Download fileserver

Seafile.conf 配置存储空间容量设置用户默认空间上限[quota]# 单位为 GBdefault = 2这个设置对所有用户生效. 如果你想对某一特定用户进行容量分配, 请以管理员身份登陆 Seahub 网站,在System Admin页面中进行设置.从企业版 10.0.9 开始，您可以设置资料库允许的最大文件数量，当资料库中的文件数量超出这个限制时，无法上传文件到这个资料库。默认没有限制。[quota]library_file_limit = 100000默认历史记录设置对所有的资料库设置一个默认的文件历史保留天数：[history]keep_days = days of history to keep资料库回收站清理周期对于删除的资料库，不会被立即彻底清除。它们会被临时存放到“已删除的资料库”和“系统管理”中的“资料库回收站”中，默认会在30天后自动清除这些数据。如果需要修改保留天数：[library_trash]expire_days = 60缓存（专业版）Seafile 会将一些信息保存到缓存中以提高性能，一些会话信息也会被保存到缓存中以在集群节点之间共享。Seafile 支持使用 Memcached 作为缓存。如果您使用的是memcached:[memcached]# 如果您使用的不是本地的 memcached，替换 localhost 为相应地址# POOL-MIN 和 POOL-MAX 用于控制连接池的大小。通常使用例子中给的值即可。memcached_options = --SERVER=localhost --POOL-MIN=10 --POOL-MAX=100Seafile fileserverSeafile 监听的端口号 (不要修改该设置)[fileserver]# Seafile tcp 端口 (不要修改该设置)port = 8082从社区版 6.2 和企业版 6.1.9 开始，你可以设置用于服务 http 请求的线程数。默认值是10个线程。这个默认值适用于大多数应用场景。[fileserver]worker_threads = 15上传/下载大小限制：[fileserver]# 上传文件最大限制为200M，默认是无限制.max_upload_size=200 # 最大下载目录限制为200M, 默认是100M. max_download_dir_size=200 通过Web界面或客户端中的云端浏览器上传文件后，需要将其分成固定大小的块并存储到后端存储。我们称这个过程为“索引”。默认情况下，文件服务器使用1个线程顺序索引文件并逐个存储块。这适用与大多数情况。但是如果您使用的是 S3/Ceph/Swift 后端，则在存储后端可能会有更多带宽来并行存储多个块。我们提供了一个选项来定义索引中并发线程的数量；[fileserver]max_indexing_threads = 10当用户在Web界面上传文件时，文件服务器将文件分割成固定大小的块。Web上传文件的默认块大小为8MB。块大小可以在这里设置。[fileserver]#Set block size to 8MBfixed_block_size=8当用户上传文件时，文件服务器分配一个令牌来授权上传操作。该令牌默认有效期1小时。通过WAN上传大型文件时，上传时间可能会超过1小时，您可以将令牌到期时间更改为更大的值。[fileserver]#Set uploading time limit to 3600s web_token_expire_time=3600您可以从Web界面下载文件夹为zip存档，但是Windows上的一些zip软件不支持UTF-8，在这种情况下，您可以使用"windows_encoding"设置来解决此问题。[zip]# The file name encoding of the downloaded zip file.windows_encoding = iso-8859-1“ httptemp”目录包含在文件上载和zip下载期间创建的临时文件。在某些情况下，文件传输中断后，临时文件不会被清除。从7.1.5版本开始，文件服务器将定期扫描“ httptemp”目录以删除很久以前创建的文件。[fileserver]# After how much time a temp file will be removed. The unit is in seconds. Default to 3 days.http_temp_file_ttl = x# File scan interval. The unit is in seconds. Default to 1 hour.http_temp_scan_interval = x自 Pro 7.1.16 和 Pro 8.0.3 中的新功能：您可以设置 Seafile 客户端可同步的资料库中包含的最大文件数。默认值为 100000。下载资料库时，Seafile 客户端会请求 fs id 列表，您可以通过 fs_id_list_request_timeout 选项控制请求的超时时间，默认为 5 分钟。添加这两个选项是为了防止长时间的 fs-id 列表请求导致服务器超载。自专业版 8.0.4 起，您可以将这两个选项都设置为-1，以允许无限制的大小和超时。[fileserver］max_sync_file_count = 100000fs_id_list_request_timeout = 300从 Seafile Pro 8.0.6版本开始，您可以用移动端浏览器下载文件。[fileserver]# 允许所有文件上传、下载 token 被多次访问。# 默认下载的 token 不允许多次访问，这造成了对一些移动端浏览器的不兼容。web_token_reusable = false如果您使用对象存储作为存储后端，当一个大文件频繁下载时，需要从存储后端获取相同的块到 Seafile 服务器。这可能会浪费带宽并导致内部网络的高负载。从 Seafile Pro 8.0.5 版本开始，我们添加了块缓存来改善这种情况。请注意，此配置仅对通过网页或 API 下载文件有效，对同步文件无效。要启用此功能，请在 [fileserver] 组中设置 use_block_cache 选项。默认情况下未启用。block_cache_size_limit 选项用于限制缓存的大小。其默认值为 10GB。这些块缓存在 seafile-data/block-cache 目录中。当缓存文的总大小超过限制时，seaf-server 将清理旧文件，直到大小减少到限制的 70%。清理间隔为 5 分钟。您必须很好地估计缓存目录需要多少空间。否则，在频繁下载时，这个目录会很快被填满。block_cache_file_types 配置用于选择缓存的文件类型。 block_cache_file_types 默认值为 mp4;mov。use_block_cache = true# 将块缓存大小限制设置为 100MB block_cache_size_limit = 100# 针对特定的文件后缀使用缓存，默认 mp4 和 mov 格式block_cache_file_types = mp4;mov当有大量文件通过网页、API上传时，基于文件内容计算块的哈希值会有很大的计算开销。从Seafile Pro 9.0.6版本开始，您可以通过增加skip_block_hash选项来使用一个随机的字符串作为块ID，从而减少计算开销。注意，这个选项会影响 fsck 程序，使得无法检查块的内容是否损坏。因此设置了这个选项后，执行 fsck 的时候需要指定 --shallow 选项，不验证块内容的完整性。[fileserver]skip_block_hash = true如果你想在上传文件时限制文件类型，从 Pro 10.0.0 版本开始，您可以在 [fileserver] 组中设置 file_ext_white_list 选项。该选项是一个文件类型列表，只有列表中的文件类型才允许上传。默认情况下不启用。[fileserver］file_ext_white_list = md;mp4;mov从 10.0.1 版本开始，使用 go fileserver 时，您可以在 [fileserver] 组中设置 upload_limit 和 download_limit 选项，以限制文件上传和下载的速度。默认情况下是不启用的。[fileserver］# 单位是 KB/s。upload_limit = 100download_limit = 100更改MySQL连接池大小当您将seafile服务器配置为使用MySQL时，默认连接池大小为100，这对于大多数用例应该是足够的。您可以通过在seafile.conf中添加以下选项来更改此值：[database]......# Use larger connection poolmax_connections = 200使用 unix_socket 认证插件从社区版 10.0.5 和企业版 10.0.7 开始，您可以使用 MariaDB/MySQL 提供的 unix_socket 认证插件。您可以通过增加 unix_socket 选项，并且去掉 user 和 password 选项，来开启该功能。[database]......# unix socket 文件路径unix_socket = /var/run/mysqld/mysqld.sock开启 Slow LogSeafile-pro-6.3.10 开始，Seafile增加了 seaf-server 的 RPC 慢请求查询日志，便于管理员更好的做性能分析。该功能是默认开启的，如果您想要自主配置相关选项，可以在 seafile.conf 中添加如下配置：[slow_log]# 默认为 trueenable_slow_log = true# 所有慢请求日志阈值的单位为毫秒。# 默认为5000毫秒，这意味着只有处理超过5000毫秒的RPC查询才会被记录。rpc_slow_threshold = 5000在 logs/slow_logs 目录下，可以找到 seafile_slow_rpc.log；并且该日志文件支持使用 log-rotate 做日志切割，只需要向 seaf-server 进程发送 SIGUSR2 信号，进程就会关闭并重新打开日志文件。自 9.0.2 Pro 起，触发日志旋转的信号已更改为SIGUSR1 。此信号将触发 seaf-server 打开的所有日志文件的轮换。您应相应地更改日志轮换设置。启用访问日志尽管Nginx记录了所有具有某些详细信息的请求，例如url，响应代码，上游进程时间，但有时需要有更多关于请求的上下文，例如每个请求的用户ID。此类信息只能从文件服务器本身记录。从 9.0.2 Pro 开始，访问日志功能已添加到文件服务器。要启用访问日志，请将以下选项添加到 seafile.conf[fileserver]# default to false. If enabled, fileserver-access.log will be written to log directory.enable_access_log = true日志格式如下：start time - user id - url - response code - process time可用于SIGUSR1触发日志轮换。Go Fileserver性能分析从Seafile 9.0.7开始，您可以增加下面的选项来开启 go fileserver 的 profile 功能。# profile_password是必须配置的，您可以按需修改。[fileserver]enable_profiling = trueprofile_password = 8kcUz1I2sLaywQhCRtn2x1这个接口可以通过 Go 语言提供的pprof工具来使用。pprof的详细信息，请参阅 请注意，您在使用以下命令之前，必须先安装 Go。 password 参数应该与您在配置中设置的相匹配。go tool pprof tool pprof Seafile 10.0.0 开始，您可以通过添加以下配置选项来启用通知服务：# jwt_private_key 是必需的。您应该手动生成它。[notification]enabled

JMeter is one of the most popular load testing tools on the market. When using it, there are a wide array of features available to testers. One of those features is JMeter File Upload/Download.This blog will serve as a step-by-step guide to JMeter File Upload/Download.Time to get started!Back to topUsing JMeter File Upload/DownloadFile UploadNow that we have given a brief overview of JMeter performance testing, let us dive into uploading files to JMeter. A form-based File Upload in HTML is described in RFC-1867, which identifies “multipart/form-data” media type for file attachments. JMeter's HTTP Request Sampler in its turn offers a “Use multipart/form-data for POST” checkbox. Ticking this checkbox indicates that the request contains file data. The file(s) needs to be supplied through “Files Upload”' input. In HTML, the source upload form looks like this:method="post" action="UploadServlet" enctype="multipart/form-data"> Welcome to Blazemeter upload test servlet, select file to upload: type="file" name="dataFile" id="fileChooser"/> type="submit" value="Upload" />It is critical to mention “dataFile” as the name of the “file” input, since this must be the parameter that needs to be passed as “Parameter Name”. The HTTP Request should look like the following:Pay attention to the following parameters:File Path: should be the location of the file to upload.Parameter Name: relevant form of “file” type name.MIME Type: the file's Internet media type.The easiest way to get these fields populated for you is using JMeter's Superpower: The HTTP Proxy Server. JMeter is smart enough to detect and populate values during the recording process. Just make sure to copy the file(s) you will be uploading to your JMeter installation’s “bin” folder before uploading the file in the browser, as this is the only way that JMeter can properly capture the request. File Path There are two options for providing a file path:Full path, i.e. c:\somefolder\myFile.txtRelative path, i.e. myFile.txt only When using the full path, JMeter will upload the given file as it is. When providing a relative path such as file name only, JMeter will look for it in the FileServer base folder. You can check the jmeter.log file to see which folder is being investigated by searching for lines that contain “FileServer.”2022-05-01 09:36:59,585 INFO o.a.j.s.FileServer: Default base='C:\jmeter\bin'You can get or set the value using any suitable JSR223 Test Element with Groovy language, such as:import org.apache.jmeter.services.FileServer;FileServer fs = FileServer.getFileServer(); //instantiate FileServer classlog.info(fs.getBaseDir()); // get FileServer base dir and write it to jmeter.logfs.setBasedir("/path/to/test/files/location"); // set a new base dirYet it is important to consider the following when choosing a file path:It is preferable to use full paths when running tests locally. Relative paths may differ depending on where the .jmx file is and/or how you launch JMeter.If you are uncertain regarding where to put files, it is recommended to set up

Adding an SPN (service principal names) to an account may entice the attacker to attempt Kerberos the account which should also trigger an alert from Defender for Identity. As an example, around a SQL server environment, you could create a SPN for SQL (again be sure to remove privileges to server\database). One other great option is to put them in fake groups called something like sec.right.AllServerAdmins. This will make them appear to have broader admin access.Just as important as the type of account you use and the meta data surrounding it, is how visible your Honeytokens are to attackers that may have penetrated initial access controls and begun to do internal discovery. Below are a list of techniques that you may want to leverage to ensure these fake accounts are easily uncovered by common attack strategies. One way to plant breadcrumbs leading to these accounts is to place the account and password in a Domain Admin Passwords.xslx on a discoverable \\fileserver\fileshare. When using this method be sure again to use an existing .xslx with an older creation date versus creating a new one. Tools like Mockaroo can help you to quickly mockup data as well to put into the file.The above concept allows discovery in cleartext, some other examples of location of file/data might be.SharePoint Site (file listed as passwords)Code RepositoriesMapped DrivesLocal documentsYou can also leverage other techniques, for example placing the account in a PowerShell script, and attaching the script locally or in a \\fileserver\fileshare to use as a

It on any GP. I suppose in a large environment you may want to limit unnecessary processing, but I’d rather have a tiny bit of extra overhead processing and less chance of issues. Granted, I’ve only been doing this for about a decade, and go back a few years further and things might be different. I’m having a very similar issue (Windows 10 Pro version 1803 desktops, Server 2016) we have a group policy user config preferences set to map two network drives with a path such as “\fileserver\data\admin” and these map fine. We also have a group policy user config preferences set to map the users home drive with a path such as “\fileserver\users\staff%username%” and this drive is quite often missing, restarting seems to be the only way to get it to come back.Both policies are set with “Update” as the action and does not have a tick in the “Apply Once” check box.Did you find out if it’s best practice to use “Replace” rather than “Update”?Our drives are SSD’s and I’ve thought about them booting too quickly before network ready so I’m tempted to try and find the GPO’s to make them wait till the network is ready - have you tried enabling this and has it made any difference?I’m wondering why any network would be setup to map user’s home drives through group policy as opposed to mapping the home drive through the Active Directory Users profile tab? robpeterson2 (RobTheGiant) March 20, 2019, 12:17pm 11 This may