Ubuntu Servers :: GUI For Several Thousand Simultaneous Users
Jan 30, 2011
GUI for several thousand simultaneous users. I know this post may at first seem a little out of place here when you begin to read it but please bear with me and I think you will understand why I am seeking feedback here. The company I work for (whose name I am withholding for now) is currently running a little over 5,000 Linux servers. Each of these servers has a minimum of 500 users, some as many as 3,000 users on them during a business day.
These servers are serious pieces of hardware. They have hot swappable raid arrays, redundant well most everything is redundant on them, frankly. Our conversion to Linux took place about 8 years ago. We have a proprietary database and development environment that was originally built in the 1970s. All but the most recent applications run in this proprietary environment, whereas our newer applications are using PostgreSQL and executables that are delivered using Apache web server and run inside each users Windows Explorer session on their local PC.
Really, our situation is rather typical of any organization that has legacy software that still gets the job done - with the possible exception of the **massive** number of total users we keep up-and-running every day. more, our next Linux servers are expected to support somewhere between 6,000-8,000 users [exact figure yet to be determined]. You may ask this is all very interesting but why are you posting this here?.
I am researching alternate ways of presenting attractive and functional GUI to users of Linux servers. I find the Ubuntu GUI to be attractive and very functional. the user input these forums get. Finally I am looking for points of view outside my own organization so I don't fall into the 'but we don't do it that way here' mental trap and miss a really great concept.
I have an idea about how to solve this growing challenge, and I want to present it here for your review and comments. For those who might think I have never worked on systems of this size; I can't imagine my opinion would be valuable remember this: The Ark was built by a passionate volunteer; The Titanic was built by paid professionals.....
How can I configure a single computer to work for multiple simultaneous users? I would like to have multiple keyboards and monitors on a single PC ... Is it possible? This set up I'm planning for CyberCafe, For cyberCafe business ROI is very slow... so was planning to have something like this....
-single CPU (Which would act as a server) -Multiple monitors, keyboards, Headphones etc....
If this works out then I would save a lot on my investments (OS cost, Hardware devices, electricity etc)
I'm writing a client-server program. There are more than 500 clients. I start a thread to process and response to each client and the processing needs some MySQL query. I'm looking for any possible hazards on my server!
1- Any limitation on "Maximum Simultaneous Socket Connection"?
2- Any limitation on using mysql?
3- As socket on Linux are file, Any limitation on number of sockets or threads?
I'm using a Linux server (Centos or Fedora or Ubuntu) and clients are both Linux and Windows.
If I want to add Windows & Mac users as Samba users, must I first add them all as Ubuntu users? If so, since none of the other users will actually be working on the Ubuntu Server, how do I disable the other non-admin users on the Ubuntu Server login screen. I am using Webmin to administer some server settings, and command line for others.
I wonder if other people have this problem. sometimes I need to duplicate most or all of a server from one machine to another. part where I most tend to have issues is with recreating the users. Sometimes the UIDs don't match up because of differences between the two systems, which gives funky results, like files owned by the wrong user.
Or, sometimes, I just need to create the same 10 users on x number of machines, and it's tedious to do this 10x times. Plus, then, if you need to change a password or delete a user, you have to do it x times.Any recommendations on how to centralize users/auth?
following a new motherboard install I've just done a clean install of 10.10. At first, I was using the default network manager wifi to connect to my router - I had the upturned radar beam type icon in the top panel showing me the strength of the signal and access, by right clicking on it, to things like Enable Networking, Enable Wireless, etc. I read that installing WICD via synaptic would install wicd and remove the default network manager, so this what I did. My wifi is OK but I still have my upturned radar type icon and a new WICD icon in the top panel. I was expecting the upturned radar icon to disappear - is this correct? If I right click on the radar icon I can click on the Edit Connections option and change things - it appears I have both the default network manager and WICD working at the same time.
I am using the latest version of ubuntu(10.04 LTS) Desktop edition 32Bit. My problem is, after I open the firefox browser for first time after booting, a chain of new firefox windows start opening one after the another. In a matter of few minutes, there are like 15-20 instances of the browser opened.
I reinstalled the OS, but it is happening again and again. Is there some flaw in OS, something related to OS not compatible with my computer or some problem with the keyboard where some key(combination) is in a "press" mode (however, I dont see that happening on Laptop, the keyboard looks fine, plus Windows XP is working just fine).
This problem is forcing me to spend a larger time on Windows (which is bit cringy, after you get used to Ubuntu!).
I've got two soundcards in my new machine - Realtek ALC8892 onboard, with headphones plugged in, and an M-Audio Audiophile 192 with speakers. There's also HDMI sound out on the graphics card, but I don't use that.
Audio with PulseAudio's output through the sound card is fine. Audio with PulseAudio's output through the onboard sound is fine. If I set the output to Simultaneous Output, however, I get significant lag. Almost half a second. It's quite distracting.
Is this just a problem with PulseAudio's simultaneous output implementation, or is there a fix that I've overlooked? Running 64-bit Natty, btw.
I have created my own server with php, mysql and pma. I want to creat users and disable ssh so they can only user their username and password to get access to /home/<username> with a ftp program.
I have a ubuntu server 8.10 running on my dedicated server. How do I allocate IPs for users ? I have 3 IPs and 3 users running, I want each user to use a Single unique IP for all inbound and outbound traffic Also disallow a user from using the IP other then the one allocated to him. I tried to work out a solution using Iptables, but it does not work I suppose
My problem is quite simple, when I set up squirrelmail I only had one user on the server, and now that I have created new users, I can not log in to their accounts via squirrelmail I am using Maildir System is 10.04 server, running postfix and dovecot. I origionaly set up 'remote' as a user, I can send and recive mail on that account and log into squirrelmail
any new accounts can send email to remote(via outlook/thunderbird), to be seen in Squirrelmail, but get an invalid username/password when I try to log into squirrelmail in short, new accounts can not log into squirrelmail
I really hope someone could help me with this problem. I've been stuck on this for a month.I am using the sftp command to upload files using a bash script.The problem is that it is extremely slow to do it this way. as many of you would know if you have shared server somewhere. I would use scp if remote server supported it, but it doesn't.
Anyway, If any of you have ever used FileZilla, in the Settings, if you go to "Transfers" there is a place where you can set the number of "maximum simultaneous transfers". This feature works wonders with SFTP (and FTP too). It really speeds things up. How do I accomplish this same thing with the sftp command... because I don't want to use a GUI. I don't even mind using FileZilla through the command line if possible... but it does not seem to be possible. I've been stuck on this for a month!!! I've searched everywhere and tried a lot of things with no avail...
I'm trying to launch CoverGloobus (a fantastic little album art and current song display) and Rhythmbox simultaneously through a BASH script. Problem is that when I attempt to run the script the first time, CoverGloobus doesn't launch, but Rhythmbox does. Here's my script as it is right now:
Code:
#! /bin/bash #Opens Rhythmbox and CoverGloobus rhythmbox && covergloobus done
I've double-checked the commands in gnome-terminal, and they're the right ones for launching the applications. Here's where it gets odd, though. The second time I run the script, CoverGloobus and Rhythmbox launch simultaneously. I've reproduced this scenario multiple times.
sat in an internet cafe with my music on loud using my headphones when my cell phone vibrated and upon taking off my headphones I noticed that pulseaudio was channeling all my music not only to my headphones but also to my regular monitor out!
So I was actually listening to music using my headphones when in fact the sound was LOUD on my laptop speakers...
Is there a way to configure the thing to not do that? I want the sound to be muted as soon as I plug in my headphones.
As the title says, I'm interested in two (or more) simultaneous keyboards, each with a unique layout. Everything I've researched indicates this should be possible, but I'm unable to get it to work. Regardless of how I set things up, all keyboards consistently operate as if they share the CoreDevice setting in xorg.conf. In the above config, both keyboards behave as if they're set for the USA layout. Changing the layout in gnome-keyboard-properties affects all of the connected keyboards equally, but has no provision for altering individual devices.
I work for a company that uses Ubuntu exclusively. Without going too much into detail, we need a single sound-card to do up to 3 things at one time. What I mean by this, basically, is that there needs to be 3 instances listening to a card's line-in port at any one time. For example (these are made-up), Audacity listening to the line-in jack and making a recording of what it hears, a custom script that streams the input to the internet, and a third instance that listens to this input and runs different scripts based on what it hears.
Currently, our company sends out machines with 3 sound cards to do this. One card for each job. Of course, we'd like to save money and only send out one. Whenever we try to set all 3 programs to run off of one card, they error out and say something along the lines of "device already in use". Is there a way to go around this? If so, how? We use the command-line version of Ubuntu 8.04. It is NOT x64. The hardware may vary a bit, but is for the most part, consistent. (We order in batches of 50.)
i am using ubuntu 10.04 lts "lucid lynx", and has been encountering problems with audio recently.
i found out for some reason that my system fails to produce sound through pulseaudio, especially when an application is using it (e.g audacious).
the details of the problem:
1. i am using the pulseaudio equalizer from the repository, and i set a program (e.g. audacious) to use pulseaudio so that i can activate the equalizer.
2. while playing, all of my other applications lose the ability to produce any sound, and changing back to alsa during playback (which is using pulseaudio) in audacious indicates that the device is busy. examples include the loss of sound while playing a video from videos.
rough breakdown of what i want to (be able to) do: have sound for all applications using it, for example being able to play a video in videos from my browser and play music in audacious simultaneously.
is there a workaround for this? i'd like to think that there is, but so far my search returned nil.
I need to setup a simple server to make available Quickbooks accounting data to just a few stations. Can this be done with Ubuntu desktop version? Would it be recommended due to the small number of users? Or should I use a server platform?I have never used Linux before but have used command line interfaces and am looking forward to it. I installed Ubuntu Server 9.10 32 bit and have even gone as far as logging in. Amazing I know. I installed it on a separate disk so I can switch between Windows and Linux startup disks at boot via cmos.as a first step into linux would like to be able to browse the web from Ubuntu Server 9.10 if the Server edition is in fact the route I should be taking.I would also like a good source to learn not only command line commands but also all the stuff that I see after the commands, you know, all the stuff with - signs all over the place.
I'm using vsftpd on my server. When I connect (using file-zilla) from other computers on the same network I can't download any files. I can upload, create directories, and delete stuff, but I can't download. I've disabled anonymous access and enabled local user log-in. My /etc/vsftpd.conf
Code: # Example config file /etc/vsftpd.conf # # The default compiled in settings are fairly paranoid. This sample file # loosens things up a bit, to make the ftp daemon more usable. # Please see vsftpd.conf.5 for all compiled in defaults.
I have two ubuntu 10.04 64-bit servers running samba (3.4.7) and openLDAP (2.4.21). The LDAP directory is successfully replicating between the two servers. These servers also serve as LDAP servers for sudo, pam, nss, and other services for a dozen servers without issues. The BDC samba is configured to use itself for LDAP. I connected to the BDC using the samba ldap credentials and verified I could a) see the Computer object b) read NTPassword and LMPassword. The workstations can authenticate to the domain successfully against the PDC. If a workstation boots and connects to the BDC, they login fails with:
Code: [2010/07/18 11:46:23, 0] rpc_server/srv_netlog_nt.c:336(get_md4pw) get_md4pw: Workstation MACHINENAME$: no account in domain [2010/07/18 11:46:23, 0] rpc_server/srv_netlog_nt.c:584(_netr_ServerAuthenticate3) _netr_ServerAuthenticate3: failed to get machine password for account MACHINENAME$: NT_STATUS_ACCESS_DENIED
Successful authentication against the PDC shows: Code: [2010/07/18 11:59:20, 1] smbd/service.c:1063(make_connection_snum) MACHINENAME (192.168.2.145) connect to service netlogon initially as user username (uid=30000, gid=512) (pid 1727) [2010/07/18 11:59:20, 1] smbd/service.c:1063(make_connection_snum) MACHINENAME (192.168.2.145) connect to service data initially as user nobody (uid=65534, gid=65534) (pid 1727) .....
installed Ubuntu Server Edition and I've found that my first user has a bash history and I can turn on a coloured prompt by editing my .bashrc etc but new users don't have that!I did : useradd -d /home/newb -m newbpasswd newband the correct looking .bashrc file appears to be in /home/newb but it is being ignore by bash when logged in as newb. Instead I am presented with just a dollar prompt instead of "newb@server"how can I sort out my users with proper prompts?
I am moving a webserver from a Gutsy to Lucid server. The webserver works fine, but I am having a problem with the users. I moved the home directories, along with passwd, group, shadow, and gshadow. The users can login fine, and their home directories are fine. The problem is when they try to logout, they get
How do I migrate my user list to a new server installation? I thought that just copying /etc/passwd and /etc/shadow would be enough. I recently tried upgrading a lightweight bazaar server to 11.04, and the upgrade crashed. I had to perform a new installation of Ubuntu to recover operation of the system. I installed on a different hard drive. I had a dozen other people with user logins, so that they can get or modify bazaar repositories.
The new installation is running, and I can read all of the files on my old system's / drive. I tried copying the /etc/passwd and /etc/shadow files onto my new /etc -- but that did not seem to work. (When I looked using the System->Administration->Users-and-Groups GUI tool, there were no users visible, so I restored the old files.) What I intended to do was to clone the /etc and /home trees over to the new system. What am I missing?
Not sure if this is the right place fo this but....I'm just learning to write bash scripts. I'm experimenting on a ubuntu server 11.04 virtual machine.As part of the script I want to automate user creation.Is there a way to check if a user already exists in a script. so i can do something like
Code: if userexists; then do this else do this instead fi
I was looking around on Wikipedia, and ran across the UltraSPARC T2 processor. This is a processor with 8 cores (8 physical processors) and 8 threads per core (64 logical processors!)I'm aware of Intel's Hyper-Threading and other technologies, and I know collisions between the threads have the possibility of decreasing the processor's throughput.From what I know, the O(1) scheduler (and surely the current CFS scheduler) is aware of Hyper-Threading, and is careful with processor affinity for threads (mainly, it tries to fill every physical processor before giving any multiple threads, and it tries to keep processes that much move between logical processors on the same physical processor.)
However, although that certainly helps somewhat, I believe the largest gain can come if the compiler is allowed to re-arrange instructions, so that more than two threads can be run at once without collisions in one physical processor. For this to happen, I believe we'd need to come up with a new executable format that can store multiple threads in one file, so the scheduler can schedule them as one block.For example, with the current scheduler and executable formats, the UltraSPARC would appear as 64 logical processors. However, if we consider a system that consists entirely of thread-aware systems, it might look at it as 8 logical processors, and schedule eight processes at a time. Since only one process would exist on each physical processor, the compiler would be able to optimize and significantly reduce resource contention (which I mention below.)
We can look at one physical processor with 8 threads of execution (or any number >2.) If you run 8 different processes, as Linux would currently do, there would likely be a lot of resource contention. My guess is that most of the time, several threads would be waiting for the unit they need to become available. I feel that this is supported because I saw a number from Sun (I forgot where) that the correct optimization settings on their compiler can yield a 300% speedup, on a practical workload.Is there any infrastructure available for Linux that supports processors like the UltraSPARC T2, that have a high number of hardware threads per core? Also, I know this would be a large task to implement, and there would be hurdles (how do you mix multithreaded processes with single-threaded processes on one machine?), and I am wondering if others think there would be enough benefit to justify the effort.
We just setup a Moodle server in Apache. We would like to track how close (if at all) we are getting to our MaxClients limit. Is there a log that will show simultaneous connections?If there is, can it be set to only log when specific numbers are hit?
I've setup an HTPC with xbmc, and I'd like to output the audio from the system to both my digital and analog outputs on my mother board. How do I do it? Can their be a script that duplicates the output of one to the other? Here's what aplay -l said:
I have a server setup with all my web development stuff in /var/www and in several sub-folders within that. (each project having it's own folder)It works great with one FTP account. But recently I've been getting help on a projects from a buddy of mine that freelances, and have made him an FTP user account as well. All is fine, except for when he tries to edit a file and gets a permissions error.
Here's the issue, I don't want us to have the same FTP login, but all the files are currently owned by my user name. So, when he logs in to edit a file, he can't because I'm the owner, and the files are set to 744. Will I cause any harm by adding both users to the same group (www-data) and chmod'ing the files to 775 so that we can both access and modify the files?
