Of course, this is in the end just my personal opinion, and it's mostly
guesswork based off my experience running a similar system in a small
environment. Others may have been much more successful or may think my
estimates are overly optimistic.
--
Craig Ringer
---
S,
http://pxes.sf.net/ , though I haven't tried it myself. LTSP works great
here.
--
Craig Ringer
---
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Win
On Mon, 2004-11-08 at 14:39, robinboby . wrote:
> hai all
>
> if any one have idea how to restart or turn off a thin client form server...
> please share the ideas
Use the local apps mechanism, such that the server runs the 'shutdown'
or 'reboot' command on
hare, you should be able to map a drive using
ancient DOS network drivers and the lanman client. We're talking dark
ages here ;-) . After that, depending on the app it's quite likely to
"just work".
--
Craig Ringer
---
This S
ally the --uid-owner and --gid-owner
restrictions.
That way, even if they do manage to run a the browser, they won't get
very far.
--
Craig Ringer
---
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for
apologies for the duplicate comments. I keep on forgetting that many
mail clients tend to break threading, and really shouldn't rely on it.
--
Craig Ringer
---
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download
gotiate a session, and they behave indistinguishably
thereafter.
--
Craig Ringer
---
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
n for a lot of this network traffic?
They do chat constantly, as mentioned below, but the level of traffic
involved is fairly minimal.
--
Craig Ringer
---
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT product
e
> more ports you must open, consult "man ssh".
Nope, ssh tunnels everything required back through its connection.
--
Craig Ringer
---
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your bu
e
network is pegged at any point, spot odd drop-outs, etc. You do need to
keep careful track of what you did when in the testing session though.
This thread has been broken several times and I didn't have time to dig
back to find it all so I hope I haven't repeated anything.
--
Crai
e terminal may be
running out of memory. Are you using NFS swap?
--
Craig Ringer
---
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your
> Anyone know how to use this video card with 800x600 resolution?
I'd try using the 'vesa' driver. While slow, it'll work with most or all
standards-compliant video cards.
> I`m sorry my bad english =(
And must apologise for my t
> motherboard).
Hmm. Asus onboard NIC an an off-the-shelf consumer parts server. I'd
seriously consider dropping in a better NIC - Intel EtherExpress Pro
10/100 or 10/100/1000 NICs are good choices in my experience.
--
Craig Ringer
-
card, maybe)?
I can't see how, but I've been in the computer industry long enough now
to never, ever, ever say the "impossible" word.
> 3) Could it be some network problem (switches, cables)? Maybe some strange
> packet that makes the server lock?
I'd be very surpris
rmally displayed) rather than a "real" bootloader like GRUB, I can't
help you with that.
--
Craig Ringer
---
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your
couple of 15m cables with a FireWire
repeater in the middle.
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http:
DM on
mandrake by default. Check out /etc/X11/kdm/ or use the KDE control
panel to change the rules about shutdown.
I'm very surprised it permits remote clients to shut down the server.
gdm will let you allow that, but by default only shows the
reboot/shutdown buttons for locally controlle
f them is causing the monitor to work too hard (crap monitors often
go blurry with high vrefresh in my experience). LCD with different
resolutions being chose, causing it to scale with one and not with
another. I guess something could be different in the way they actually
drive the hardware, too, but I
eway, the LTSP server is routing packets between
the LTSP clients and the gateway router, and the gateway router knows to
route packets for 192.168.2.0/24 addresses back to the LTSP server.
Right?
--
Craig Ringer
---
SF.Net email is sponsor
ng IPTables on the server. The /proper/ fix
is to fix your organisation's routing so that the subnet the LTSP
clients are on gets routed properly.
--
Craig Ringer
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk S
mote X. Either ssh in to the client or set the client to
start a shell as well as X.
--
Craig Ringer
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
hcpd.conf .
Personally, I'm just maintaining DHCP, forward and reverse DNS manually
- I don't have enough clients for it to be a significant bother.
--
Craig Ringer
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Med
cond NIC in the server that can be on the LTSP
VLAN.
I could easily be misunderstanding what you want or just wrong, but
hopefully this will point out some possible avenues of investigation.
--
Craig Ringer
IT Manager
POST Newspapers
Western Australia
---
o the trick, as should
`tar` with the -p option.
--
Craig Ringer
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner
rly significant (growing?) interest in issues with scaling
LTSP setups and improving availability. I know there was a fairly big
discussion about it a while ago, but IIRC it was directed toward a
fairly specific purpose rather than general scaling/availability
improvements. Is there a wiki or someth
ld grab a copy of the XF86Config it generates to find out what server
and version it's using?
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the pas
rience, though,
it really makes a difference - even for a normal desktop, not just thin
clients.
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the
or screensavers, though. If you want to do bandwidth
testing, running a screensaver like xflame seems to be a good way to do
it - but I wouldn't recommend it on the network normally.
--
Craig Ringer
---
This SF.Net email is sponsored by OST
ll, that means it's almost certainly not a server RENDER bug, because
the X server doesn't support it. Which is odd - I could've sworn that
the s3 driver in Xorg supported RENDER. Are you using a version 3.3
server or the Xorg server? (your lts.conf would be useful).
--
Craig Ringer
ownload cache with the pre-downloaded packages,
which is a bit clumsy.
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big chan
driver? Knowing
that information, especially when X crashes are involved, will help
anybody who tries to answer your question.
If you have a Debian 3.0 box (or other machine with a pre-xft2/RENDER X
environment) I'd try connecting to that from your thin client
deo - the
SiS cards are NOTHING on a GeForce 4 MX though) so hopefully I'll be
able to upgrade our terminals to not quite so terminally slow video soon
:-)
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the ch
created fairly trivially by bundling up the package cache and
the ltsp utils RPM with mkisofs. OTOH, it'd be nice to have a CD with
all the www.ltsp.org docs, etc on it too.
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you not
mv /opt/ltsp4/pkg_cache/* /opt/ltsp4/pkg_cache_old/
should do it.
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more
hing like
wget -r -A .tgz,.ltsp -l 0 http://www.ltsp.org/ltsp-4.1/
IIRC. I'd try exporting those two environment variables:
export LANG=C
export LC_ALL=C
and then trying ltspadmin again before resorting to wget, though.
--
Craig Ringer
---
make sure you don't have a
firewall active and use netstat to make sure that kdm is actually
listening on the right interface(s) .
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, IT
LC_ALL=C
also seemed to help. Perl utf-8 issues, perhaps?
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one more big change to ann
by the way. It lets me drop my
ugly hacked-in snmp support I added locally to v3.
--
Craig Ringer
---
This SF.Net email is sponsored by OSTG. Have you noticed the changes on
Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
one
ernel might not be a bad plan either, as you'll get a
newer version of the NFS server. I'd also consider upgrading the NFS
userspace (portmap, etc) but this is not as easy.
Out of interest, when the server crashes do you see anything on the
console? Do the keyboar
.
Sounds like you've changed from automounting /home to automounting
/home/$USER . Check /etc/auto.master and look at the automount config
file for /home as referenced there; see if that tells you more.
> Possibly because of thes
for authentication?
Ouch. That is a lot of entries, but I'd hope not enough to cause issues.
I'd be inclined to pursue options like LDAP authentication (which is
IHMO nicer anyway) before thinking about dedicated hardware, though.
--
Craig Ringer
---
signed
in in at present) and helps maintain compatibility of LTSP-build
programs between systems.
I'm no expert on this, of course, but I do tend to think the 'download
precompiled binaries' approach is probably the nicest for your
situation.
--
Craig Ringer
and complex lts.conf, especially between multiple LTSP
servers...".
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
superior on my home machine, largely because the RENDER
acceleration (when enabled) seems to make things quite a bit snappier.
It's certainly not smart to just write off the OSS driver.
--
Craig Ringer
---
This SF.Net email is sp
ing - I'd investigate entirely NFS-less clients first
(if you don't also need full LTSP functionality, that is).
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE develop
highly likely suspect. I've learned never
to rule anything out (and never, ever, ever say "that's impossible")
when working with computers, though.
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
(VNC is poorly unsuited to my
environment, and I find it unpleasant at the best of times), but it's
worth mentioning for those it might suit - such as, perhaps, the OP.
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic
al.com/en/products/Products.asp?DriveID=65 )
and RAID controllers (eg
http://www.3ware.com/products/serial_ata9000.asp ) out there, and not
write off SATA based storage as a serious option.
--
Craig Ringer
---
This SF.Net email is sponsored by B
my impression that it's not something
that LTSP can really make happen - AFAIK it needs to be done at the
level of the X server, X libraries, and client toolkits like gtk and qt.
--
Craig Ringer
---
This SF.Net email is sponsored by BEA
nd from the disks? It'd be
incredibly helpful here, but I've been unable to find anything that can
associate disk I/O with the processes that caused it. I understand it's
not an easy process because of the way the vfs and I/O systems are
designed, but I'm hoping there's something
y making client suspend
possible would make client session migration possible too, but I don't
think we'll be seeing it any time soon.
Of course, as I said I'm no expert...
--
Craig Ringer
---
This SF.Net email is sponsored by BEA W
m really fast (happy at @1600x1200x32bit on a
> > P133) with the open source 2D-only driver
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA
the major factors - and I suspect the only
way to really test that is to try some cards.
The NVidia ones do seem really fast (happy at @1600x1200x32bit on a
P133) with the open source 2D-only driver, and I've had good results
from ATi Rage series
t
parameter in the dhcp configuration? ISC DHCPd, at least, permits the
creation of "groups" of hosts that would make this fairly simple.
--
Craig Ringer
---
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2E
servers in general, but it's
my opinion that those are the points that really matter. LTSP works so
well, with so little fuss, that you can forget it's there and instead
spend your time making sure your users' environment is right for them
without having to worry about how they acc
main challenge will
probably be recompiling LTSP for PPC - you'll probably need a ppc
cross-compiler for i386, and I don't know if the build scripts are set
up to handle cross-compiling (or if the programs build properly with
arch=ppc). Anybody able to answer this, or is the answer "
ith one service per LV (mail, database,
etc) then it wouldn't be so bad. Even so, I'm pretty happy with syncing
the disks and snapshotting the mounted volumes. Remounting read-only
would be another option, but still suffers from problems with things
holding long-life write file handles.
--
On Tue, 2004-07-13 at 14:44, Joshua N Pritikin wrote:
> On Tue, Jul 13, 2004 at 02:12:17PM +0800, Craig Ringer wrote:
> > If you're thinking of running 2.6, I'd stay away from LVM for now. LVM2
> > isn't quite finished yet - for example, snapshot support only ju
o - I think you just need to evaluate your priorities before deciding
on storage technology. For us, hardware SATA RAID with consumer SATA
disks had the advantage by a massive margin, but that won't be true for
everybody. Some will want SCSI / FC, some will want SATA RAID on
"Enterprise"
with LVM, not least because of
the ability to use 'reiserfs_resize' to expand a partition while it's
mounted read/write.
You'll also never understand how you lived without LVM snapshots. They
makes things like mail spool backups so much easier (and l
Perhaps something similar can be done for USB HID devices. I'll have a
play with it if I get a chance.
It always comes down to that, doesn't it...
--
Craig Ringer
---
This SF.Net email sponsored by Black Hat Briefings & Training
o do it.
(I accidentally sent a copy of this before re-subscribing. I'd appreciate
it if the list admin would kindly kill it from the queue next time they
check the held spam^H^H^H^Hmail in the list queue. Sorry.)
--
Craig Ringer
---
I use it here even though we don't have thin clients
deployed (yet, working on it) to control login to the linux server or
the SCO box. Yeah, its cross-platform too 'cos XDMCP's a standard
protocol - really cool.
--
Craig Ringer
GPG Key Fingerprint: AF1C ABFE 7E64 E9C8 FC27 C16E D3C
63 matches
Mail list logo
