On 07/12/06, Eilert <[EMAIL PROTECTED]> wrote: > Looking for this entry in the Wiki, but didn't find it - can you give me > a hint what I've got to look for? >
Look at running two dhcpd ltsp boxes on the same network. When the client machine is witched on the server which replies first gets the client process load. Generally the servers even out in terms of client loads as the server which is least loaded offers dhcp lease fastest. Something like round robin response. > I'm just about planning such a setup here, but it's for 40+ clients. > Currently we've got only one server with two Xeons and 4 GB Ram, but if > more than about 35 or 40 clients are used (yesterday there were 44 for a > while), it begins to get somewhat slow. > > The most important thing is that it's dangerous to make everyone > dependent on only one machine. That is why my suggestion of running two ltsp servers with each running its own dhcpd service for the same subnet. The rough schematics is as explained above. One server will have its IP as 192.168.0.254 and the other as 192.168.0.253 If you can not find anything post back and I will explain with example. Common directory structure was created by first creating /home/user and /home/back paths on both servers. Users accounts are in sub path below /home/user Server1 should export its home directory by adding something like this in /etc/export file: /home 192.168.0.0/255.255.255.0(rw,no_root_sqaush,async) And Server2 should mount this on its /home/user path by and entry in /etc/fstab like: 192.168.0.254:/home/user /home/user ext3 defaults 0 2 This way which ever server the user logs on the home directory is same. Ofcourse this also is assuming that Server1 and Server2 are both having identical copy of /etc/passwd /etc/shadow and /etc/group and they are frequently synced. If there is another scheme of authentication at play then both point to same auth-server. Now the server2 acts as a back up server where Server1:/home/user is rsynced to Server2:home/back regularly through cron job. If Server2 goes down there is no issu as Server1 will continue to work and users will experience some slow down. If Server1 goes down then the admin has to on Server2 quickly "ln -sf /home/back /home/user" so that users can continue to work from Server2. It is best if you can run this NFS export/mount another physical subnet (say 192.168.1.254 and 192.168.1.253) so that NFS export traffic is not slowed by general traffic on 192.168.0.0/24 subnet used by LTSP clients and other machines. You may automate these steps by use of scripts and some ping response check on server2 to see status of server1. -- Regards, Sudev Barar ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _____________________________________________________________________ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.freenode.net
