| --- |
| title: Usage and Performance Notes |
| --- |
| |
| <!-- |
| Licensed to the Apache Software Foundation (ASF) under one or more |
| contributor license agreements. See the NOTICE file distributed with |
| this work for additional information regarding copyright ownership. |
| The ASF licenses this file to You under the Apache License, Version 2.0 |
| (the "License"); you may not use this file except in compliance with |
| the License. You may obtain a copy of the License at |
| |
| http://www.apache.org/licenses/LICENSE-2.0 |
| |
| Unless required by applicable law or agreed to in writing, software |
| distributed under the License is distributed on an "AS IS" BASIS, |
| WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| See the License for the specific language governing permissions and |
| limitations under the License. |
| --> |
| |
| Optimize the cache and region snapshot feature by understanding how it performs. |
| |
| ## <a id="concept_49775DA33EFF4534950E1A60158F4B36__section_D1AB128682CA41D2BEFB429BED892ABE" class="no-quick-link"></a>Cache Consistency and Concurrent Operations |
| |
| Importing and exporting region data is an administrative operation, and certain simultaneous runtime conditions can cause the import or export operation to fail such as when you are rebalancing partitioned region buckets or experience a network partition event. This behavior is expected, and you should retry the operation. Redoing an export overwrites an incomplete snapshot file, and redoing an import updates partially imported data. |
| |
| The snapshot feature does not guarantee consistency. Concurrent cache operations during a snapshot import or export can cause data consistency issues. If snapshot consistency is important, we recommend that you take your application offline before export and import, to provide a quiet period ensures data consistency in your snapshot. |
| |
| For example, modifications to region entries during an export can result in a snapshot that contains some but not all updates. If entries { A, B } are updated to { A', B'} during the export, the snapshot can contain { A, B' } depending on the write order. Also, modifications to region entries during an import can cause lost updates in the cache. If the region contains entries { A, B } and the snapshot contains { A', B' }, concurrent updates { A\*, B\* } can result in the region containing { A\*, B' } after the import completes. |
| |
| The default behavior is to perform all I/O operations on the node where the snapshot operations are invoked. This will involve either collecting or dispersing data over the network if the region is a partitioned region. |
| |
| ## <a id="concept_49775DA33EFF4534950E1A60158F4B36__section_2C90FE1552D94AD08D1E5F3035153E0E" class="no-quick-link"></a>Performance Considerations |
| |
| When using the data snapshot feature, be aware of the following performance considerations: |
| |
| - Importing and exporting cache or region snapshots causes additional CPU and network load. You may need to increase CPU capacity or network bandwidth depending on your applications and infrastructure. In addition, if you export regions that have been configured to overflow to disk, you may require additional disk I/O to perform the export. |
| - When exporting partitioned region data, allocate additional heap memory so the member performing the export can buffer data gathered from other cache members. Allocate at least 10MB per member to your heap in addition to whatever configuration is necessary to support your application or cache. |
| |
