zhangzf875
zhangzf875
快照无法恢复
我有一个集合已经删除了,但是有做快照,但是进行快照的时候却报错误: { "method": null, "endpoint": null, "reqBody": " {\n \"location\": \"http:\n}\n", "error": "Fix the Position brackets to run & check the json" } 具体恢复快照的语句是: PUT /collections/69169_bak/snapshots/recover { "location": "http://127.0.0.1:6333/collections/69169_bak/snapshots/69169_bak-8430046310527005-2023-12-08-01-41-03.shapshot" }
I have 40 collections on qdrant installed with version 1.7.2, and the memory of the server is 5G. I will optimize the collection through the official document and update the...
Why does the data file in qdrant take up very little space, and the snapshot file takes up so much space, and each snapshot file is about 65MB, what is...
## 错误报告 **Datart版本号** 1.0.0.rc.1 **错误描述** datart上面所有的数据都无法正常查看,出现提示:wait millis 100000, active 0, maxActive 100, creating 0, createErrorCount 1 然后需要重启才能正常查看看板数据 在里面报错日志基本都是:Caused by: java.sql.SQLSyntaxErrorException: Backend not found. Check if any backend is down or...
因为想要给qdrant的集合备份,然后将备份数据传送至,另外的地方进行存储,但是目前就只有建立快照的方式去备份集合数据,导致服务器的磁盘空间很容易被快照文件撑爆,请问一下有没有好的备份方式进行备份。 如果进行全量备份是不是只要将文件qdrant_storage备份下来即可, 但是如果以集合为单位进行备份有没有比较好的方法,或者是可以直接将快照文件传送别的地方,但是又怕想要恢复的时候找不到对应的快照路径,导致无法恢复数据,请大佬帮忙看一下
我在linux上面部署了qdrant1.7.2,使用一段时间后,发现内存随着集合和点数在增加,内存也在一点点增加;后面我将索引的配置修改为存储在磁盘中,发现内存由22G降到了11.7G,现在又将内存配置由存储在磁盘中,改为到存储在内存中,发现内存没有从11.7G变回到22G左右,这是什么原因导致的 修改的方式: curl -X PATCH http://localhost:6333/collections/${collections} \ -H 'Content-Type: application/json' \ --header 'api-key: ' \ -d '{ "hnsw_config": { "on_disk": false } }'
### Current Behavior When milvus-backup is installed, run. /milvus-backup server to start the service. Then run the. /milvus-backup check command. An error is detected: 0.4.24 (Built on 2024-09-20T02:08:15Z from Git...
Because each user created a collection separately before, the number of collections has greatly increased to more than 5000. Now the system memory can no longer meet the current demand....