新集群3.3是和旧集群2.6配置一样，节点数保持一直
通过快照备份迁移恢复的方式，迁移数据
快照数据拷贝到新集群 一启动就给删了
目前只有工具一种方式。。。请严格按照文档操作，不要跳过任何步骤。
1 个赞
不同的版本通过快照的方式不一定靠谱。数据格式有可能都变了。
是说2.6 升级到 3.3.0 然后迁移的？还是 2.6 直接迁移？ 后者肯定不行
我的旧集群是2.6，新集群是3.3，现在是用快照的方式直接迁移到新集群，这种方式不行吗？这种方式一启动就给删了
是需要旧集群2.6–快照迁移–新集群2.6–新集群2.6升级3.3–迁移数据 这种方案才行吗？
不行的， 因为大版本之间（2.6到3.0）之间，底层的存储结构变化了（因为有一些优化是要趁着大版本机会做掉的，那时候数据可以不兼容）
那目前有什么方案支持我这种场景吗？
数据不兼容，就只能数据导出，然后导入到新版本？
可以用升级的程序去升级数据文件哈，文档里有的哈~~
这种就可以，如果你新集群已经装好了高版本，你可以把 2.6 搞到一个中间集群2.6上，升级它的数据，再搞到 3.3 的新集群里。
可以试着用二进制包里的 binary upgrade 工具，放到 storaged 里执行，可以在测试环境先演练一下
是。这个没专门验证过，不过理论上是可行的。数据解析的格式改了
1 个赞
在2.6的服务的storaged执行以下命令
./db_upgrader
–src_db_path=/usr/local/nebula/data/storage
–dst_db_path=/usr/local/nebula/data/storage-v3
–upgrade_meta_server=192.168.66.79:49232,192.168.66.79:49233,192.168.66.79:49228
–upgrade_version=2:3
upgrade_meta_server是3.3的
db_upgrader也是3.3的
数据路径：
日志
===========================PARAMS============================
meta server: 192.168.66.79:49232,192.168.66.79:49233,192.168.66.79:49228
source data path: /usr/local/nebula/data/storage
destination data path: /usr/local/nebula/data/storage-v3
The size of the batch written: 100
upgrade data from version: 2:3
whether to compact all data: true
maximum number of concurrent parts allowed:10
maximum number of concurrent spaces allowed: 5
===========================PARAMS============================
I20221220 03:15:17.504247 560 DbUpgraderTool.cpp:112] Prepare phase begin
I20221220 03:15:17.504603 560 MetaClient.cpp:80] Create meta client to "192.168.66.79":49233
I20221220 03:15:17.504608 560 MetaClient.cpp:81] root path: , data path size: 0
E20221220 03:15:18.513824 561 ThriftClientManager-inl.h:70] Failed to resolve address for 'metad1': Name or service not known (error=-2): Unknown error -2
I20221220 03:15:19.533504 560 MetaClient.cpp:3108] Load leader of "storaged0":9779 in 3 space
I20221220 03:15:19.533535 560 MetaClient.cpp:3108] Load leader of "storaged1":9779 in 3 space
I20221220 03:15:19.533538 560 MetaClient.cpp:3108] Load leader of "storaged2":9779 in 3 space
I20221220 03:15:19.533541 560 MetaClient.cpp:3114] Load leader ok
I20221220 03:15:19.536188 560 MetaClient.cpp:162] Register time task for heartbeat!
I20221220 03:15:19.536207 560 DbUpgraderTool.cpp:171] Prepare phase end
I20221220 03:15:19.536209 560 DbUpgraderTool.cpp:174] Upgrade phase begin
I20221220 03:15:19.536991 568 DbUpgraderTool.cpp:185] Upgrade from path /usr/local/nebula/data/storage to path /usr/local/nebula/data/storage-v3 begin
I20221220 03:15:19.537659 568 DbUpgrader.cpp:1169] Upgrade from path /usr/local/nebula/data/storage to path /usr/local/nebula/data/storage-v3 in DbUpgrader run begin
E20221220 03:15:19.537685 568 MetaClient.cpp:1393] Space 806 not found!
E20221220 03:15:19.537737 568 DbUpgrader.cpp:76] Space id 806 no found
E20221220 03:15:19.537744 568 DbUpgrader.cpp:59] Init /usr/local/nebula/data/storage space id 806 failed
W20221220 03:15:19.537756 568 DbUpgrader.cpp:1180] Upgrade from path /usr/local/nebula/data/storage space id 806 to path /usr/local/nebula/data/storage-v3 init failed
W20221220 03:15:19.537766 568 DbUpgrader.cpp:1182] Ignore upgrade /usr/local/nebula/data/storage space id 806
E20221220 03:15:19.537781 568 MetaClient.cpp:1393] Space 0 not found!
E20221220 03:15:19.537787 568 DbUpgrader.cpp:76] Space id 0 no found
E20221220 03:15:19.537799 568 DbUpgrader.cpp:59] Init /usr/local/nebula/data/storage space id 0 failed
W20221220 03:15:19.537806 568 DbUpgrader.cpp:1180] Upgrade from path /usr/local/nebula/data/storage space id 0 to path /usr/local/nebula/data/storage-v3 init failed
W20221220 03:15:19.537812 568 DbUpgrader.cpp:1182] Ignore upgrade /usr/local/nebula/data/storage space id 0
E20221220 03:15:19.537818 568 MetaClient.cpp:1393] Space 796 not found!
E20221220 03:15:19.537827 568 DbUpgrader.cpp:76] Space id 796 no found
E20221220 03:15:19.537834 568 DbUpgrader.cpp:59] Init /usr/local/nebula/data/storage space id 796 failed
W20221220 03:15:19.537840 568 DbUpgrader.cpp:1180] Upgrade from path /usr/local/nebula/data/storage space id 796 to path /usr/local/nebula/data/storage-v3 init failed
W20221220 03:15:19.537845 568 DbUpgrader.cpp:1182] Ignore upgrade /usr/local/nebula/data/storage space id 796
E20221220 03:15:19.537853 568 MetaClient.cpp:1393] Space 609 not found!
E20221220 03:15:19.537860 568 DbUpgrader.cpp:76] Space id 609 no found
E20221220 03:15:19.537866 568 DbUpgrader.cpp:59] Init /usr/local/nebula/data/storage space id 609 failed
W20221220 03:15:19.537878 568 DbUpgrader.cpp:1180] Upgrade from path /usr/local/nebula/data/storage space id 609 to path /usr/local/nebula/data/storage-v3 init failed
W20221220 03:15:19.537883 568 DbUpgrader.cpp:1182] Ignore upgrade /usr/local/nebula/data/storage space id 609
I20221220 03:15:19.537889 568 DbUpgrader.cpp:1196] Max concurrent spaces: 0
I20221220 03:15:19.537894 568 DbUpgrader.cpp:1206] Upgrade from path /usr/local/nebula/data/storage to path /usr/local/nebula/data/storage-v3 in DbUpgrader run end
I20221220 03:15:19.537896 568 DbUpgraderTool.cpp:193] Upgrade from path /usr/local/nebula/data/storage to path /usr/local/nebula/data/storage-v3 end
I20221220 03:15:19.537964 560 DbUpgraderTool.cpp:202] Upgrade phase end
[root@nebula-cluster-storaged-3 bin]#
1 个赞
meta启动了吗？
你正常登录下你的服务，看看这些space还在不在。meta好像都不认为他们存在
–upgrade_meta_server=192.168.66.79:49232,192.168.66.79:49233,192.168.66.79:49228
是指的3.3服务的meta吗？
我是在2.6的环境下进行db_upgrader，2.6的服务都是正常的，space能正常操作
如果我没记错，你需要启动3.X的meta (graph有必要也可以启动 但一定不要启动storage)，然后用3.X的db_upgrader来升级2.6的storage，升级成功才能启动storage