思路SMB 掉速和多方面因素有关：磁盘性能网络吞吐SMB 设置现在将从这三方面介绍如何测试瓶颈，以及提升。一、磁盘性能提供一种最简单的方法，不需要额外工具。条件是：同一个共享目录存在一个大文件（10GB以上为佳）测试方法：在该目录下，将大文件复制一份。原理：同一目录下直接复制，整个复制过程都是在 NAS 中直接完成，也就是说，这个文件不会先下载到 Windows，然后再上传到 NAS。其速度可以反映磁盘整列（或逻辑卷）的最大性能。我的全 SSD 磁盘整列下，该复制速度可以稳定达到 400MB/s （仅供参考）如果你的速度低于 100MB/s 或者拷贝速度不稳定（时快时慢）则需要考虑磁盘性能问题。二、网络吞吐这里需要借助第三方工具iperf3。QNAP的应用商店不直接提供[1]，需要借助第三方商店[2]。主要步骤如下：QNAP 的准备根据你的CPU架构下载iperf3，例如我的 453DMini CPU 是Intel(R) Celeron(R) J4125 CPU ，至高 2700 MHz (4 核心，4 线程))那么就属于 x86_64，下载TS-NASX86_64版本即可。下载地址qnapclub store – iperf3 download登录 NAS，关掉签名验证，安装第三方插件，具体方法参考How to install iperf3 utility in QT ...继续阅读 (17)

新域名启用通知自 2023年3月 起，本博客迁至新域名https://www.lesca.cn与此同时，原域名lesca.me保留使用至 2023年12月。starting Mar 2023, this blog enables the new domain – www.lesca.cn ; The old domain keeps working until Dec 2023. ...继续阅读 (17)

引子Wiki JS 推荐使用 PostgreSQL 作为默认数据库，但是其 PostgreSQL 的词典配置中是没有 Chinese 选项的。由于是第一次接触PostgreSQL，经过一番调查，发现官方的 PostgreSQL 镜像是没有中文断词支持的。为了解决这个问题，需要自己编译中文断词插件，我选择的是 pg_jieba。由于是 Docker 环境，因此还需要生成一个 Docker 镜像。项目和镜像已经开源，请前往我的 GitHub：https://github.com/lesca/postgres-jieba用法1. 配置容器version: "3" services: db: image: lesca/postgres-jieba container_name: wiki_db environment: POSTGRES_DB: wiki POSTGRES_PASSWORD: "xxx" POSTGRES_USER: xxx logging: driver: "none" restart: unless-stopped volumes: - ./dbdata/:/var/lib/postgresql/data wiki: image: requar ...继续阅读 (17)

引子Wiki JS 推荐使用 PostgreSQL 作为默认数据库，但是其 PostgreSQL 的词典配置中是没有 Chinese 选项的。由于是第一次接触PostgreSQL，经过一番调查，发现官方的 PostgreSQL 镜像是没有中文断词支持的。为了解决这个问题，需要自己编译中文断词插件，我选择的是 pg_jieba。由于是 Docker 环境，因此还需要生成一个 Docker 镜像。项目和镜像已经开源，请前往我的 GitHub：https://github.com/lesca/postgres-jieba用法1. 配置容器version: "3" services: db: image: lesca/postgres-jieba container_name: wiki_db environment: POSTGRES_DB: wiki POSTGRES_PASSWORD: "xxx" POSTGRES_USER: xxx logging: driver: "none" restart: unless-stopped volumes: - ./dbdata/:/var/lib/postgresql/data wiki: image: requar ...继续阅读 (14)

/!\ 本文已更新，请前往“火腿百科”查看最新版：*火腿百科–通过 SpyServer 远程连接 AirSpy目标在 NanoPi M4 （树莓派亦可）上连接 AirSpy R2 并通过 SDRSharp 远程访问下载 SPY Server – SDR Server for Linux ARM64https://airspy.com/download/编译https://github.com/airspy/airspyone_host#how-to-build-the-host-software-on-linux# prerequisites sudo apt-get install build-essential cmake libusb-1.0-0-dev pkg-config # Build host software on Linux: wget https://github.com/airspy/airspyone_host/archive/master.zip unzip master.zip cd airspyone_host-master mkdir build cd build cmake ../ -DINSTALL_UDEV_RULES=ON make sudo make install sudo ldconfig配置# spyserver.con ...继续阅读 (17)

/!\ 本文已更新，请前往“火腿百科”查看最新版：*火腿百科–通过 SpyServer 远程连接 AirSpy目标在 NanoPi M4 （树莓派亦可）上连接 AirSpy R2 并通过 SDRSharp 远程访问下载 SPY Server – SDR Server for Linux ARM64https://airspy.com/download/编译https://github.com/airspy/airspyone_host#how-to-build-the-host-software-on-linux# prerequisites sudo apt-get install build-essential cmake libusb-1.0-0-dev pkg-config # Build host software on Linux: wget https://github.com/airspy/airspyone_host/archive/master.zip unzip master.zip cd airspyone_host-master mkdir build cd build cmake ../ -DINSTALL_UDEV_RULES=ON make sudo make install sudo ldconfig配置# spyserver.con ...继续阅读 (15)

引言WHfB 数据是用户到设备的1对1绑定关系，如果电脑换用户、或者其他特殊情况，可以使用此方法重置WHfB的绑定，以允许用户重新设置WHfB。如何重置 Windows Hello for Business服务端进入Azure AD – Users – 找到指定用户 – Authentication Methods删除 “Windows Hello for Business” 项如果是 Hybrid Azure AD 的环境，该操作会通过 Azure AD Connector 写回本地用户对象具体是删除msDS-KeyCredentialLink属性。因此，该操作需要等待本地DC的该用户对象修改后才会最终生效。建议等生效后执行客户端操作。客户端打开命令行 cmd (普通用户权限即可)运行如下命令certutil.exe -DeleteHelloContainer注销用户或重启登录 ...继续阅读 (9)

引言WHfB 数据是用户到设备的1对1绑定关系，如果电脑换用户、或者其他特殊情况，可以使用此方法重置WHfB的绑定，以允许用户重新设置WHfB。如何重置 Windows Hello for Business服务端进入Azure AD – Users – 找到指定用户 – Authentication Methods删除 “Windows Hello for Business” 项如果是 Hybrid Azure AD 的环境，该操作会通过 Azure AD Connector 写回本地用户对象具体是删除msDS-KeyCredentialLink属性。因此，该操作需要等待本地DC的该用户对象修改后才会最终生效。建议等生效后执行客户端操作。客户端打开命令行 cmd (普通用户权限即可)运行如下命令certutil.exe -DeleteHelloContainer注销用户或重启登录 ...继续阅读 (9)

引言重置Windows Hello for Business后仍然无法在登录选项中添加、删除指纹时，可以参考本文。Windows Hello for Business 无法删除指纹的解决方法重启进入BIOS删除指纹数据登录Windows – services.msc – 停止 Windows Biometric Service进入C:\Windows\System32\WinBioDatabase，删除指纹数据（可先备份）重启 Windows Biometric Service进入 设置 – 账户 – 登录选项 – 指纹设置此时可能仍然无法删除，先添加指纹，成功后即可删除。 ...继续阅读 (9)

引言重置Windows Hello for Business后仍然无法在登录选项中添加、删除指纹时，可以参考本文。Windows Hello for Business 无法删除指纹的解决方法重启进入BIOS删除指纹数据登录Windows – services.msc – 停止 Windows Biometric Service进入C:\Windows\System32\WinBioDatabase，删除指纹数据（可先备份）重启 Windows Biometric Service进入 设置 – 账户 – 登录选项 – 指纹设置此时可能仍然无法删除，先添加指纹，成功后即可删除。 ...继续阅读 (9)

引言最近，访问外网出现不稳定的频率有所提升，随之而来的是 Windows 10 客户端 和 Office 365 客户端也出现了奇奇怪怪的问题：Windows 10 右下角显示“地球”图标，并提示“无Internet”，但是实际上国内网站都正常访问。Office 365 客户端无法登录，Outlook 无法收发邮件，其登录依赖于上述检测结果Windows 10 是如何判断是否有互联网访问的Windows 10 检测是否有互联网访问，通过一种叫做 NCSI (Network Connection Status Indicator) 的检测机制，当网卡连接后，会发起“主动”检测，检测该URL的可访问性：Win10:http://www.msftconnecttest.com/connecttest.txt如果该文件无法获得，则认定为“无Internet”。此时，如果用户手动访问以下地址，则有机率被检测到，并认为网络恢复。这种方式称为“被动”检测机制。Win10:http://www.msftconnecttest.com/redirect该链接会重定向到MSN网站。原因如果出现“地球”图标后，你可以手动访问这个链接：http://www.msftconnecttest.com/connecttest.txt目前，我的测试结果是，WebKit内核的浏览器，大约5次中有一次可以访问；IE浏 ...继续阅读 (15)

引言最近，访问外网出现不稳定的频率有所提升，随之而来的是 Windows 10 客户端 和 Office 365 客户端也出现了奇奇怪怪的问题：Windows 10 右下角显示“地球”图标，并提示“无Internet”，但是实际上国内网站都正常访问。Office 365 客户端无法登录，Outlook 无法收发邮件，其登录依赖于上述检测结果Windows 10 是如何判断是否有互联网访问的Windows 10 检测是否有互联网访问，通过一种叫做 NCSI (Network Connection Status Indicator) 的检测机制，当网卡连接后，会发起“主动”检测，检测该URL的可访问性：Win10:http://www.msftconnecttest.com/connecttest.txt如果该文件无法获得，则认定为“无Internet”。此时，如果用户手动访问以下地址，则有机率被检测到，并认为网络恢复。这种方式称为“被动”检测机制。Win10:http://www.msftconnecttest.com/redirect该链接会重定向到MSN网站。原因如果出现“地球”图标后，你可以手动访问这个链接：http://www.msftconnecttest.com/connecttest.txt目前，我的测试结果是，WebKit内核的浏览器，大约5次中有一次可以访问；IE浏 ...继续阅读 (11)

场景将重要的照片、文件等备份到云盘，有一定泄露风险。为了将重要文件进行加密，可以使用gpg或者openssl命令的对称加密方式（更高效率）然而，原生命令不支持对目录及子目录进行批量加密目的使用一种高效的方法，对一个目录进行递归读取后加密如果文件大小超过2GB，则对文件进行分段处理某些云服务可能对文件大小和下载速度进行限制，分段大小请根据自己的云服务自行调节不建议使用过大的分段，一旦损坏将造成整个分段的文件丢失解密时，如果遇到损坏的分段，则仅该分段无法解密，不影响后续分段使用方法：将以下脚本中您所需要的函数放入 .bashrc 或者复制所需函数到终端，粘贴运行即可导入Windows 平台可以使用 WSL使用说明passfile – 密码文件。只读取第一行，长度/复杂度任意。该文件必须位于命令执行目录，即工作目录。sslen / sslde – 使用openssl方式进行加密、解密。效率较好。推荐gpgen / gpgde – 使用gpg命令进行加密、解密。相比openssl效率较低。脚本# openssl - better performace sslen() { [ "$2" = "" ] && des=encrypted || des=$2 ; mkdir -p $des ; tar -c $1 | openssl enc -aes-256-c ...继续阅读 (11)

场景将重要的照片、文件等备份到云盘，有一定泄露风险。为了将重要文件进行加密，可以使用gpg或者openssl命令的对称加密方式（更高效率）然而，原生命令不支持对目录及子目录进行批量加密目的使用一种高效的方法，对一个目录进行递归读取后加密如果文件大小超过2GB，则对文件进行分段处理某些云服务可能对文件大小和下载速度进行限制，分段大小请根据自己的云服务自行调节不建议使用过大的分段，一旦损坏将造成整个分段的文件丢失解密时，如果遇到损坏的分段，则仅该分段无法解密，不影响后续分段使用方法：将以下脚本中您所需要的函数放入 .bashrc 或者复制所需函数到终端，粘贴运行即可导入Windows 平台可以使用 WSL使用说明passfile – 密码文件。只读取第一行，长度/复杂度任意。该文件必须位于命令执行目录，即工作目录。sslen / sslde – 使用openssl方式进行加密、解密。效率较好。推荐gpgen / gpgde – 使用gpg命令进行加密、解密。相比openssl效率较低。脚本# openssl - better performace sslen() { [ "$2" = "" ] && des=encrypted || des=$2 ; mkdir -p $des ; tar -c $1 | openssl enc -aes-256-c ...继续阅读 (8)

scpcan copy directory over ssh, butsparseparameter is not supported (unlikecp).To efficiently copy directory with sparse files, usetarandsshinstead.# Copy "images" directory to remote server tar -cvSf - images/ | ssh user@remoteHost "(cd /mnt/pve/vmdata/ ; tar -xf -)"S – stands forsparsetarscp.sh scriptUsetarscp.shto transfer directory over network:Usage:tarscp srcDir/ user@host:/dest/dir/You can find the script from mygithubDirect download link for raw file:tarscp.sh ...继续阅读 (19)

scpcan copy directory over ssh, butsparseparameter is not supported (unlikecp).To efficiently copy directory with sparse files, usetarandsshinstead.# Copy "images" directory to remote server tar -cvSf - images/ | ssh user@remoteHost "(cd /mnt/pve/vmdata/ ; tar -xf -)"S – stands forsparsetarscp.sh scriptUsetarscp.shto transfer directory over network:Usage:tarscp srcDir/ user@host:/dest/dir/You can find the script from mygithubDirect download link for raw file:tarscp.sh ...继续阅读 (15)

By understanding my previous posts:Hybrid Azure AD Join – How a computer device is recognized as Hybrid device ?Azure AD Connect: How to manually synchronize using import, syncronize, export?Now it is easy to find out how to make hybrid join happen immediately:Setup the hybrid AAD auto join infrastructure, i.e. AAD, SCP configuration, rollout plan (by GPO), etc.Make sure theuserCertificateattribute of the computer object existing.Delta import from on-premise AD (runDelta Importon the on-premise AD connector)Delta synchronize on the on-premise AD connector (runDelta Synchronizationon the on-pre ...继续阅读 (12)

By understanding my previous posts:Hybrid Azure AD Join – How a computer device is recognized as Hybrid device ?Azure AD Connect: How to manually synchronize using import, syncronize, export?Now it is easy to find out how to make hybrid join happen immediately:Setup the hybrid AAD auto join infrastructure, i.e. AAD, SCP configuration, rollout plan (by GPO), etc.Make sure theuserCertificateattribute of the computer object existing.Delta import from on-premise AD (runDelta Importon the on-premise AD connector)Delta synchronize on the on-premise AD connector (runDelta Synchronizationon the on-pre ...继续阅读 (11)

You may have noticed that:if you remove a hybrid domain joined device from AAD, it comes up again.if you revert the machine or shut it down, then remove the hybrid device from AAD again, still it comes up again.Why the hybrid domain joined device comes up again?When the device auto join the hybrid domain to AAD, it writes something to the device (computer) object in on-premise AD.What is the attribute to identify hybrid device?That is theuserCertificateattribute.You can verify it by the following steps:copy the hex code anddecode HEX to PEM certificate.Decode the PEM certificateCompare thecomm ...继续阅读 (9)

You may have noticed that:if you remove a hybrid domain joined device from AAD, it comes up again.if you revert the machine or shut it down, then remove the hybrid device from AAD again, still it comes up again.Why the hybrid domain joined device comes up again?When the device auto join the hybrid domain to AAD, it writes something to the device (computer) object in on-premise AD.What is the attribute to identify hybrid device?That is theuserCertificateattribute.You can verify it by the following steps:copy the hex code anddecode HEX to PEM certificate.Decode the PEM certificateCompare thecomm ...继续阅读 (12)

Recommended Steps:In short, the recommended steps to manually do the syncronization is:(Full/Delta) Import from on-premises directories, such as Active Directory(Full/Delta) Import from Azure AD(Full/Delta) Synchronization from on-premises directories, such as Active Directory(Full/Delta) Synchronization from Azure ADExport to Azure ADExport to on-premises directories, such as Active DirectoryUnderstanding the stepsThe keys to understand the steps:Import(step 1,2) imports things from AAD or on-premise AD to the AAD Connect server for comparing purpose.Synchronize(step 3,4) doesn’t really ...继续阅读 (9)

Create a scheduled taskCommand:%SystemRoot%\system32\WindowsPowerShell\v1.0\powershell.exeParameter:-ExecutionPolicyByPass-file "C:\VHD\mountVHD.ps1"Powershell Script to mount VHDMethod1: Mount-DiskImageMount-DiskImage -ImagePath "C:\VHD\DriveD.vhdx"Method2: Mount-VHDIf you were usingMount-VHDto mount VHD disks like this:Mount-VHD –Path "C:\VHD\DriveD.vhdx"You would be possible to get this error message in future:Mount-VHD : The Hyper-V Management Tools could not access an expected WMI class on computer 'LESCA-Computer'. This may indicate that the Hyper-V Platform is not installed on the compu ...继续阅读 (16)

IntroductionYou might have encountered this error message many times: “The trust relationship between this workstation and the primary domain failed”.The general solution is dis-joining and re-joining the domain. Any other ways without doing so? The answer is yes.Error messageThe trust relationship between this workstation and the primary domain failed.Solution: without rejoining the domainLogin as local administrator account, run Powershell with elevated privileges and type this command:Reset-ComputerMachinePassword -Server "DC1.contoso.com" -Credential domain\AdministratorAfter r ...继续阅读 (11)

Issue Description:The problem server installed the follwing roles:WDSDHCPWSUSRRAS (Route and Remote Access Service)The WDS service is configured with Active Directory integrated, everything seems configed correctly.Buta Hyper-V Gen2 VM client is not possible to boot with PXE. Message is"DHCP failed".If you redirect the vNIC and connect to the WDS server with a physical machine, you get a message “PXE-E16: No offer received“.Trouble shootingMy steps to trouble shooting:I deploy a new WDS server with only DHCP and WDS service. This WDS service works fine.I capture the packages with W ...继续阅读 (6)

Error MessageDebian or Raspbian: nfs-kernel-server not starting portmapper is not running … (warning) rpcbind.targetFix StepsHere is my first try, but it doesn’t work for me. It just reset the configs. So if you messed it up, try this first.apt-get purge rpcbind nfs-kernel-server nfs-common apt-get install nfs-kernel-serverNext, I have a look into/etc/init.d/nfs-kernel-serverand find its start level is2 3 4 5. However,nfs-commonandrpcbindhave differenct run level, I change it to the same, i.e.2 3 4 5.# Have a look at these files' runlevel /etc/init.d/nfs-kernel-server /etc/init.d/n ...继续阅读 (62)

Error MessageDebian or Raspbian: nfs-kernel-server not starting portmapper is not running … (warning) rpcbind.targetFix StepsHere is my first try, but it doesn’t work for me. It just reset the configs. So if you messed it up, try this first.apt-get purge rpcbind nfs-kernel-server nfs-common apt-get install nfs-kernel-serverNow check the service startup dependencies by command below. And make sure you findrpcbindbefore nfs service.systemctl list-dependencies nfs-kernel-server.serviceIf it is not, unfortunately you should trypurgeandinstallthem again as mentioned above, which should ...继续阅读 (216)

The Virtual Indicators allow you to monitor, and in some cases control, the state of indicators on the host system, including the Unit Identification Light (UID).The Unit Identification Light (UID)The UID helps youidentify and locate a system, especially in high-density rack environments. Additionally, the UID is used toindicate that a critical operation is underway on the host, such as Remote console access or ROM flash.In general, you control the state of the UID using the UID controls or using the physical UID button on the front of your system (or the back, if supported).The “current ...继续阅读 (37)

由于chrome v30开始有些特性不讨人喜欢，所以很多人喜欢停留在v29，只是在保留老习惯的同时，一些新的扩展就不能使用了。但是总不能让谷歌任行其道对吧？Lesca将介绍如何绿化v29并在v29上使用Chrome Web Store上的新版本扩展，如Proxy SwitchyOmega。下面先奉上做好的绿色版。做好的绿色版百度网盘下载：Chrome-bin.zip下载好后解压缩到当前目录，有两种方法可以运行：双击chrome.exe直接运行这样，chrome会将用户配置文件存放在%localappdata%\Google\Chrome\User Data\目录下。双击start_chrome.bat运行此时，chrome会将用户配置文件目录 User Data 存放在当前目录。这种方法的好处是，便于打包备份。你也可以将chrome.exe发送快捷方式到桌面，然后编辑快捷方式，在“目标”最后添加--user-data-dir="../User Data"即可。Chrome v29绿化之旅首先找到statble的安装文件：Chrome v29.0.1547.76 Stable.exe安装完后，在安装目录找到C:\Program Files (x86)\Google\Chrome\Application\29.0.1547.76\Installer\chrome.7z这个文件。由于这个 ...继续阅读 (275)

Hi! I am writing to explain how to test your network connection between your client (Windows or Linux) and another client (say, a remote server). The tool isiperfversion 3.You can find some useful explanations and examples on this French site:https://iperf.fr/But the version stays at 2 which does not have some amazing new features.Withiperf3you will not need to set parameters on the server side because the client will negotiate the parameters with the server.Here are some useful examples you may want to test withiperf3:# Classic mode (10 seconds) iperf3 -c 10.10.10.1 -w 300k # 2 Connections i ...继续阅读 (178)

上一节中我们介绍了如何在CentOS 7上配置MySQL的主从复制模式，今天Lesca将介绍如何备份和还原MySQL的从服务器，平台仍然是CentOS 7。使用mysqldump进行备份和还原使用mysqldump进行备份mysqladmin stop-slave -uroot -p mysqldump --all-databases > fulldb.dump mysqladmin start-slave -uroot -p tar -czf /tmp/dbdump.tar.gz ./fulldb.dump./mysql-relay-log.info我们除了要备份了整个数据库的dump以外，还需要备份relay-log.info文件（上例为mysql-relay-log.info），该文件包含类似如下信息：/var/lib/mysql/mysql-relay-bin.000002 720mysql-bin.000002 3968红色高亮部分，指明了当前MySQL主服务器上二进制日志的执行状态。这个数据在还原从服务器的时候至关重要。使用mysqldump进行还原mysql -uroot -p < /root/dbdump.db stop slave; CHANGE MASTER TO MASTER_HOST='192.168.10.201', MASTER_USER='slave_ ...继续阅读 (32)