178 | + * timeout: 200 seconds |
| 179 | + * Verify no replicator errors |
| 180 | +6. Check that all device databases have the same content: |
| 181 | + * Each device should have exactly 10 documents |
| 182 | + * All databases should have identical content including blobs |
| 183 | +7. Stop multipeer replicator on all devices. |
0 commit comments