fix: link and error

minerdao · Jul 27, 2023 · 90d30a8 · 90d30a8
1 parent 9a38a03
commit 90d30a8
Show file tree

Hide file tree

Showing 33 changed files with 56 additions and 34 deletions.
diff --git a/README.md b/README.md
@@ -11,29 +11,29 @@ MinerDAO是面向中文社区的挖矿技术交流学习组织，我们尝试以
 
 ## 技术文章
 #### Spacemesh
-- [SpaceMesh 集群P盘教程](./posts/spacemesh/cluster-mining-tutorial.md) - 如何使用postcli搭建集群P盘的详细教程
+- [SpaceMesh 集群P盘教程](./spacemesh/cluster-mining-tutorial.md) - 如何使用postcli搭建集群P盘的详细教程
 
 #### Filecoin
-- [Filecoin封装流程详解](./posts/filecoin/lotus-mining-process.md)
-- [Filecoin挖矿硬件推荐配置](./posts/filecoin/hardware-configuration.md)
-- [扇区生命周期详解，扇区状态管理](./posts/filecoin/sector-life-cycle.md)
-- [Filecoin节点搭建及启动](./posts/filecoin/daemon-deployment.md)
-- [节点管理，节点常用操作，节点同步常见问题](./posts/filecoin/daemon-operation.md)
-- [钱包管理及使用场景](./posts/filecoin/wallet-management.md)
-- [Lotus-ops部署运维工具使用说明](./posts/filecoin/ansible-deploy-tool-usage.md)
-- [Filecoin同构集群搭建流程](./posts/filecoin/mining-cluster-deployment.md)
-- [Filecoin存储管理，存储i/o性能分析，网络分析](./posts/filecoin/storage-manage.md)
-- [新矿工节点上线CheckList](./posts/filecoin/new-miner-checklist.md)
-- [Filecoin监控报警系统搭建](./posts/filecoin/monitoring-deployment.md)
-- [Filecoin 订单配置](./posts/filecoin/deals-configuration.md)
-- [分布式Miner模块配置和使用](./posts/filecoin/distributed-miner-configuration.md)
-- [如何避免掉算力](./posts/filecoin/miner-keep.md)
-- [Lotus日常运维CheckList](./posts/filecoin/lotus-ops-checklist.md)
-- [常用环境变量说明](./posts/filecoin/environment-usage.md)
-- [自动Pledge脚本使用说明](./posts/filecoin/auto-pledge.md)
-- [常见运维问题定位与解决](./posts/filecoin/questions.md#1-顽固扇区如何删除) - 顽固扇区如何删除, Worker掉线如何处理
-- [Filecoin Datacap申请](./posts/filecoin/lotus-datacap-application.md) - Datacap如何申请及注意事项
-- [Snap Deals操作详解](./posts/filecoin/snap-deals.md) - Snap Deals操作流程及遇到的问题
+- [Filecoin封装流程详解](./filecoin/lotus-mining-process.md)
+- [Filecoin挖矿硬件推荐配置](./filecoin/hardware-configuration.md)
+- [扇区生命周期详解，扇区状态管理](./filecoin/sector-life-cycle.md)
+- [Filecoin节点搭建及启动](./filecoin/daemon-deployment.md)
+- [节点管理，节点常用操作，节点同步常见问题](./filecoin/daemon-operation.md)
+- [钱包管理及使用场景](./filecoin/wallet-management.md)
+- [Lotus-ops部署运维工具使用说明](./filecoin/ansible-deploy-tool-usage.md)
+- [Filecoin同构集群搭建流程](./filecoin/mining-cluster-deployment.md)
+- [Filecoin存储管理，存储i/o性能分析，网络分析](./filecoin/storage-manage.md)
+- [新矿工节点上线CheckList](./filecoin/new-miner-checklist.md)
+- [Filecoin监控报警系统搭建](./filecoin/monitoring-deployment.md)
+- [Filecoin 订单配置](./filecoin/deals-configuration.md)
+- [分布式Miner模块配置和使用](./filecoin/distributed-miner-configuration.md)
+- [如何避免掉算力](./filecoin/miner-keep.md)
+- [Lotus日常运维CheckList](./filecoin/lotus-ops-checklist.md)
+- [常用环境变量说明](./filecoin/environment-usage.md)
+- [自动Pledge脚本使用说明](./filecoin/auto-pledge.md)
+- [常见运维问题定位与解决](./filecoin/questions.md#1-顽固扇区如何删除) - 顽固扇区如何删除, Worker掉线如何处理
+- [Filecoin Datacap申请](./filecoin/lotus-datacap-application.md) - Datacap如何申请及注意事项
+- [Snap Deals操作详解](./filecoin/snap-deals.md) - Snap Deals操作流程及遇到的问题
 - Filecoin 订单市场操作详解 - 分离订单市场、离线订单 [TODO]
 - FVM合约开发及部署 [TODO]
 

diff --git a/posts/aleo/.gitkeep → aleo/.gitkeep b/posts/aleo/.gitkeep → aleo/.gitkeep
diff --git a/posts/aleo/how-testnet3-start.md → aleo/how-testnet3-start.md b/posts/aleo/how-testnet3-start.md → aleo/how-testnet3-start.md
diff --git a/posts/filecoin/ansible-deploy-tool-usage.md → filecoin/ansible-deploy-tool-usage.md b/posts/filecoin/ansible-deploy-tool-usage.md → filecoin/ansible-deploy-tool-usage.md
diff --git a/posts/filecoin/auto-pledge.md → filecoin/auto-pledge.md b/posts/filecoin/auto-pledge.md → filecoin/auto-pledge.md
diff --git a/posts/filecoin/daemon-deployment.md → filecoin/daemon-deployment.md b/posts/filecoin/daemon-deployment.md → filecoin/daemon-deployment.md
diff --git a/posts/filecoin/daemon-operation.md → filecoin/daemon-operation.md b/posts/filecoin/daemon-operation.md → filecoin/daemon-operation.md
diff --git a/posts/filecoin/datacap-process.md → filecoin/datacap-process.md b/posts/filecoin/datacap-process.md → filecoin/datacap-process.md
diff --git a/posts/filecoin/deals-configuration.md → filecoin/deals-configuration.md b/posts/filecoin/deals-configuration.md → filecoin/deals-configuration.md
diff --git a/...lecoin/distributed-miner-configuration.md → filecoin/distributed-miner-configuration.md b/...lecoin/distributed-miner-configuration.md → filecoin/distributed-miner-configuration.md
diff --git a/posts/filecoin/environment-usage.md → filecoin/environment-usage.md b/posts/filecoin/environment-usage.md → filecoin/environment-usage.md
diff --git a/posts/filecoin/hardware-configuration.md → filecoin/hardware-configuration.md b/posts/filecoin/hardware-configuration.md → filecoin/hardware-configuration.md
diff --git a/posts/filecoin/lotus-datacap-application.md → filecoin/lotus-datacap-application.md b/posts/filecoin/lotus-datacap-application.md → filecoin/lotus-datacap-application.md
diff --git a/posts/filecoin/lotus-mining-process.md → filecoin/lotus-mining-process.md b/posts/filecoin/lotus-mining-process.md → filecoin/lotus-mining-process.md
diff --git a/posts/filecoin/lotus-multisig.md → filecoin/lotus-multisig.md b/posts/filecoin/lotus-multisig.md → filecoin/lotus-multisig.md
diff --git a/posts/filecoin/lotus-ops-checklist.md → filecoin/lotus-ops-checklist.md b/posts/filecoin/lotus-ops-checklist.md → filecoin/lotus-ops-checklist.md
diff --git a/posts/filecoin/lotus-upgrade.md → filecoin/lotus-upgrade.md b/posts/filecoin/lotus-upgrade.md → filecoin/lotus-upgrade.md
diff --git a/posts/filecoin/miner-keep.md → filecoin/miner-keep.md b/posts/filecoin/miner-keep.md → filecoin/miner-keep.md
diff --git a/posts/filecoin/mining-cluster-deployment.md → filecoin/mining-cluster-deployment.md b/posts/filecoin/mining-cluster-deployment.md → filecoin/mining-cluster-deployment.md
diff --git a/posts/filecoin/monitoring-deployment.md → filecoin/monitoring-deployment.md b/posts/filecoin/monitoring-deployment.md → filecoin/monitoring-deployment.md
diff --git a/posts/filecoin/myscheduler-configuration.md → filecoin/myscheduler-configuration.md b/posts/filecoin/myscheduler-configuration.md → filecoin/myscheduler-configuration.md
diff --git a/posts/filecoin/new-miner-checklist.md → filecoin/new-miner-checklist.md b/posts/filecoin/new-miner-checklist.md → filecoin/new-miner-checklist.md
diff --git a/posts/filecoin/questions.md → filecoin/questions.md b/posts/filecoin/questions.md → filecoin/questions.md
diff --git a/posts/filecoin/raid + xfs .md → filecoin/raid + xfs .md b/posts/filecoin/raid + xfs .md → filecoin/raid + xfs .md
diff --git a/posts/filecoin/sector-life-cycle.md → filecoin/sector-life-cycle.md b/posts/filecoin/sector-life-cycle.md → filecoin/sector-life-cycle.md
diff --git a/posts/filecoin/snap-deals.md → filecoin/snap-deals.md b/posts/filecoin/snap-deals.md → filecoin/snap-deals.md
diff --git a/posts/filecoin/storage-manage.md → filecoin/storage-manage.md b/posts/filecoin/storage-manage.md → filecoin/storage-manage.md
diff --git a/posts/filecoin/wallet-management.md → filecoin/wallet-management.md b/posts/filecoin/wallet-management.md → filecoin/wallet-management.md
diff --git a/posts/phala/cluster-architecture.md → phala/cluster-architecture.md b/posts/phala/cluster-architecture.md → phala/cluster-architecture.md
diff --git a/posts/phala/cluster-deployment.md → phala/cluster-deployment.md b/posts/phala/cluster-deployment.md → phala/cluster-deployment.md
diff --git a/posts/phala/phala-ops.md → phala/phala-ops.md b/posts/phala/phala-ops.md → phala/phala-ops.md
diff --git a/posts/phala/problem.md → phala/problem.md b/posts/phala/problem.md → phala/problem.md
diff --git a/posts/spacemesh/cluster-mining-tutorial.md → spacemesh/cluster-mining-tutorial.md b/posts/spacemesh/cluster-mining-tutorial.md → spacemesh/cluster-mining-tutorial.md
@@ -37,7 +37,15 @@ make build
 smcli也可以直接去官方下载编译好的：https://github.com/spacemeshos/smcli/releases。
 
 ## 2 集群P盘流程
-这里主要分享多机器集群P盘的详细操作。简单来说，先初始化，然后用postcli算好需要P盘的文件数，然后计算index，再分配到集群的多台机器开始P盘，最终全部P完后，再合并到一起，然后由`go-spacemesh`来扫盘并生成证明。
+这里主要分享多机器集群P盘的详细操作流程，主要包括以下步骤：
+
+- 使用`smcli`创建钱包并备份助记词和钱包密码；
+- 使用编译好的go-spacemesh进行初始化；
+- 使用postcli的 `-printNumFiles` 计算生成的`.bin`文件数量；
+- 计算分段索引，将文件分段并分发给多台机器；
+- 按照计算好的分段索引在每台机器上启动P盘；
+- 等所有机器P盘完成后，将生成的文件合并到运行go-spacemesh的机器上；
+- 重新启动go-spacemesh，开始扫盘并生成证明。
 
 ### 2.1 创建钱包
 使用上面编译好的smcli来创建钱包，注意要备份好助记词和钱包密码。
@@ -46,15 +54,15 @@ smcli也可以直接去官方下载编译好的：https://github.com/spacemeshos
 ```
 钱包创建完毕后，会输出到一个`/path/to/wallet_2023-07-21T07-17-52.946Z.json`文件中。
 
-下一步执行导入钱包，并输入创建时的密码，就能看到钱包地址（格式为：sm1xxxxxxxxx）。
+下一步执行导入钱包，并输入创建时的密码，就能看到刚才创建的钱包地址（格式为：sm1xxxxxxxxx）。
 ```shell
 ./smcli wallet read /path/to/wallet_2023-07-21T07-17-52.946Z.json
 ```
 
 ### 2.2 初始化
 先用编译好的`go-spacemesh`初始化：
 ```shell
-./go-spacemesh --config config.mainnet.json --smeshing-coinbase sm1xxxxxxx --smeshing-opts-numunits 5 --smeshing-opts-datadir /mnt/spacemesh/post_data --data-folder ~/md0/spacemesh/node_data
+./go-spacemesh --config config.mainnet.json --smeshing-coinbase sm1xxxxxxx --smeshing-opts-numunits 5 --smeshing-opts-datadir /mnt/spacemesh/post_data --data-folder /mnt/spacemesh/node_data
 ```
 该命令会启动一个Spacemesh节点，并同时启动P盘，参数说明：
 - `--config`: 节点配置文件，通过`wget https://smapp.spacemesh.network/config.mainnet.json`获取；
@@ -94,26 +102,34 @@ SpaceMesh是以`numUnits`为基本的存储单元，每个`numUnits = 64GB`，P
 
 Subset是把要P的文件分段并分发给多台机器来跑，通过`-fromFile`，`-toFile`来设置开始及结束的文件索引。
 
-例如，上面的`-numUnits=5`总共需要生成40个.bin文件，如果平均分配到4台机器上，则每台机器的启动命令分别为：
+例如，上面的`-numUnits=5`总共需要生成80个.bin文件，如果平分给4台机器，则每台机器的`-fromFile`和`-toFile`分别为：
+机器| -fromFile | - toFile
+------|-------:|------:
+机器1 | 0  | 19
+机器2 | 20 | 39
+机器3 | 40 | 59
+机器4 | 60 | 79
+
+启动命令分别为：
 
 - 机器1 (0 - 19)
 ```shell
-./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=0 -toFile=19 -datadir=/mnt/spacemesh/data
+./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=0 -toFile=19 -datadir=/mnt/spacemesh/post_data
 ```
 
 - 机器2 (20 - 39)
 ```shell
-./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=20 -toFile=39 -datadir=/mnt/spacemesh/data
+./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=20 -toFile=39 -datadir=/mnt/spacemesh/post_data
 ```
 
 - 机器3 (40 - 59)
 ```shell
-./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=40 -toFile=59 -datadir=/mnt/spacemesh/data
+./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=40 -toFile=59 -datadir=/mnt/spacemesh/post_data
 ```
 
 - 机器4 (60 - 79)
 ```shell
-./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=60 -toFile=79 -datadir=/mnt/spacemesh/data
+./postcli -provider=0 -commitmentAtxId=9eebff023abb17ccb775c602daade8ed708f0a50d3149a42801184f5b74f2865 -id=[hex_decoded_id] -numUnits=5 -fromFile=60 -toFile=79 -datadir=/mnt/spacemesh/post_data
 ```
 
 ### 2.5 启动P盘
@@ -129,15 +145,19 @@ Subset是把要P的文件分段并分发给多台机器来跑，通过`-fromFile
 - `-datadir` P盘完成后的文件保存目录。
 
 ### 2.6 合并P盘文件
-等所有subset的机器P盘完成后，需要将每台机器上生成的文件，合并到运行`go-spacemesh`服务机器的`--smeshing-opts-datadir`路径下，然后重新启动`go-spacemesh`就可以开始扫盘并生成证明了。如果文件完整，就会输出类似下面的日志：
+等所有subset的机器P盘完成后，需要将每台机器上生成的文件，合并到运行`go-spacemesh`服务机器的`--smeshing-opts-datadir`路径下，然后重新启动`go-spacemesh`就可以开始扫盘并生成证明了。
+
+**注意**`.bin`文件合并后，需要在每个运行postcli机器的`postdata_metadata.json`文件里，找到一个全局最小的nonce，作为`go-spacemesh`的`--smeshing-opts-datadir`路径下`postdata_metadata.json`的nonce （有的postcli进程可能找不到nonce）。
+
+启动后如果文件完整，扫码文件时会输出类似下面的日志：
 
 ```shell
 2023-07-23T15:57:24.849+0800	INFO	fb26a.post	initialization: file already initialized	{"node_id": "fb26a9d2da5626ded24027da14054bf0fbf8886bd7ec4a29d05ee2fdd44edddd", "module": "post", "fileIndex": 0, "currentNumLabels": 268435456, "targetNumLabels": 268435456, "startPosition": 0}
 2023-07-23T15:57:24.849+0800	INFO	fb26a.post	initialization: file already initialized	{"node_id": "fb26a9d2da5626ded24027da14054bf0fbf8886bd7ec4a29d05ee2fdd44edddd", "module": "post", "fileIndex": 1, "currentNumLabels": 268435456, "targetNumLabels": 268435456, "startPosition": 268435456}
 ...
 ```
 
-### 2.7 扫盘及提交证明
+### 2.7 扫盘并提交证明
 文件扫描完毕后，开始读取文件并生成证明，生成证明开始的日志类似：
 ```shell
 2023-07-23T15:57:59.995+0800	INFO	fb26a.post calculating proof of work for nonces 0..144
@@ -146,11 +166,13 @@ Subset是把要P的文件分段并分发给多台机器来跑，通过`-fromFile
 扫盘持续时间根据机器配置和磁盘速率而定，可通过每隔5秒钟输出磁盘I/O：
 `iostat -dmt /dev/md0 5`
 
-据我个人的测试观察，整个扫盘并生成证明的过程，在刚开始会有10-20分钟的准备时间(20T数据)，然后开始有磁盘I/O。
+据我个人测试观察，整个扫盘并生成证明的过程，在刚开始会有10-20分钟的准备时间(20T数据)，然后开始有磁盘I/O。
+
+**扫盘时间 = (总容量 / 磁盘读速率) + 30分钟的准备时间**
 
-可以根据自己的磁盘速率，估算扫盘及生成证明的时间。
+可根据自己的磁盘速率，估算扫盘及生成证明的时间。
 
-比如磁盘读写为2G/s，那么20T数据的扫盘+证明时间大致为：`(20 * 1024) / 2 /3600`约为2.8小时，加上准备时间，大约为3个小时。
+比如磁盘读写为2G/s，那么20T数据的扫盘+证明时间大致为：`(20 * 1024) / 2 /3600`约为2.8小时，加上准备时间，大约3小时。
 
 扫盘完成后，会输出类似下面的日志：
 ```shell