I'll make my best efford with my english:-)

I am using LTSP for about three years. Lan's topology consist of 15 diskless 
workstastions (using 3COM's builtin eproms, 8139's with floppy from 
rom-o-matic and 4 machines with builtin network that can networkboot). All 
works great in this config but... I add 5 new machines that are 180 meters 
away from my LAN, in a another building.
Solution was bridging "two" networks with a pair of Linksys WAP54G bridges:



 xxxx    xxxx                                       xxxx    xxxx
 x  x    x  x                                       x  x    x  x
 xxxx    xxxx              (               )        xxxx    xxxx
xxxxxx  xxxxxx            (        +    +   )      xxxxxx  xxxxxx
xxxxxx  xxxxxx          |(       +  + +      )|    xxxxxx  xxxxxx
  |        |      [1]   | (    +     +      ) |  [2]  |        |
--+--------+---+BRIDGE+-+  (               )  +-wap54-------------
  |        |    WAP54
 xxxx    xxxx            perfect line sight    "New segment added"
 x  x    x  x            no other is using        6-port switch
 xxxx    xxxx              the channel           3 Duron 2500+ in
xxxxxx  xxxxxx                                     Biostar mobo w/
xxxxxx  xxxxxx                               sis900/via onboard networkboot
                                              2 terminals with 3COM905C-TX-M
"ACTUAL TOPOLOGY"                            NETWORK: 192.168.0.0 (the same)
15 workstations
P4 Server w/768 RAM
HUB 100Mbps
NETWORK: 192.168.0.0

So, bridging unifies everything in one network. Transmit/Receive power
is excellent (I changed omni antenas for 24dB grid antennas). In the new 
building,
no machine has hard disk, when I power on a terminal, in 100 seconds kdm login
appears (think that in normal conditions, it takes 25 secs). Performance
is just fine, but graphically all is more slowly.
Every time machine boots, I receive "nfs server not responding, still trying",
althought it can boot, it is due to network congestion (or packet loss in 
air).

I would like to get more performance in that new machines, I thought to add
a Linux router+nfs+tft to server only the new machines, and for secure the Lan
that has the machine where users has their accounts. This new machine will 
server
the kernel and filesystem for that 5 machine in one interface and provide a
connection to the bigger Lan (in this case 192.168.0.0 the bigger network and
192.168.1.0 for the remote's machines).

With this, I think I would get more performance if I reduce MTU for remote's 
machine if my problem is packet loss, so nfs and kdm connections will perform 
better, but I remember that a long time ago, I did it that and then receive a 
"Fragmented kernel". Maybe playing with rsize/wsize in nfs mounts wolud help. 
I also add security and reduce innecesary traffic in the wireless segment (if 
all is the same network).

Do you think it can be a good solution? is there another one?

Luciano Andino



-------------------------------------------------------
This SF.Net email is sponsored by: SourceForge.net Broadband
Sign-up now for SourceForge Broadband and get the fastest
6.0/768 connection for only $19.95/mo for the first 3 months!
http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click
_____________________________________________________________________
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

Reply via email to