 My name is Masaki Burakami and today's presentation is Just Scale-Operation Rebreak Song Night in 2019 My name is Masaki Burakami and I'm from Kobe city in Japan My main office is Kobe and my branch office is Tokyo, Vietnam and Mongolia My task is several networks for IT infrastructure administration I can tell you something about LibrePC or LibreBox LibrePC is a PC for Linux and other open source tools in PC LibreBox is an integrated of LibreBox Online and Nexsquare installation It is new to management in these tasks for many months Last year, I presented a certain object as Scale-Operation Rebreak Song Night But it is not good for presentation scale One year later, I ran a few sense of the operation LibrePC Song Night And I tried to re-approach it This presentation is the same approach as I presented at LibrePC Asia Conference at Tokyo this year The operation for many users is the same For many users using LibrePC Song Night, it is considered the same One-fifteen, many users at the same time The same time operation for many users And for many users, five or ten or twenty users Such as an infrastructure-changing or no coding skill Building and editing source code is a very hard way to do it Many servers building and load balancing is suitable for scale of themselves It is an easy way to do it It has tried to configure load balancing for several different online servers than usual Cap-Stick planning is important for creating new version-type servers for many other servers It is a medium-paragraph open-source document for Cap-Stick planning It is that bit, and CPUs for one-share per ten active users Memory is ten megabyte per one active user And network bandwidth is 100kbps per one active user Example for fifty active users usually use CPUs for five-share And memory prepares to 500 megabytes And network bandwidth preparing for 500 megabytes And it is so easy to prepare it for some of these servers Configuration is very important for using servers Editing external files in install directory and configuring it Logging is very important for operating servers Log file is temp directory by default for it But it usually servers not using temp directory Another directory to change is linux Logging is not by default and changing settings SSL session is for certification part The port is self-signature and meantime use its press directory to edit it And for this session is external servers only for the web server And load balancing is very important for continue to provide the service Many Japanese customers really have to continue to provide the service And load balancing is usually made to continue to provide the services First step is configuring two developers on-line servers And connecting from the next cloud with a follower on-line apps This is the testing environment Network load balancer is at first I tried to use a load balancer Provided by cloud server provider it is simple function And it is made to load balancing to access from the next cloud server One file access to developers on the server, two developers on-line at the same time And it is there And I know that it is a function of identity of a load thing about file And it is not to use network appliance It is too expensive about a file becoming And check access load check at Apache load file And to find the key of identity of each file Access load and get the hint I got the hint SSL access load And access token or we can use this Tool for load balancing is If you use hf proxy services You can use load balancing for a custom header as well And load balancing connection I think But all versions of hf proxy are center-wise It is not used as a custom header Next try to use eSentry version of hf proxy And I think that we need more ideas And otherwise the solution is problem And I will try it And I will try it