Re: [Ltsp-discuss] Ang: Re: Ang: Re: swap -again!

[Jakob Unterwurzacher]

Hi Johan, ok, if you do not get messages like this

> kernel: Out of memory: Kill process 9163 (mysqld) score 511 or sacrifice child

you are not running out of memory.

But this looks like it could be it:
https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/1573478/
And there even seems to be a fix already in xenial-proposed.

Best regards,
Jakob

On 08.05.2016 09:30, Johan Kragsterman wrote:
> 
> Hi,Jakob!
> 
> 
> Thanks for stepping in here...
> 
> 
> 
> -Jakob Unterwurzacher <jakob...@gmail.com> skrev: -
> Till: ltsp-discuss@lists.sourceforge.net
> Från: Jakob Unterwurzacher <jakob...@gmail.com>
> Datum: 2016-05-08 00:23
> Ärende: Re: [Ltsp-discuss] Ang: Re: swap -again!
> 
> Hi Johan, do you see out of memory messages in the kernel log?
> 
> 
> 
> I can't say I do, but these are some msg I got:
> 
> May  7 19:13:48 ltsp102 kernel: [90371.938999] metacity[29159]: segfault at 
> 16b1000 ip 7f2e5728f208 sp 7fff859af250 error 4 in 
> libcairo.so.2.11400.6[7f2e5722c000+10e000]
> May  7 19:13:48 ltsp102 gnome-session-binary[29107]: Unrecoverable failure in 
> required component metacity.desktop
> May  7 19:13:53 ltsp102 kernel: [90377.420913] block nbd9: NBD_DISCONNECT
> May  7 19:13:53 ltsp102 kernel: [90377.422049] block nbd9: Receive control 
> failed (result -32)
> May  7 19:14:19 ltsp102 gnome-session-binary[8610]: Entering running state
> 
> 
> Another strange thing I got is that another machine with the exact same 
> hardware is freezing, so that one needs to be hard-rebooted, while the one I 
> mainly use is rebooting to login screen.
> 
> Regards Johan
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> On Sat, May 7, 2016 at 9:09 PM, Johan Kragsterman 
> <johan.kragster...@capvert.se> wrote:
> 
> Hi, Alkis!
> 
> 
> -Alkis Georgopoulos <alk...@gmail.com> skrev: -
> Till: ltsp-discuss@lists.sourceforge.net
> Från: Alkis Georgopoulos <alk...@gmail.com>
> Datum: 2016-05-07 20:24
> Ärende: Re: [Ltsp-discuss] swap -again!
> 
>  From http://manpages.ubuntu.com/lts.conf:
> 
>  > NBD_SWAP
>  > boolean, default False
>  > Set this to True if you want to turn on NBD swap.
>  > If unspecified, it's automatically enabled for thin clients with
>  > less than 300 MB RAM and for fat clients with less than 800 MB RAM.
> 
> Your clients have more than 800 MB RAM, so they don't automatically get
> a swap unless you specifically set NBD_SWAP=True in lts.conf.
> 
> 
> 
> I already tried that, but it still look like this:
> 
> admin@ltsp102:~$ sudo cat /proc/swaps
> [sudo] password for admin:
> FilenameTypeSizeUsed
> Priority
> admin@ltsp102:~$ free
>   totalusedfree  shared  buff/cache   
> available
> Mem:7582448 1802804 2236020 1461384 3543624 
> 4224156
> Swap: 0   0   0
> admin@ltsp102:~$
> 
> 
> Shouldn't it show something here if swap was configured?
> 
> 
> 
> 
> 
> Note that even if you set a swap, by default it's just 512 MB, so it
> won't save you if you think that your 8 GB of RAM isn't enough and that
> the problem is indeed lack of free memory...
> ...which I doubt. I'd suggest you look elsewhere for fixing the crashes,
> e.g. Xorg/graphics drivers or kernel issues.
> 
> 
> 
> 
> 
> Well, graphics drivers should be fine, since I ran the exact same hardware on 
> 14.04, kernel issues I don't know about.
> 
> But I've seen a discussion about the authfile = /etc/ltsp/nbd-server.allow
> 
> People still complain about this, last post was 16.03.28. This file is 
> refered to by the swap.conf, but doesn't not exist. Of coarse I can create it 
> myself, but I don't actually know what to put in there...
> 
> 
> 
> 
> --
> Alkis Georgopoulos
> LTSP developer
> Professional LTSP support: alk...@gmail.com
> 
> 
> On 06/05/2016 11:37 , Johan Kragsterman wrote:
>>
>> Hi!
>>
>>
>> I tried asking a question about swap the other day, but nobody seems to 
>> answer...
>>
>> So I try again:
>>
>> There seems to be no swap configured for my fat clients, since it looks like 
>> this:
>>
>> admin@ltsp102:~$ sudo cat /proc/swaps
>> Filename
>> admin@ltsp102:~$
>>
>> Type  SizeUsedPriority
>> admin@ltsp102:~$ sudo free
>>totalusedfree  shared  buff/cache   
>> available
>> Mem:7582448 1976300 1930544 1416000 3675604 
>> 4099192
>> Swap: 0   0   0

Re: [Ltsp-discuss] Ang: Re: swap -again!

[Jakob Unterwurzacher]

Hi Johan, do you see out of memory messages in the kernel log?

On Sat, May 7, 2016 at 9:09 PM, Johan Kragsterman <
johan.kragster...@capvert.se> wrote:

>
> Hi, Alkis!
>
>
> -Alkis Georgopoulos  skrev: -
> Till: ltsp-discuss@lists.sourceforge.net
> Från: Alkis Georgopoulos 
> Datum: 2016-05-07 20:24
> Ärende: Re: [Ltsp-discuss] swap -again!
>
>  From http://manpages.ubuntu.com/lts.conf:
>
>  > NBD_SWAP
>  > boolean, default False
>  > Set this to True if you want to turn on NBD swap.
>  > If unspecified, it's automatically enabled for thin clients with
>  > less than 300 MB RAM and for fat clients with less than 800 MB RAM.
>
> Your clients have more than 800 MB RAM, so they don't automatically get
> a swap unless you specifically set NBD_SWAP=True in lts.conf.
>
>
>
> I already tried that, but it still look like this:
>
> admin@ltsp102:~$ sudo cat /proc/swaps
> [sudo] password for admin:
> FilenameTypeSizeUsed
> Priority
> admin@ltsp102:~$ free
>   totalusedfree  shared  buff/cache
>  available
> Mem:7582448 1802804 2236020 1461384 3543624
>  4224156
> Swap: 0   0   0
> admin@ltsp102:~$
>
>
> Shouldn't it show something here if swap was configured?
>
>
>
>
>
> Note that even if you set a swap, by default it's just 512 MB, so it
> won't save you if you think that your 8 GB of RAM isn't enough and that
> the problem is indeed lack of free memory...
> ...which I doubt. I'd suggest you look elsewhere for fixing the crashes,
> e.g. Xorg/graphics drivers or kernel issues.
>
>
>
>
>
> Well, graphics drivers should be fine, since I ran the exact same hardware
> on 14.04, kernel issues I don't know about.
>
> But I've seen a discussion about the authfile = /etc/ltsp/nbd-server.allow
>
> People still complain about this, last post was 16.03.28. This file is
> refered to by the swap.conf, but doesn't not exist. Of coarse I can create
> it myself, but I don't actually know what to put in there...
>
>
>
>
> --
> Alkis Georgopoulos
> LTSP developer
> Professional LTSP support: alk...@gmail.com
>
>
> On 06/05/2016 11:37 , Johan Kragsterman wrote:
> >
> > Hi!
> >
> >
> > I tried asking a question about swap the other day, but nobody seems to
> answer...
> >
> > So I try again:
> >
> > There seems to be no swap configured for my fat clients, since it looks
> like this:
> >
> > admin@ltsp102:~$ sudo cat /proc/swaps
> > Filename
> > admin@ltsp102:~$
> >
> > Type  SizeUsedPriority
> > admin@ltsp102:~$ sudo free
> >totalusedfree  shared  buff/cache
> available
> > Mem:7582448 1976300 1930544 1416000 3675604
> 4099192
> > Swap: 0   0   0
> > admin@ltsp102:~$
> >
> >
> > nbd-server is up and running, connections are established:
> >
> > admin@ltsp102:~$ netstat -tu | grep nbd
> > tcp0  0 192.168.20.102:55232server:nbd
>  ESTABLISHED
> > tcp0  0 192.168.20.102:33850server:nbd
>  CLOSE_WAIT
> > tcp0  0 192.168.20.102:55366server:nbd
>  CLOSE_WAIT
> >
> >
> > System is an LTSP server KVM virtual machine on OmniOS with enough
> resources, and the fat clients are quad core celerons with 8 GB DDR3
> memory. 8 GB is of coarse a lot, but there seem to be memory loss that
> causes reboot to login screen, or get the entire client to freeze, which
> forces a hard reboot.
> >
> > I checked some other resources, if I should configure swap through
> ldm.conf, but since I didn't need that in 14.04, I don't understand why I
> should need it now...? There is no ldm.conf file in
> /var/lib/tftboot/"myclient", and it was not in 14.04, so this must all be
> default then.
> >
> > Only thing I changed since 14.04 is that I now use another default GW, I
> don't go through the LTSP server anymore. But the server responds to the
> "server" ping in the client network, so I can't see that as a problem.
> >
> > So I would really appreciate some input here...
> >
> >
> > Best regards from/Med vänliga hälsningar från
> >
> > Johan Kragsterman
> >
> > Capvert
> >
>
>
>
> --
> Find and fix application performance issues faster with Applications
> Manager
> Applications Manager provides deep performance insights into multiple
> tiers of
> your business applications. It resolves application problems quickly and
> reduces your MTTR. Get your free trial!
> https://ad.doubleclick.net/ddm/clk/302982198;130105516;z
> _
> 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
>
>
>
>
> --
> 

Re: [Ltsp-discuss] Poor performance from LTSP

[Jakob Unterwurzacher]

How do you start the rdp connection? Like this
http://wiki.ltsp.org/wiki/Integration#Rdesktop ?

On Tue, Aug 18, 2015 at 1:24 PM, Funke, Martin m.fu...@olpe.de wrote:

 Hi everyone,



 im using a Xubuntu 14.04 with ltsp 5.5.1. Every client which works with
 the ltsp starts right away with a xfreerdp connection to a windows machine.



 I noticed that the scrolling in webbrowsing is working poor. Everything is
 a bit slow.



 When I connect from a regular Xununtu 14.04 PC I don’t have this poor
 performance in scrolling on the windows Machine.



 Does anyone know why this Is happening?



 Best regards

 Martin



 --

 _
 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


--
_
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

Re: [Ltsp-discuss] LTSP PXE client PC stalling @ Ubuntu splash screen

[Jakob Unterwurzacher]

On 21.09.2014 16:42, Denis Croombs wrote:
 Thanks for the reply
 We do not get the login just the splash screen, will try to disable the 
 splash screen tomorrow, thanks for the advise

Looks a lot like:
https://bugs.launchpad.net/ubuntu/trusty/+source/linux/+bug/1330671

Does

# service nbd-server restart

fix the issue temporarily? (Warning: this may reboot all thin clients).

Best regards,
Jakob

--
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports
Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper
Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer
http://pubads.g.doubleclick.net/gampad/clk?id=154622311iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] [PATCH 1/1] Bye bye No response from server - add ssh error logging

[Jakob Unterwurzacher]

 On 15.09.2014 17:30, Vagrant Cascadian wrote:
 We unfortunately don't have a lot of active developers who understand a
 patch of this size; ldm has largely been in maintenance mode for quite
 some time.

Hi Vagrant,

I looked at that thing again. Most of the complexity in the ssh handling
seems to come from the expired password handling.
However, it's broken and deemed unfixable [1, 2].
Ok to rip it out? That will simplify things a lot.

Best regards,
Jakob


[1] https://bugs.launchpad.net/ltsp/+bug/1053447
[2]
https://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg40865.html

--
Slashdot TV.  Video for Nerds.  Stuff that Matters.
http://pubads.g.doubleclick.net/gampad/clk?id=160591471iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] [PATCH 1/1] Bye bye No response from server - add ssh error logging

[Jakob Unterwurzacher]

On 15.09.2014 17:30, Vagrant Cascadian wrote:
 On 2014-09-14 11:16:52 -0500, Jakob Unterwurzacher wrote:
 ldm now reads the full output from ssh even if it dies on the way.
 Proper error messages are logged and displayed, wrapped in _() so
 they can be localized.
 
 Thanks for the (updated) patch!  It's an issue that's long been a thorn
 in the side of many users. Apologies for not getting back to you sooner
 on this.
 
 We unfortunately don't have a lot of active developers who understand a
 patch of this size; ldm has largely been in maintenance mode for quite
 some time.
 
 Would it be possible to usefully break this into maybe 2-4 smaller
 patches, and submit to the (old) launchpad bug? That might make it a
 little easier to review.

Hi Vagrant!

I already suspected it might be a bit big. I will try to split it up
into a few logical steps.

 * ssh: connect to host server port 22: Connection timed out
   For that to work, ConnectTimeout=10 is set on the ssh command line.
   Otherwise the 30 second timeout elapses before ssh tells us what is
   going on.
 
 What controls the 30 second timeout? Where exactly does it occur?

The thirty seconds are passed in from ssh_chat():

ssh.c:427: seen = expect(fd, lastseen, 30, SENTINEL, : , NULL);

That is, expect gives up after 30 seconds if it cannot find a prompt.

When the server cannot be reached, ssh by default takes longer than that
to timeout, so we would never get the error message.

 The last time we tried parsing the output like this, it turned out that
 the messages returned via ssh were highly variable depending on distro
 and local configuration. Maybe your code handles this no worse than the
 current code; I can't really say.

Parsing human-readable output is generally fragile, I understand the
concern. However, I am not actually interpreting the output, I am just
passing it on to the user in full. That is, only if a password prompt
cannot be found.
The logic for detecting the password prompt has not changed - a : at
the end of output is a prompt.

I'll take a look if the prompt can look differently, and of course I
will be there to fix any issues that may arise.

Best regards,
Jakob

--
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191iu=/4140/ostg.clktrk
_
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

[Ltsp-discuss] [PATCH 0/1] Bye bye No response from server - add ssh error logging

[Jakob Unterwurzacher]

Dear team,

I first posted the patch to add ssh error logging to ldm a while ago to
https://bugs.launchpad.net/ltsp/+bug/1325388 .

It has bit-rotted there for three months without feedback. Due to
whitespace changes it does not apply any more.

I have now un-bit-rotted and tested again on Ubuntu 14.04. It works fine.

I am posting the patch in a follow-up email in the hope to get it merged,
to have ldm finally produce useful error messages on login failure.

Best regards,
Jakob

Jakob Unterwurzacher (1):
  Bye bye No response from server - add ssh error logging

 src/logging.c |  12 +++--
 src/plugins/ssh/ssh.c | 124 +++---
 2 files changed, 95 insertions(+), 41 deletions(-)

-- 
1.9.3


--
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191iu=/4140/ostg.clktrk
_
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

[Ltsp-discuss] [PATCH 1/1] Bye bye No response from server - add ssh error logging

[Jakob Unterwurzacher]

ldm now reads the full output from ssh even if it dies on the way.
Proper error messages are logged and displayed, wrapped in _() so
they can be localized.

Messages we get:
* Permission denied (publickey,password).
  The common case of a password typo or invalid user.
  We make sure we read all the output BEFORE checking if the child has
  died. Also, we handle EINTR, as this is what select() returns when it
  gets SIGCHLD.

* ssh: connect to host server port 22: Connection timed out
  For that to work, ConnectTimeout=10 is set on the ssh command line.
  Otherwise the 30 second timeout elapses before ssh tells us what is
  going on.

* ssh: connect to host server port 22: Connection refused
  That message contains a :  like the password prompt. This is solved
  by waiting for 200ms of silence before searching for the prompt (polling
  interval reduced from 1s to 200ms to not bore the user).
  In the error case, ssh will have exited in the meantime, and we know
  for sure that it was NOT a password prompt.
---
 src/logging.c |  12 +++--
 src/plugins/ssh/ssh.c | 124 +++---
 2 files changed, 95 insertions(+), 41 deletions(-)

diff --git a/src/logging.c b/src/logging.c
index 63389dd..d93ff39 100644
--- a/src/logging.c
+++ b/src/logging.c
@@ -29,10 +29,14 @@
 #include logging.h
 
 FILE *logfile = NULL;
-char *LOG_LEVEL_NAMES[] = { EMERGENCY, ALERT, CRITICAL, ERROR,
-WARNING, NOTICE, INFO, DEBUG
-};
-
+char *LOG_LEVEL_NAMES[] = { EMERGENCY, /* 0 */
+ALERT, /* 1 */
+CRITICAL,  /* 2 */
+ERROR, /* 3 */
+WARNING,   /* 4 */
+NOTICE,/* 5 */
+INFO,  /* 6 */
+DEBUG };   /* 7 */
 int loglevel = 0;
 
 /*
diff --git a/src/plugins/ssh/ssh.c b/src/plugins/ssh/ssh.c
index 0110937..689ec43 100644
--- a/src/plugins/ssh/ssh.c
+++ b/src/plugins/ssh/ssh.c
@@ -16,6 +16,7 @@
 #include sys/ioctl.h
 #include sys/stat.h
 #include utmp.h
+#include errno.h
 #include crypt.h
 
 #include ../../ldm.h
@@ -289,7 +290,7 @@ expect(int fd, char *p, int seconds, ...)
 char buffer[BUFSIZ];
 gchar *arg;
 GPtrArray *expects;
-int loopcount = seconds;
+int loopcount = seconds * 5;
 int loopend = 0;
 
 bzero(p, MAXEXP);
@@ -313,58 +314,87 @@ expect(int fd, char *p, int seconds, ...)
  * Main loop.
  */
 
+log_entry(ssh,7,expect: entering main loop);
+
 while (1) {
-timeout.tv_sec = (long) 1;   /* one second timeout */
-timeout.tv_usec = 0;
+timeout.tv_sec  = 0;
+timeout.tv_usec = 20; /* 200ms timeout */
 
 FD_ZERO(set);
 FD_SET(fd, set);
 st = select(FD_SETSIZE, set, NULL, NULL, timeout);
 
-if (child_exited) {
-break;   /* someone died on us */
-}
-
-if (st  0) {/* bad thing */
-break;
-}
-
-if (loopcount == 0) {
-break;
-}
+log_entry(ssh, 7, expect: select returned %d, st);
 
-if (!st) {   /* timeout */
-loopcount--; /* We've not seen the data we 
want */
-continue;
-}
+/* There is data available - we want to read it in even if
+ * the child process has exited - we want to get the error
+ * message
+ */
+if (st  0)
+{
+size = read(fd, buffer, sizeof buffer);
+if (size == -1) {
+log_entry(ssh, 3, expect: read returned error %d (%s), 
errno, strerror(errno));
+break;
+}
+if (size == 0) {
+log_entry(ssh, 3, expect: read returned 0 (EOF)));
+break;
+}
 
-size = read(fd, buffer, sizeof buffer);
-if (size = 0) {
-break;
+if ((total + size)  MAXEXP) {
+strncpy(p + total, buffer, size);
+total += size;
+log_entry(ssh, 7, expect: got %d bytes so far, total);
+}
+else
+{
+log_entry(ssh, 3, expect: buffer full);
+return ERROR;
+}
 }
+else /* No data available */
+{
+/* select() was interrupted by a signal,
+ * most likely SIGCHLD. Just try again */
+if (st == -1  errno == EINTR)
+{
+log_entry(ssh, 7, expect: retrying);
+continue;
+}
+else if (st == -1) {/* other errors */
+log_entry(ssh, 3, expect: select returned error %d (%s), 
errno, strerror(errno));
+break;
+}
+/* st == 0 and there was no data for 200ms */
 
-if 

Re: [Ltsp-discuss] Ubuntu 14.04 LTSP now working?

[Jakob Unterwurzacher]

On 17.06.2014 04:29, Todd O'Bryan wrote:
 Fix released an hour after you posted the bug. Now that's what I call
 service. (Which is nice if you provide your own patch.)

Yes, the response time was awesome!

Note that the updated kernel for Trusty has not been released yet, it is
in Fix Committed state.
Subscribe to the launchpad bug to get notified when it is released (
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1330671 ).


Best regards,
Jakob

--
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing  Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_
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

Re: [Ltsp-discuss] Ubuntu 14.04 LTSP now working?

[Jakob Unterwurzacher]

On 12.06.2014 07:33, Funke, Martin wrote:
 Hi many thanks for your information. For the time being I will stick with 
 Ubuntu 12.04.

FYI, I reported the issue on launchpad, maybe this gets things moving:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1330671

Jakob

--
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing  Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_
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

Re: [Ltsp-discuss] Ubuntu 14.04 LTSP now working?

[Jakob Unterwurzacher]

On 10.06.2014 13:39, Fabrizio Carrai wrote:
 Hello Martin,
 I tryed LTSP with 14.04 (VirtualBox, to be precise) and it seems working
 fine. Since I'm going to present it, I wonder if there are hidden
 problems I didn't face yet:
 Which kernel did you use ?


So I guess we are talking about this problem

https://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg41773.html

this is the fix

net: avoid dependency of net_get_random_once on nop patching
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=3d4405226d27b3a215e4d03cfa51f536244e5de7
.

Judging from

http://kernel.ubuntu.com/git?p=ubuntu%2Fubuntu-trusty.gita=searchh=HEADst=commits=avoid+dependency+of+net_get_random_once+on+nop+patching
:

Nope, not fixed yet.

Also, I experienced that issue on real thin clients, but could not
reproduce it under VirtualBox. I have no idea why.

Best regards,
Jakob

--
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing  Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_
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

Re: [Ltsp-discuss] LTSP kernel image permissions with Ubuntu 14.04

[Jakob Unterwurzacher]

On 28.05.2014 09:48, Alkis Georgopoulos wrote:
 Στις 26/05/2014 03:12 μμ, ο/η Derek Schuurman έγραψε:
 Just a quick update regarding the file permissions issue I encountered
 in /var/lib/tftpboot/ltsp/amd64/ -- I tracked the issue down to a
 modified umask setting in /etc/login.defs. The umask settings were
 changed to 027 so that the default file permissions prevent users from
 being able to read files in other peoples home folders. Unfortunately
 this had the undesired side-effect that the default file permissions of
 the image files created by ltsp-update-image were no longer readable
 either...

 The file permissions of the vmlinuz boot image file should always be
 world-readable regardless of any changes in login.defs, no?  It seems
 this problem could be averted in the future by simply adding a line to
 the generate_image function in ltsp-update-image script to chmod o+r the
 image file... Is this a change that might be useful to submit?

 
 
 No, chmod wouldn't be useful,
 but setting umask to an appropriate value at the start of any LTSP 
 script that creates files (maybe in ltsp-server-common-functions) is a 
 topic that should be discussed by LTSP developers...
 
 So, file a bug report in upstream LTSP about setting umask:
 https://bugs.launchpad.net/ltsp/

I already did a week ago: https://bugs.launchpad.net/ltsp/+bug/1320994
Please click Affects me too so the bug can go to confirmed!

Best regards,
Jakob

--
Time is money. Stop wasting it! Get your web API in 5 minutes.
www.restlet.com/download
http://p.sf.net/sfu/restlet
_
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

Re: [Ltsp-discuss] 1st boot - OK, 2nd boot - FAIL!?!?!

[Jakob Unterwurzacher]

Wow, this is quite bad. Thank you Veli-Matti.
Workaround: Use the kernel from Precise in the chroot:
http://packages.ubuntu.com/precise/i386/linux-image-3.11.0-13-generic/download

Jakob

On Tue, May 13, 2014 at 4:11 PM, Veli-Matti Lintu 
veli-matti.li...@opinsys.fi wrote:

 Hello,

 We came across the same symptoms in our systems with kernel 3.13 or newer.
 Also our Ubuntu 12.10 based systems failed the same way when kernel 3.13
 was used. Restarting nbd-server or killing the nbd-server process serving
 the client always fixed the problem.

 We found out that there was a change in kernel 3.13 that caused the
 nbd-client to always use the same source port number when connecting the
 nbd-server. As dhcpd usually gives out the same ip for the same client, the
 nbd tcp connection had always the same srcip:srcport - dstip:dstport
 pairs. As the nbd-server had the client's connection still open, it did not
 accept a new connection from the same port.

 Bisecting revealed that commit e34c9a69970d8664a36b46e6445a7cc879111cfd
 broke it (http://marc.info/?l=git-commits-headm=138433307618907w=2) and
 the fix is now here on its way to stable kernels too:
 http://www.spinics.net/lists/netdev/msg282079.html

 So if you are experiencing the same bug, a patch should be on its way.


 Happy hacking,

 Veli-Matti


 - Original Message -
 From: Jacob Tennant jacob.tenn...@pierpont.edu
 To: LTSP list ltsp-discuss@lists.sourceforge.net
 Sent: Monday, May 5, 2014 3:36:03 PM
 Subject: [Ltsp-discuss] 1st boot - OK, 2nd boot - FAIL!?!?!

 I just booted up my first client on a fresh LTSP install on Ubuntu Server
 14.04LTS and it worked.

 However when I shutdown the thin-client machine and restarted it,
 everything went as before except I got the following error;;;

 Error: Socket failed: Connection timed out
 Exiting.
 mount: mounting /dev/nbd0 on /root failed: Invlai argument
 /scripts/init-bottom/ltsp: line 27: panic: not found
 chroot can't execute 'usr/bin/test': No such file or directory
 mount: mounting /root on rofs failed: Invalid argument
 mount: mounting /root/rofs failed: Invalid argument
 mount: mounting /dev on /root/dev failed: No such file or directory
 mount: mounting /sys on /root/sys failed: No such file or directory
 mount: mounting /proc on /root/proc failed:No such file or directory
 Target filesystem doesn't have requested /sbin/init-ltsp
 No init found. Try passing init bootarg

 BusyBox v1.21.1 (Ubuntu 1:1.21.0-1ubuntu1) built-in shell (ash)
 Enter 'help' for a list of built-in commands.

 (initramfs)

 I have tried to reboot the thin-client 3 times and get the same error
 everytime.

 Any suggestions?

 --


 Jacob W. Tennant






 IT Specialist

 Pierpont Community  Technical College

 Office: 304-333-3641

 jacob.tenn...@pierpont.edu




 Confidentiality Notice

 The information contained in this e-mail and any attachments may be
 legally privileged and confidential. If you are not an intended recipient,
 you are hereby notified that any dissemination, distribution or copying of
 this e-mail is strictly prohibited. If you have received this e-mail in
 error, please notify the sender and permanently delete the e-mail and any
 attachments immediately. You should not retain, copy or use this e-mail or
 any attachment for any purpose, nor disclose all or any part of the
 contents to any other person. Thank you!


 --
 Is your legacy SCM system holding you back? Join Perforce May 7 to find
 out:
 #149; 3 signs your SCM is hindering your productivity
 #149; Requirements for releasing software faster
 #149; Expert tips and advice for migrating your SCM now
 http://p.sf.net/sfu/perforce
 _
 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


 --
 Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
 Instantly run your Selenium tests across 300+ browser/OS combos.
 Get unparalleled scalability from the best Selenium testing platform
 available
 Simple to use. Nothing to install. Get started now for free.
 http://p.sf.net/sfu/SauceLabs
 _
 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

--
Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled scalability from the best Selenium testing platform available
Simple to use. Nothing to 

Re: [Ltsp-discuss] Server Benchmarks?

[Jakob Unterwurzacher]

On 05.02.2014 19:11, Jeff Siddall wrote:
 On 02/05/2014 11:58 AM, Steffen Schaumburg wrote:
 Since the i7 (almost certainly) cannot have ECC memory I'd certainly
 think twice about using that for a server, or for anything else that's
 important for that matter. Google published a study into memory
 reliability and.. well, it's not pretty. And they tested server-grade
 ECC memory, which is presumably higher quality than non-ECC sticks... Go
 for Xeons, Opterons or AM2/3/3+ Athlon/Phenom with a suitable board
 (afaik only Asus) to get ECC support. The extra cost of ECC over non-ECC
 is almost nothing, but unfortunately Intel cripples the i* CPUs to
 remove this feature, so that you have to get a Xeon instead.
 
 Yeah, good point.  All my servers are Phenom+Asus+ECC.  Honestly it is 
 extremely rare (once I can think of) that it reported a corrected memory 
 error but it's yet another thing that could go wrong if you don't have it.

I have had bad memory corrupt the squashfs image, causing all kinds of
funny errors. The problem is that memory corruption not only causes
instability, but will also cause disk corruption, as every write goes
through memory first. As the LTSP server is de-facto also a file server,
that was not pretty.
I have yet to upgrade to ECC, but I wrote cshatag
( https://github.com/rfjakob/cshatag )
and let it run every night so memory corruption (which will cause disk
reads to be corrupted, too) is spotted early.

Best regards,
Jakob

--
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121051231iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] flow control revisited

[Jakob Unterwurzacher]

On 03.01.2014 12:24, E Kogler wrote:
 My option would be to force the gigabitport to 100MBit :-)
 Edgar



 The issue here is that the buffer on the switch fills from data from
 its giga port
 while data is more slowly released through its 100 Mbps ports. What is
 hoped is that
 with flow control disabled the server will slow down (or pause for less
 time) to
 allow the switch's buffer time to empty enough to accommodate new data
 arriving.

Yes, the server will slow down *just this client* instead of all traffic.
Now, how does the server know which client to slow down? Because packets 
destined to this client are lost! This is perfectly okay. TCP will 
notice, retransmit the packet, and adjust the transmit speed until no 
packets are lost anymore. Losing those packets enable TCP to autodetect 
the maximum transmit speed - this is also how the internet works.

With the server on gigabit and flow control disabled, you can have ten 
clients receiving 100Mb/s *each*.
Forcing the server to 100Mb... Well, a bit less.

Jakob

--
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET,  PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] flow control revisited

[Jakob Unterwurzacher]

On 29.12.2013 01:21, rkwesk_ltsp wrote:

 As the switch's port to the client is also 100 Mps I think the client's
 100mps nic cannot be overloaded
 per se. However, a fellow member of this list, alkisg, has since
 explained to me that the buffer on the
 switch will fill up since it is receiving packets from its giga port
 but can only release them at its
 100 mbps port. It was this phenomenon (is it called buffer overrun?)
 that I hadn't thought about. This
 then forces the switch to send a pause frame to the server (in order
 for its buffer to be relieved.)

Yes, exactly.
I wouldn't call it buffer overrun - that would imply a programming bug. 
The buffer just gets full.

 Now I will show complete ignorance by asking whether if in fact the
 overload occurs even sooner (or more often)
 if an all giga switch were in place of the above described switch and
 the bottleneck was the 100 mbps nic rather
 than the 100 mbps port on the above switch?

Whether it is a 100Mb port or a gigabit port operating at 100Mb does not 
make a difference.
The bottleneck will be the switch port (not the NIC!) that cannot get 
the data out as fast as the gigabit port gets new data in.


Jakob

--
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET,  PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] flow control revisited

[Jakob Unterwurzacher]

On 28.12.2013 12:18, rkwesk_ltsp wrote:
 I thought I had this figured out but I'd like to confirm:

 Configuration 1

 unmanaged switch w 1 giga port and 16 100 ports not connected directly
 w router
 Server with two nics, one gigabit to giga port on switch and one 100
 bit to router directly.
 Clients, mixed thin and fat but all with 100 bit nics to same switch to
 100 bit ports.

 Configuration 2

 unmanaged switch w 1 giga port and 16 100 ports, one of 100 bit ports
 connected w router
 Server with one nic (giga) to giga port on switch
 Clients, mixed thin and fat but all with 100 bit nics to same switch to
 100 bit ports.

 In both cases (I think) there should be no overload and so the possible
 complications with flow control are avoided.

 Is this correct?

No, I think you will experience head-of-line blocking if the switch has 
flow control enabled.

Most of the traffic in LTSP is from the server to the clients. Say a 
client watches an HD youtube video. This will easily generate more than 
100Mb/s of traffic and will overload the client's 100Mb/s port. The 
switch will send pause frames to the server, effectively throttling down 
it's 1Gb/s port to a slower speed. This will unnecessarily
affect all other clients' traffic, too - this is what is called 
head-of-line blocking.

That article explains it quite well (just multiply the speeds by a 
factor of 10 ;) )
http://virtualthreads.blogspot.co.at/2006/02/beware-ethernet-flow-control.html


Jakob

--
Rapidly troubleshoot problems before they affect your business. Most IT 
organizations don't have a clear picture of how application performance 
affects their revenue. With AppDynamics, you get 100% visibility into your 
Java,.NET,  PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] No ssh into clients due to missing client host keys

[Jakob Unterwurzacher]

On 24.07.2013 07:22, Alkis Georgopoulos wrote:
 If my assumption above is true, that a leaked ssh server private key
 means that the ssh connections are no longer private, then this applies
 to key-based authentication as well.


Ok, thanks for the clarification!

After the Debian weak key issue (the private keys were predictable), 
they did warn about the issue you describe:

http://wiki.debian.org/SSLkeys#End_User_Summary
 Note that this last point means that passwords transmitted over ssh to
 a server with a weak dsa server key could be compromised too
[...]
 This is due to an 'attack' on DSA that allows the secret key to be
 found if the nonce used in the signature is known or reused.

However, this seems to be limited to DSA keys. These are no longer used 
nowadays because of security issues like this.

With RSA, the sessions should be protected using a random per-session 
key exchanged using diffie-hellmann that does not depend on the private 
key for its security.

I will try to find a definitive source for that and follow-up here.

 One other possible implementation that I was thinking of, would be to
 use some values that are unique for each client as the random seed for
 the ssh private keys.
 For example, the client MAC, its output of `dmidecode` etc.
 Then the LTSP client ssh server host keys would always be the same, and
 the sysadmin would only need to trust them once.

This is a very good idea! Has to be done carefully, though, to make the 
key unpredictable enough.

Best regards,
Jakob

--
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] No ssh into clients due to missing client host keys

[Jakob Unterwurzacher]

On 24.07.2013 09:11, Jakob Unterwurzacher wrote:
 With RSA, the sessions should be protected using a random per-session
  key exchanged using diffie-hellmann that does not depend on the
 private key for its security.

 I will try to find a definitive source for that and follow-up here.

Found it. The buzzword is perfect forward secrecy (PFS), and SSH - 
both version 1 and 2 - have it.

It's explained very nicely in [1] and [2], but for
the definitive source I'm just gonna quote the sshd_config manpage:

 KeyRegenerationInterval
 In protocol version 1, the ephemeral server
 key is automatically regenerated after this many seconds (if it has
 been used). The purpose of regeneration is to prevent decrypting
 captured sessions by later breaking into the machine and stealing the
 keys. The key is never stored anywhere.

SSHv2 does it differently but achieves the same goal.


Conclusion:
SSH to thin clients is secure against passive eavesdropping, even if the 
host key is shared.
However, do NOT use a password to log into the thin client, this is 
insecure!


Alkis, I'd suggest to not delete the host key, instead to make sure that 
passwords are not allowed in sshd_config, i.e.

PasswordAuthentication no


Best regards,
Jakob


References:
[1] http://utcc.utoronto.ca/~cks/space/blog/tech/SshForwardSecrecy
[2] http://docstore.mik.ua/orelly/networking_2ndEd/ssh/ch03_05.htm

--
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] No ssh into clients due to missing client host keys

[Jakob Unterwurzacher]

On 21.07.2013 22:51, Alkis Georgopoulos wrote:
 Στις 21/07/2013 11:02 μμ, ο/η Rüdiger Kupper έγραψε:
 Now given this change, and accepted that it is a security measure,
 let me rephrase my question:

 - Since ssh login to running clients is a security risk, what
 other measure can I take to allow remote shutdown of a running
 client?

 If you run ssh-keygen on client boot, I don't think there are security
 issues anymore.

This means that you won't know the client ssh host key. That's not 
better than a publicly-known one - either way you can't be sure the box 
you are connecting to is not spoofed.

But then, if you want to shut the thing down, who cares?

Make sure you are using key-based authentification to log into the thin 
clients. Never use passwords in the thin client - its /etc/shadow is 
also publicly viewable, and can be used to crack your password. This 
could then be used to log into thin clients and install keyloggers etc.

If done that way I don't see any security issues, even with shared ssh 
host keys.

Jakob

--
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831iu=/4140/ostg.clktrk
_
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

Re: [Ltsp-discuss] Restricting Internet Access

[Jakob Unterwurzacher]

On 04.07.2013 14:46, David Burleigh wrote:
 Thank you, Rolf. I have tried a few settings with iptables, but I must be
 doing something wrong because it has the effect of blocking the desktop
 environment from being set up on the client after login. I'm sure it's because
 I really don't understand iptables properly. Perhaps its a matter of which
 network interface is being controlled. My client-side interface is eth1
 (192.168.0.1), while the router side is eth0 (192.168.1.1).

 On Thursday, July 04, 2013 01:10:59 PM Sprachschule Eilert wrote:

Wow, thanks Alikis, I didn't know this was possible with iptables!

David, I think the command for blocking internet access for student123 
should look like this:

iptables -A OUTPUT -o eth0 -m owner --uid-owner student123 -j REJECT

This should not interfere with traffic to/from the thin clients (eth1 in 
your case) as far as I can see.

--
This SF.net email is sponsored by Windows:

Build for Windows Store.

http://p.sf.net/sfu/windows-dev2dev
_
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

Re: [Ltsp-discuss] How is graphics divided between server and thin client?

[Jakob Unterwurzacher]

On 05.03.2013 22:56, Lachele Foley wrote:
 I read the Wikipedia article and I searched virtualgl ltsp.  It
 looks good!  I believe that virtualgl is not being used in ltsp.  Am I
 correct?

 Is this something a mere user like me can add?  (I'm a chemist, and I
 can spell computer graphics.)  If not, is it likely to be added by
 someone more qualified any time soon?  Can I help accelerate that
 process?

It's not included in LTSP, no. But in theory, it should be entirely 
transparent to the X Server, so LTSP doesn't even have to know about it. 
I have no first-hand experience, and I don't know what distro you are 
running, but here
http://comments.gmane.org/gmane.comp.video.opengl.virtualgl.user/771
somebody got it working on Ubuntu 12.04 (after fixing some command line 
arguments, in a setup without LTSP).

I would
1) Install virtualgl on the server, restart
2) Try it using a normal PC (that is, no LTSP thin client) as the
virtualgl client as explained here
http://www.virtualgl.org/vgldoc/2_2_1/#hd007
3) Try it from a thin client (use a terminal window that was started on
the thin client as a localapp, i.e. ltsp-localapps xterm ). If you
are lucky, it will just work.

Do report back if you have any results!

Best regards,
Jakob

--
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and remains a good choice in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev
_
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

Re: [Ltsp-discuss] How is graphics divided between server and thin client?

[Jakob Unterwurzacher]

Hi Lachele!

Opengl commands are sent over the network and executed by the client. Even
on gigabit, the latency is probably very high compared to local execution.

Virtualgl promises to solve that problem - and its wikipedia article is a
good read on the topic!

Best regards,
Jakob
Am 05.03.2013 17:00 schrieb Lachele Foley lf.l...@gmail.com:

 In a response to Jim a day or so ago, I gave reasons why localapps
 can't solve all our problems.  To recap that, a brief example:  The
 programs that generate our graphics can, for the large systems we
 treat, fill the 3-4 GB of RAM available on a client.  While we could
 potentially move some things to our clients, that won't solve all our
 troubles.  I used the video (youtube) example because it is something
 simple that most people do every day.  We are actually pretty
 unconcerned about being able to view videos over the internet.

 You mention network latency...  I might not have the tools I need to
 properly assess things, but if I open atop on both client and server
 and run an extended part of one of our benchmarks (not a video), the
 ethernet load remains solidly below 10%.  So, for our normal
 benchmark, best I can tell, ethernet is not an issue.

 Even if I stray from our usual benchmarking and give it a more serious
 workout (same software and system), the load only pushes 85%.  In that
 situation, the client has one CPU running at 50% load, and the server
 reports the process as taking about 32% of a CPU.  Other info from
 atop:  packets from server to client: 710,000-720,000.  From client
 back to server:  170,000-175,000.  The speed between them is ~0.85
 Gbps (hence, I presume, the 85% load).  The default Ubuntu atop
 install doesn't come with a kernel patch for networking, so that's
 all I can get without a custom install.

 If I start a full-screen video on hulu (filling as much as it can of a
 wide format 24 inch monitor), then the ethernet load goes up to 90+%.
 I can accept that streaming video is more/different work than
 rendering molecular graphics.  I just know we need something better
 than what we have.



 On Tue, Mar 5, 2013 at 8:01 AM, Richard Kweskin rkwesk_l...@hellug.gr
 wrote:
  Quoting James McQuillan:
 
  Lachele,
 
  I'll try and explain how the graphics works in a thin client
 environment.
 
  snip
 
  I've been focusing most of my rather limited time on running things on
 the
  thin client.  To me, everything is moving towards being web-based and
 the
  thin client is becoming a platform for running a web browser.  Running
 the
  browser locally makes some things much better.  Sound and video perform
  very well this way. But,  it brings new challenges, like printing.  If
 run
  the browser on the thin client, you need to also have a printing
 sub-system
  installed like CUPS.
  Also, where's the users home directory?  Probably on the server, which
  means you need to use NFS or sshfs to have access to it.
 
  I should note that what I've just described in the last paragraph is
 really
  just a fat client.
 
  If you've got powerful enough thin clients, moving the processing down
 to
  the thin client can really boost the number of thin clients that a
 single
  server can support.  basically, the server is just a boot server. A
 decent
  server should be able to handle hundreds of thin clients.
 
  I hope my rather long-winded explanation helps.
 
  Jim McQuillan
 
  What if these powerful clients ran the browsers locally (and indeed
  any other app that is laboring under network latency.) That way the
  client still remains thin and resources from the server are available
  when needed, rather than running the clients whole hog fat. Of course
  experimenting one way or another takes time. Perhaps start with just
  two or three clients and experiment with altering just their
  configurations.
 
  Richard
 
 
 
 --
  Everyone hates slow websites. So do we.
  Make your web apps faster with AppDynamics
  Download AppDynamics Lite for free today:
  http://p.sf.net/sfu/appdyn_d2d_feb
  _
  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



 --
 :-) Lachele
 Lachele Foley
 CCRC/UGA
 Athens, GA USA


 --
 Everyone hates slow websites. So do we.
 Make your web apps faster with AppDynamics
 Download AppDynamics Lite for free today:
 http://p.sf.net/sfu/appdyn_d2d_feb
 _
 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

Re: [Ltsp-discuss] performances problem: a partial solution (qt4)

[Jakob Unterwurzacher]

On 02.11.2012 11:14, Thierry Dumont wrote:
 I recall our problem:

 - migrating from Ubuntu 10.04 to Ubuntu 12.04 = many application
 became very slow, and KDE extremely slow.

 A partial solution:

 When migrating qt4 switched from 4.6 to 4.8 the graphicssystem changed
 from native to raster. This generates a *huge* X11 traffic on the
 network.

 Question: how to fix graphicssystem to native system wide? where?

 Yours
 t.d.

Now, that seems to be a pointer in the right direction!
There is a bugreport about this:
https://bugzilla.redhat.com/show_bug.cgi?id=830533

Try (in a terminal)

 export QT_GRAPHICSSYSTEM=native
 kile

and see if that brings Kile's speed back to normal. Then you can place 
the export command in a startup script, I'd suggest
/etc/X11/Xsession.d/99qtfix

Regards,
Jakob

--
LogMeIn Central: Instant, anywhere, Remote PC access and management.
Stay in control, update software, and manage PCs from one command center
Diagnose problems and improve visibility into emerging IT issues
Automate, monitor and manage. Do more in less time with Central
http://p.sf.net/sfu/logmein12331_d2d
_
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

Re: [Ltsp-discuss] I'm really soured on Edubuntu

[Jakob Unterwurzacher]

On Tue, Sep 4, 2012 at 4:08 AM, rob r...@fantinibakery.com wrote:

 I saved this to try sometime:

 [default]
LDM_SESSION=gnome-session --session=ubuntu-2d


Yes, something like that should work.

However, that the default install is just broken, and there is no
information in the release notes
https://wiki.edubuntu.org/PrecisePangolin/ReleaseNotes/Edubuntu
makes me wonder: Has anybody reported that problem to Ubuntu? (i.e.
launchpad.net)
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_
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

Re: [Ltsp-discuss] tried to change to fat clients and made a mess

[Jakob Unterwurzacher]

On 27.08.2012 02:49, Robert Lefebvre wrote:
 Perhaps it may be of value to mention I had backed up /opt/ltsp/i386

 I did that because the fat client install procedure kept getting an
 error that i386 was already present.

Okay that's good. I assume you have the original i386 folder in place now.

 I tried to replace the current new
 one with the original but still got the same error and nothing changed.

/opt/ltsp/i386 is compressed into /opt/ltsp/images/i386.img when you run 
ltsp-update-image.

That means you probably still have the broken fat client i386.img lying 
there. Run ltsp-update-image create a fresh one from the i386 folder.

Then you should be back to slow-but-working thin clients.

Jakob



--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_
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

Re: [Ltsp-discuss] Need to disable lock screen for all users (12.04)

[Jakob Unterwurzacher]

On Sun, Aug 12, 2012 at 9:59 PM, David Trask dtr...@vcsvikings.org wrote:

 ** **
 Thanks for the help, but it doesn't work for Ubuntu 12.04 (Precise).  I
 tried dconf-editor and that gave me access to some keys, but it still
 didn't work.  I'm stumped and a little desperate


Brute-force temporary clutch: Rename the screen lock executable so it can't
start?
I can't check right now, but it's probably /usr/bin/gnome-screensaver, so
I'd rename to gnome-screensaver.DISABLED  or something.

Jakob
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_
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

Re: [Ltsp-discuss] Choppy scrolling in Chromium on terminal

[Jakob Unterwurzacher]

On 11.08.2012 02:46, John Hupp wrote:
 With fully updated Lubuntu 12.04 on a server with an AMD 3400+ and
 512MB, and a terminal with a PIII 600 MHz and 128MB, I find that the
 scrolling in Chromium 18 on the terminal is choppy.

 The same result on a terminal with Celeron 500 MHz and 196MB.

 This is the case even with mostly text content, like a page of Google
 results.

 Scrolling on the server is very smooth.

 This is a test setup with just one server and client, nothing open on
 the server, and a 100 Mbps connection.

 Anyone know why and if I can fix it?


Yes, I have also noticed that.

There is a bugreport,
http://code.google.com/p/chromium/issues/detail?id=28544 , quoting:

 Comment 25 by e...@chromium.org , Jan 26, 2010
 Chrome has a different drawing model than other browsers (due to the sandbox, 
 web
 pages render as images).  It means that it is unlikely to ever perform as 
 well over
 remote X as other browsers.

So basically, you'd rather use Firefox on LTSP.

Jakob

--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_
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

Re: [Ltsp-discuss] installation issues

[Jakob Unterwurzacher]

On 09.08.2012 22:43, Robert Lefebvre wrote:
 Thanks Rob,

 Here is what lts.conf says

 [default]
 LDM_THEME=edubuntu
 LANG=en_US.UTF-8
 LANGUAGE=en_US.UTF-8
 LDM_LANGUAGE=en_US.UTF-8


 I added to it
 XSERVER = vesa
 X_MODE_0= 1280x1024

 but that was a no-no (the server wouldn't boot afterwards). As I am
 reinstalling Edubuntu for yet another time I'll mention what may be
 another symptom/cause. When installing we are given the option to
 download updates as we install. Updating seems to cause edubuntu to not
 be able to access the Internet afterwards (not positive but sure seems
 like it).


Editing the lts.conf should not matter for booting the server at all. 
That it loses a NIC when you update is also not a good sign. And you 
really should do those updates.
I understand that your server machine has lots of Nvidia chips in it, 
these have given me some trouble recently.
I'd suggest to try using one the P4 clients as the (temporary) server. 
And get a real second NIC, I doubt the USB thingy will work reliably.

I also wonder, none of these machines have PXE builtin?


Jakob

--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_
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

Re: [Ltsp-discuss] How to deal with dead hard drive?

[Jakob Unterwurzacher]

 2. What do people suggest for speediness and reliability? I run a
 30-seat fat client lab, so the server hard drive only gets used for
 DHCP, loading the OS, programs, authentication, and routing internet
 to the clients. Should I consider an SSD (if I put /tmp and /var on
 some other drive) or is it not worth it? Same question with so-called
 enterprise drives? Are they worth the increased cost or should I
 just get a well-regarded consumer drive?


If the disk image the clients boot off fits into RAM, it will be cached
anyway, so I wouldn't bother getting an SSD.

And if I still got an SSD, I wouldn't bother putting /tmp and /var
somewhere else. Hell, most users have the swapfile on the SSD, they
hopefully can take a few logfiles by now.

Jakob
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_
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

Re: [Ltsp-discuss] Computer Lab with LTSP

[Jakob Unterwurzacher]

Wifi will not work as you cannot boot from it. You need cabled ethernet.

2012/7/4 Abraham Close amclos...@gmail.com

 Hello,

 So I just found out that my school will have to use LTSP over wifi
 because of costs. I have some questions:
 How will I do this?
 Will it be cheaper to use ethernet?
 What hardware specs do I need?

 Thanks,
 Abraham Close


--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_
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

Re: [Ltsp-discuss] slow scrolling in chrome on ltsp

[Jakob Unterwurzacher]

On 27.03.2012 16:18, Evan Ingram wrote:
 anyone else experienced slow laggy scrolling in chrome on ltsp? even on
 just simple text web pages. on the same thin client firefox runs with no
 lag when scrolling through even image heavy pages

Yes, Chrome is nearly unusable over LTSP - or any remote session.
See for example http://code.google.com/p/chromium/issues/detail?id=28544 
, though I think I have also seen a specific bugreport for LTSP or remote X.

--
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_
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

Re: [Ltsp-discuss] is LDM_LIMIT_ONE_SESSION still supported under ubuntu lucid?

[Jakob Unterwurzacher]

Am 09.02.2011 17:55, schrieb Alkis Georgopoulos:
 Στις 09-02-2011, ημέρα Τετ, και ώρα 10:32 +0100, ο/η Jakob
 Unterwurzacher έγραψε:
 I still use gnome-watchdog that kills dangling session automatically.
 
 An LTSP developer implemented a much better alternative to
 gnome-watchdog, called xexit. It's available in his PPA:
 https://launchpad.net/~sbalneav/+archive/ppa/+packages
 
 I don't think there's any wiki page describing its use, but I think
 there are some examples in its source code.

Now that's a proper solution!
Looking at the source, seems like simply installing the package will do it.

--
The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
Pinpoint memory and threading errors before they happen.
Find and fix more than 250 security defects in the development cycle.
Locate bottlenecks in serial and parallel code that limit performance.
http://p.sf.net/sfu/intel-dev2devfeb
_
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

Re: [Ltsp-discuss] is LDM_LIMIT_ONE_SESSION still supported under ubuntu lucid?

[Jakob Unterwurzacher]

Am 08.02.2011 22:25, schrieb john:
 Hi Peter and Matt!
 
 On Mon, Feb 7, 2011 at 4:47 PM, Peter Matulis
 peter.matu...@canonical.com wrote:
 
 I read that the proper way to kill stale user sessions is no longer
 gnome-watchdog but an lts.conf switch called LDM_LIMIT_ONE_SESSION

Proper... This will only kill the session when the user logs in again.
So you can have lots of dangling sessions for a long time. I still use
gnome-watchdog that kills dangling session automatically.

Jakob

--
The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
Pinpoint memory and threading errors before they happen.
Find and fix more than 250 security defects in the development cycle.
Locate bottlenecks in serial and parallel code that limit performance.
http://p.sf.net/sfu/intel-dev2devfeb
_
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

Re: [Ltsp-discuss] Google Chrome Lockdown

[Jakob Unterwurzacher]

Am 04.01.2011 17:00, schrieb Sam Deeljore:
 You should consider opera... since it has a kiosk mode which takes
 care of most things...

Chrome has a --kiosk switch that removes the UI and goes fullscreen,
only possibility to exit is Alt-F4. So the settings are basically
inaccessible for the user.

Jakob

 
 Sam
 
 On Tue, Jan 4, 2011 at 4:30 AM, Leon Hauck l...@fulcrum-hosting.com wrote:
 I was thinking about trying out Chrome instead of FireFox for my LTSP setup.

 Is there any documentation out there to on how to lock down and optimize
 chrome for an LTSP installation like there is for FireFox?

 I'm looking for such things as forcing settings for using Squid,
 disabling local cache, locking out the bookmark editing, connection
 settings, etc.

 Thanks in advance.

 -- Leon

 --
 Learn how Oracle Real Application Clusters (RAC) One Node allows customers
 to consolidate database storage, standardize their database environment, and,
 should the need arise, upgrade to a full multi-node Oracle RAC database
 without downtime or disruption
 http://p.sf.net/sfu/oracle-sfdevnl
 _
 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

 
 
 


--
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and, 
should the need arise, upgrade to a full multi-node Oracle RAC database 
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_
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

Re: [Ltsp-discuss] LTSP for exam setting at school

[Jakob Unterwurzacher]

Am 17.11.2010 08:46, schrieb Rolf-Werner Eilert:
 Am 17.11.2010 01:13, schrieb sm8ps-ltsp...@yahoo.com:
 Hello all!

 
 Hello Stefan,
 
 
 I would like to use LTSP specifically for exam settings in an introductory
 computer science class. The students have learned some basic system
 
 My goal is to have the students work individually on one and the same
 (LTSP-)server without means of communication or file-sharing, just as if they
 were working on individual machines without any connectivity. Googling did 
 not
 
 Basically, that is what it should be by default. You can make sure they 
 cannot peek into each others' $homes by chmod 700. This is what I did 
 here. Maybe your distribution has already done it anyway.

Yes, but they can change it back to 777 easily.
chown root, chmod 070 should do it, as Stefan has suggested.

Then, you have the world-writeable directories /tmp, /dev/shm and
probably others ( `find / -perm -o+w` to find all ).

For /tmp, chmod 000 and set TMPDIR=/home/$user/tmp
( http://en.wikipedia.org/wiki/TMPDIR ) in the session init (
/etc/X11/Xsession.d/ ?)

Making /dev/shm unreadable will kill pulseaudio, but well.

What about the wall command? Probably chmod g-w /dev/pts/* will do it
(in a loop !), haven't tested.

Lots of corner cases... Probably fat clients as suggested by Alkis would
be easier in the end.

Jakob

--
Beautiful is writing same markup. Internet Explorer 9 supports
standards for HTML5, CSS3, SVG 1.1,  ECMAScript5, and DOM L2  L3.
Spend less time writing and  rewriting code and more time creating great
experiences on the web. Be a part of the beta today
http://p.sf.net/sfu/msIE9-sfdev2dev
_
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

Re: [Ltsp-discuss] LTSP for exam setting at school

[Jakob Unterwurzacher]

Am 17.11.2010 12:53, schrieb Xavier Brochard:
 Le mercredi 17 novembre 2010 12:01:28, Jakob Unterwurzacher a écrit :
 Am 17.11.2010 08:46, schrieb Rolf-Werner Eilert:
 Am 17.11.2010 01:13, schrieb sm8ps-ltsp...@yahoo.com:
 I would like to use LTSP specifically for exam settings in an
 introductory computer science class. The students have learned some
 basic system

 My goal is to have the students work individually on one and the same
 (LTSP-)server without means of communication or file-sharing, just as if
 they were working on individual machines without any connectivity.
 Googling did not

 What about the wall command? Probably chmod g-w /dev/pts/* will do it
 (in a loop !), haven't tested.
 
 And what about the mail command?...

Good catch! But
/etc/init.d/postfix stop
will take care of that (if there is a MTA installed at all)

 There is so much to do with the shell, 
 that it looks faster to disable virtual consoles in lts.conf (SCREEN_O1, 
 SCREEN_02, see LTSP manual) and to remove all the terminals applications 
 (don't forget embeded terminals in applications like kate, konqueror, ...). 

As this is a computer science class, the students will be meant to use
the shell, can't disable it...

 The best option is to use a restricted shell together with a chroot jail for 
 each user. 
 
 Don't forget to lockdown the desktop! for that purpose you can use Pessulus 
 on 
 Gnome (and also edit the registry) and the Kiosk framework on KDE.
 
 Xavier
 xav...@alternatif.org - 09 54 06 16 26
 
 --
 Beautiful is writing same markup. Internet Explorer 9 supports
 standards for HTML5, CSS3, SVG 1.1,  ECMAScript5, and DOM L2  L3.
 Spend less time writing and  rewriting code and more time creating great
 experiences on the web. Be a part of the beta today
 http://p.sf.net/sfu/msIE9-sfdev2dev
 _
 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


--
Beautiful is writing same markup. Internet Explorer 9 supports
standards for HTML5, CSS3, SVG 1.1,  ECMAScript5, and DOM L2  L3.
Spend less time writing and  rewriting code and more time creating great
experiences on the web. Be a part of the beta today
http://p.sf.net/sfu/msIE9-sfdev2dev
_
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

Re: [Ltsp-discuss] LTSP for exam setting at school

[Jakob Unterwurzacher]

Am 17.11.2010 13:47, schrieb Xavier Brochard:
 And what about the mail command?...

 Good catch! But
 /etc/init.d/postfix stop
 will take care of that (if there is a MTA
 
 AFAIK, Bsd mailx or Gnu Mailutils doens't need an MTA for local delivery.

Hmm, where would they deliver the mail to? /var/mail/ isn't
world-writeable. It's not SUID either, and at least on Ubuntu, the
mailutils package explicitely depends on an MTA.
Anyway, should be tested before depending on it :)

--
Beautiful is writing same markup. Internet Explorer 9 supports
standards for HTML5, CSS3, SVG 1.1,  ECMAScript5, and DOM L2  L3.
Spend less time writing and  rewriting code and more time creating great
experiences on the web. Be a part of the beta today
http://p.sf.net/sfu/msIE9-sfdev2dev
_
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

Re: [Ltsp-discuss] advices needed before buying a new server

[Jakob Unterwurzacher]

Am 05.11.2010 18:37, schrieb David Burgess:
 On Fri, Nov 5, 2010 at 11:24 AM, Xavier Brochard xav...@alternatif.org 
 wrote:
 And what about to copy all the system in Ram, like some liveCD does?
 
 Interesting idea for the speed demons. It's not something I've ever
 tried. I once read a forum post from a guy who made a mirrored pair,
 one member being the root fs, the other being a ramdisk. Of course he

I'm not sure this is going into the right direction. This all boils down
to trying to be smarter than the kernel's file cache that will keep
frequently accessed files in memory anyway.

If you want to speed up the first access after reboot, read-ahead like
https://launchpad.net/ureadahead included in Ubuntu should do what you want.

Jakob

--
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book Blueprint to a 
Billion shares his insights and actions to help propel your 
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_
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

Re: [Ltsp-discuss] Thin-client screen resolution problem

[Jakob Unterwurzacher]

Am 03.10.2010 15:43, schrieb stef guyonne:
 Hi
 
 My setup was working with ubuntu 8.04, since I switch to ubuntu 10.04
 one of my thin-client work only with 800x600 resolution (resolution I
 should set in the lts.conf with this:
 [MAC ADRESS]
 X_MODE_0 = 800x600
 
 This happen only with one thin-client, the 2 others works correctly
 (the 3 machines have differents hardwares)
 
 When I boot upthe thin-client with the lts.conf directive commented, I
 got the splash screen, but ldm won't start. If I connect on a console
 as root and try ldm start, I got this message:
 
 (process:1786): GLib-CRITICAL **: g_strsplit: assertion 'string != NULL' 
 failed
 date: invalid date '+%s'
 [: 2: Illegal number:
 bs=1024, sz=581560
 
 (ldmgtkgreet:1788): Gtk-WARNING **: cannot open display:
 

Hi, have you looked at /var/log/Xorg.0.log on the thin client? Looks
like X has some problem selecting a valid resolution.

Jakob

--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_
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

Re: [Ltsp-discuss] Issue with dbus-daemon

[Jakob Unterwurzacher]

Am 22.09.2010 03:25, schrieb David Hopkins:
 I believe I have tracked down the cause of this problem ...
 dbus-daemon exceeds the nofile limit:
 https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/381063

[...]

 
 How can I increase the nofile limit for dbus?   In RHEL I could edit
 the rc script as a last resort but Ubuntu Karmic uses upstart.  Any
 assistance is greatly appreciated as I don't know where to turn next
 on this.

Hi!

Add the line

limit nofile 1 1

to /etc/init/dbus, this should fix it. At least it works here on Ubuntu
10.04.

Side note: There is dbus-daemon started for the system and one for each
user, I looked at the user one for some time and was puzzled why the
limit didn't take effect, so better get the PID from
service dbus status


Regards,
Jakob

--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_
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

Re: [Ltsp-discuss] Issue with dbus-daemon

[Jakob Unterwurzacher]

Am 22.09.2010 13:37, schrieb David Hopkins:
 /etc/init/dbus does not exist on my system.  /etc/init/dbus.conf does
 though ... it is the script that starts dbus.  /etc/init.d/dbus is a
 symlink to the /lib/init/upstart-job command.
 
 Do I add the limit line to the dbus.conf file?  Or create the dbus
 file?  More than happy to make the change but since I'll have to log
 all the users off to restart the system I want to be sure first.

Sorry, meant to write dbus.conf - the existing file /etc/init/dbus.conf !

 
 And I also had the concern about which dbus I was looking at. I am
 looking at the system dbus.
 
 Thanks!
 Dave ...
 
 On Wed, Sep 22, 2010 at 6:56 AM, Jakob Unterwurzacher
 jakob...@gmail.com wrote:
 Am 22.09.2010 03:25, schrieb David Hopkins:
 I believe I have tracked down the cause of this problem ...
 dbus-daemon exceeds the nofile limit:
 https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/381063

 [...]


 How can I increase the nofile limit for dbus?   In RHEL I could edit
 the rc script as a last resort but Ubuntu Karmic uses upstart.  Any
 assistance is greatly appreciated as I don't know where to turn next
 on this.

 Hi!

 Add the line

 limit nofile 1 1

 to /etc/init/dbus, this should fix it. At least it works here on Ubuntu
 10.04.

 Side note: There is dbus-daemon started for the system and one for each
 user, I looked at the user one for some time and was puzzled why the
 limit didn't take effect, so better get the PID from
 service dbus status


 Regards,
 Jakob



--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_
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

Re: [Ltsp-discuss] Issue with dbus-daemon

[Jakob Unterwurzacher]

Am 22.09.2010 14:18, schrieb David Hopkins:
 Success!

Glad to hear! I added the workaround to
https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/381063
in case someone else hits this.

 
 /etc/init/dbus does not exist on my system.  /etc/init/dbus.conf does
 though ... it is the script that starts dbus.  /etc/init.d/dbus is a
 symlink to the /lib/init/upstart-job command.

 Do I add the limit line to the dbus.conf file?  Or create the dbus
 file?  More than happy to make the change but since I'll have to log
 all the users off to restart the system I want to be sure first.

 Sorry, meant to write dbus.conf - the existing file /etc/init/dbus.conf !
 
 Thanks! adding this to dbus.conf reset the nofile limits as needed!
 
 Snippet from file:
 
 start on local-filesystems
 stop on runlevel [06]
 
 limit nofile 1 1
 
 expect fork
 respawn
 ..
 
 SIncerely,
 Dave ...
 


 And I also had the concern about which dbus I was looking at. I am
 looking at the system dbus.

 Thanks!
 Dave ...

 On Wed, Sep 22, 2010 at 6:56 AM, Jakob Unterwurzacher
 jakob...@gmail.com wrote:
 Am 22.09.2010 03:25, schrieb David Hopkins:
 I believe I have tracked down the cause of this problem ...
 dbus-daemon exceeds the nofile limit:
 https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/381063

 [...]


 How can I increase the nofile limit for dbus?   In RHEL I could edit
 the rc script as a last resort but Ubuntu Karmic uses upstart.  Any
 assistance is greatly appreciated as I don't know where to turn next
 on this.

 Hi!

 Add the line

 limit nofile 1 1

 to /etc/init/dbus, this should fix it. At least it works here on Ubuntu
 10.04.

 Side note: There is dbus-daemon started for the system and one for each
 user, I looked at the user one for some time and was puzzled why the
 limit didn't take effect, so better get the PID from
 service dbus status


 Regards,
 Jakob





--
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_
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

Re: [Ltsp-discuss] firefox local_app associated file types

[Jakob Unterwurzacher]

On 07/06/10 03:24, Michael Pope wrote:
 I'm running LTSP 5.2.1 and tried to use the ltsp-remoteapps on one of my 
 clients and it didn't start the application.  Here is what I did:
 
 On the thin client:
 Opened up a terminal
 $ ltsp-localapps xterm
 
 With in xterm window
 $ ltsp-remoteapps gnome-terminal
 
 Do I have to get a more recent version of LTSP or is the ltsp-remoteapps 
 not working at all yet?

Are you sure gnome-terminal is installed on the thin client?

Jakob


--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_
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

Re: [Ltsp-discuss] firefox local_app associated file types

[Jakob Unterwurzacher]

On 07/06/10 16:43, Jakob Unterwurzacher wrote:
 With in xterm window
 $ ltsp-remoteapps gnome-terminal

 Do I have to get a more recent version of LTSP or is the ltsp-remoteapps 
 not working at all yet?
 
 Are you sure gnome-terminal is installed on the thin client?

Well. Please disregard this email.

--
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_
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

Re: [Ltsp-discuss] How to supress boot banner ?

[Jakob Unterwurzacher]

Am 31/05/10 14:13, schrieb Thierry Dumont:
 In our lab, we have 2 sorts of ltsp clients. We are using Ubuntu 10.04 
 and the ltsp packages which come with it.
 During the boot of the  clients a banner Ubuntu with blinking dots is 
 displayed. This works fine on one sort of clients, but not on the other 
 one, where the banner looks completely broken (but excepted this, ltsp 
 works fine on both clients).
 
 Do you know some way:
 -to improve the display during boot,

This is done by the kernel framebuffer driver and plymouth... I have no
idea.

 -or one way to suppress this banner ?

This should do it:

 disable the Ubuntu bootsplash for the chroot environment, so we can see 
 bootup messages scroll by:
 
 sudo vim /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
 Remove the word quiet and splash. Save and exit, and reboot the thin-client. 
 You should now see all bootup messages instead of the Ubuntu graphical 
 bootsplash.

https://help.ubuntu.com/community/UbuntuLTSP/ClientTroubleshooting#Troubleshooting%20at%20the%20client%20itself


Jakob

--

_
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

Re: [Ltsp-discuss] Bonding 3 networkcards and 3 switches.

[Jakob Unterwurzacher]

On 18/05/10 21:43, jan nilsen wrote:
 So, I have a server with a 1gbit NIC, and 3 16-ports switches (all
 100mbit ports), that are linked togheter the old switch-in-switch
 style, and I have about 45 clients (mix of thinclients, workstations
 and laptops).
 
 When most of the machines are in use, I notice that having only
 100mbit out of the server is not enough.
 
 So I can either buy myself a 48 port switch with 1gbit uplink with
 money we don't have, or I thought I could try this bonding thing.
 
 I can put 3 network cards in the server, and connect each of those
 networkcards with it's own 16-port switch, that way I would have
 300mbit out of the server.
 
 What kind of bonding mode should I choose?
 
 Or is there some other way?
 

I can't find it right now on the list, but with this kind of setup it
was suggested to put the 3 NICs in different subnets and configure dhcpd
to hand out IPs from the subnet the NIC is in.
You can't and needn't use bonding when you have 3 seperate switches.


Jakob

--

_
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

Re: [Ltsp-discuss] Firefox as localapps

[Jakob Unterwurzacher]

On 18/05/10 17:51, Verner Kjærsgaard wrote:
 
 
 Xavier Brochard skrev:
 Le mardi 18 mai 2010 17:03:44, Verner Kjærsgaard a écrit :
 Hi list,

 - when FF is running as a localapps and, say, 20 students are running FF
 from their TC, how many (if any) instances in the servers ps/top should
 I see?

 One for each initiated localapps firefox or none?

 none
 if it is running as a local aplication, that mean it runs on the thin client 
 hardware, not the server

 Xavier
 xav...@alternatif.org - 09 54 06 16 26

 
 YES! Exactly! That's what I thought too...I'm very glad that you
 confirmed this.
 
 Now...I'll start hunting down why FF doesn't run as a localapp when
 indeed I have the correct entries in lts.conf.
 
 Further, when FF runs localapp (started manually from within a shell), I
 don't have correct localization and so. I'll dig into it.

Just a guess: Do you have language-support-da installed in the thin
client? (or however the language pack for your language is called)

Jakob

--

_
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

Re: [Ltsp-discuss] Firefox as localapps

[Jakob Unterwurzacher]

On 19/05/10 22:28, Verner Kjærsgaard wrote:
 
 
 Jakob Unterwurzacher skrev:
 On 18/05/10 17:51, Verner Kjærsgaard wrote:

 Xavier Brochard skrev:
 Le mardi 18 mai 2010 17:03:44, Verner Kjærsgaard a écrit :
 Hi list,

 - when FF is running as a localapps and, say, 20 students are running FF
 from their TC, how many (if any) instances in the servers ps/top should
 I see?

 One for each initiated localapps firefox or none?
 none
 if it is running as a local aplication, that mean it runs on the thin 
 client 
 hardware, not the server

 Xavier
 xav...@alternatif.org - 09 54 06 16 26

 YES! Exactly! That's what I thought too...I'm very glad that you
 confirmed this.

 Now...I'll start hunting down why FF doesn't run as a localapp when
 indeed I have the correct entries in lts.conf.

 Further, when FF runs localapp (started manually from within a shell), I
 don't have correct localization and so. I'll dig into it.

 Just a guess: Do you have language-support-da installed in the thin
 client? (or however the language pack for your language is called)

 Jakob

 
 Hi
 
 re: 'language-support-da'
 Well, I didn't build my own image. I just used the pre-compiled standard
 image. So...in order for me to have language support for danish in the
 thin client - what would I do?

What distribution do you use? I'm afraid i can only help with Ubuntu,
where you would do

 sudo -s
 chroot /opt/ltsp/i386
 apt-get install language-support-da
 exit
 ltsp-update-image


Jakob

--

_
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

Re: [Ltsp-discuss] mount usb ntfs read/write

[Jakob Unterwurzacher]

Am 29/04/10 12:15, schrieb Chris Roberts:
 On Thursday 29 Apr 2010, Wim De Geeter wrote:
 No we use nfs,

 But it is still really not clear to me when to use ltsp-update-image
 So I  thought that every time you add some packages in the chroot you
 need to update your ltsp-update-image

 No, with Debian and NFS you never use this command, delete it from your mind.

 Indeed I think you probably need to remove the image it has created, which I
 think is something like /opt/ltsp/images or something like that.

 Obviously I don't use ltsp-update-image, so I'm working from vague
 recollection, but certainly I don't have /opt/ltsp/images on my system, so if
 you do, then it probably should be removed.

 It would be good if someone who is more confident of their facts could
 confirm.

Yep, that's the chroot image.

/opt/ltsp/images$ ls -lh
total 350M
-rwxr--r-- 1 root root 350M Feb 26 20:02 i386.img


 Careful not to remove your /opt/ltsp/i386 directory though!

 There must be a lot of Debian LTSP users diligently running this command,
 could it not be modified to check for whether nfs/nbd is in use and politely
 educate the user?



--
_
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

Re: [Ltsp-discuss] Problems with twinkle / openoffice crashing and leaving a PID opened, how to kill in combination of thin/fat client. Log out log in doesnt work!!

[Jakob Unterwurzacher]

Am 23/04/10 06:35, schrieb David Taveras:
 Hello,

 I have a new LTSP setup with openoffice running at the server (ubuntu
 9.10), and twinkle running on the thin client.

 a.) If I right click the gnome deskptop and create an empty file and
 then try to open it with Word from OpenOffice... then OpenOffice will
 not startup ... it leaves a PID running and even if I log out of the
 user and log in the PID (with the userid of the user) still remains
 running.

 Question: This one from an admin level is easy to take care off... I
 would simply grep the users processes and kill the openoffice match.
 Not practicla though. What is the best way to deal with this problem
 or at least correctly kill it without hassle?

Openoffice is running at the server, right?
Then, gnome-watchdog is the solution to the lingering processes problem:
https://help.ubuntu.com/community/UbuntuLTSP/GnomeWatchdog
(site down right now, google cache: 
http://66.102.9.132/search?q=cache:D4WQAi85bL0J:https://help.ubuntu.com/community/UbuntuLTSP/GnomeWatchdog+gnome-watchdogcd=1hl=dect=clnk
 
)

 b.) Some users that use twinkle manage to crash it and make it quit.
 When they try to start it again it complaints it cannot establish a
 socket.. it is already in use... I have to go to the thin client and
 do alt + ctrl + f2 and grep and kill the twinkle process for the user
 to start it ok.

 Question: How can the user kill that process by himself... else is it
 possible for me to create a script that will kill this process RUNNING
 ON THE THIN CLIENT?   What is the best way to deal with PIDs that
 refuse to die even if i log off and log in... PIDS that run on the
 thin client?

An ad-hoc solution would be to create a script for the user to 
double-click Reset twinkle that does something like
ltsp-localapps pkill -9 twinkle
.

Killing all lingering processes on the thin client on logoff would be a 
good idea though


Jakob

--
_
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

Re: [Ltsp-discuss] Bypass Autologin

[Jakob Unterwurzacher]

Am 20/03/10 20:28, schrieb Mark Ehle:
 Thanks for the reply -  can a user other than guest log in? I want to
 log in as an admin.

 vagr...@freegeek.org mailto:vagr...@freegeek.org wrote:
 will display a login as guest button, but allow an alternate
 username to be
 specified, and after 10 seconds will autologin.

I think so.

--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_
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

Re: [Ltsp-discuss] Firefox is indeed the high disk IO culprit was Re: recommend a disk setup for LTSP install?

[Jakob Unterwurzacher]

Am 02/04/10 03:54, schrieb Joseph Bishay:
 Hello,

 On Thu, Apr 1, 2010 at 5:37 PM, Jim McQuillanj...@mcquil.com  wrote:
 John,

 Try disabling the browser cache.  That should stop it from keeping
 copies of pages in users home directories.

 Jim McQuillan
 j...@ltsp.org

 Is there a way to do this for all users globally or is this a per-user
 setting that I would have to go into each account to achieve?

Yes, you do it it globally in /etc/firefox-3.X/pref/firefox.js . Details:
https://help.ubuntu.com/community/UbuntuLTSP/Firefox3Optimize

Jakob

--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_
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

Re: [Ltsp-discuss] recommend a disk setup for LTSP install?

[Jakob Unterwurzacher]

Am 31/03/10 01:58, schrieb john:
 Hi David et al,

 On Tue, Mar 30, 2010 at 3:59 PM, David Burgessapt@gmail.com  wrote:


 One SSD to rule them all. You could keep /home on a platter if you
 need a lot of space for it.


 For context I have 16 Gig's of of RAM and 4 Cores ~3.Ghz.

 Are  SSD write times comparable with SAS? I wasn't aware of that. I'll
 take a look.

 Here's what iostat told me during one recent bit of misery. When I see
 nearly 30K Blks_wrtn/s is it
 reasonable to conclude that 30K * 4092 (ext 3 blk size) 122.8 m bytes
 sec? That seems pretty high. Am I crazy?

A block in iostat is 512 Bytes [1] - so you have 15MB/s, that seem 
reasonable.

I wonder what is doing all those writes. The browser disk cache? If you 
don't want to buy new hardware, it might be worth trying to disable it, 
https://help.ubuntu.com/community/UbuntuLTSP/Firefox3Optimize

Jakob

[1] From man iostat

   Blk_read/s
  Indicate   the  amount  of  data  read  from  the  device
  expressed in a number of blocks per  second.  Blocks  are
  equivalent  to  sectors  with  kernels  2.4 and later and
  therefore have a size of 512 bytes. With older kernels, a
  block is of indeterminate size.

   Blk_wrtn/s
  Indicate  the  amount  of  data  written  to  the  device
  expressed in a number of blocks per second.


--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_
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

Re: [Ltsp-discuss] Edit user accounts from clients on Karmic / policykit-1

[Jakob Unterwurzacher]

Okay, i think i got it. I am able to unlock the dialog through a ssh 
connection, i assume that it will be the same for remote X.

Apply this patch: 
http://launchpadlibrarian.net/39471473/polkit-systemtools-remote-allow.patch
# sudo cp -a 
/usr/share/polkit-1/actions/org.freedesktop.SystemToolsBackends.policy 
/usr/share/polkit-1/actions/org.freedesktop.SystemToolsBackends.policy.ori
# sudo patch 
/usr/share/polkit-1/actions/org.freedesktop.SystemToolsBackends.policy 
polkit-systemtools-remote-allow.patch

Then kill polkitd, it will be restarted automatically when you open the 
user-manage dialog again:
# sudo pkill polkitd



Jakob

--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_
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

Re: [Ltsp-discuss] Edit user accounts from clients on Karmic / policykit-1

[Jakob Unterwurzacher]

Am 2010-02-19 08:57, schrieb Jussi P:

I get this:
isoju...@kantti-ltsp:~$ dpkg --get-selections | grep policykit
policykit   install

[...]

Same for me.

Ok, now. When you klick the unlock button (the one that is greyed out 
when connecting remotely) you get a dialog (screenshot attached).

The action there is org.freedesktop.systemtoolsbackends.set .

This

$ pkaction --verbose --action-id org.freedesktop.systemtoolsbackends.set

org.freedesktop.systemtoolsbackends.set:
   description:   Manage system configuration
   message:   System policy prevents modifying the system 
configuration

   vendor:
   vendor_url:
   icon:
   implicit any:  no
   implicit inactive: no
   implicit active:   auth_admin_keep

gives some info about this action and the description looks like it's 
the right one.


What remains to be done is to find out how to allow this action for 
remote users. I don't have the time to do this at the moment, but i 
think it's not too difficult.


Actually, i think i found the file this is defined in:
/usr/share/polkit-1/actions/org.freedesktop.SystemToolsBackends.policy

Looking at the file and
http://hal.freedesktop.org/docs/polkit/polkit.8.html
adding the line

allow_anyauth_self/allow_any

after allow_activeauth_self/allow_active looks primising.


Jakob

attachment: screenshot.png--
Download Intel#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev_
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

Re: [Ltsp-discuss] Edit user accounts from clients?

[Jakob Unterwurzacher]

Am 2010-02-16 16:25, schrieb David Burgess:
 On Tue, Feb 16, 2010 at 6:33 AM, Appiahmupp...@gmail.com  wrote:
 Make sure that the user has sudo access.

 This will allow administration from the shell, but doesn't give GUI
 access. The only solution I've seen for this in 9.10 was proposed by
 Jordan on the list on January 13. You may want to search that post, as

This is the thread:
http://thread.gmane.org/gmane.linux.terminal-server.general/27419

 it was a decently lengthy thread. This link was offered as the
 workaround:

 https://help.ubuntu.com/community/UbuntuLTSP/GDMVNCInetdssh

Looks like a bit of overkill...

The workaround in 
https://bugs.launchpad.net/ubuntu/+source/policykit/+bug/221363 doesn't 
work for you either?
I will update that bug, then.


Jakob

--
SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW
http://p.sf.net/sfu/solaris-dev2dev
_
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

Re: [Ltsp-discuss] Edit user accounts from clients?

[Jakob Unterwurzacher]

Am 2010-02-16 19:16, schrieb David Burgess:
 On Tue, Feb 16, 2010 at 9:18 AM, Jakob Unterwurzacher
 jakob...@gmail.com  wrote:

 Looks like a bit of overkill...

 The workaround in
 https://bugs.launchpad.net/ubuntu/+source/policykit/+bug/221363 doesn't
 work for you either?
 I will update that bug, then.

 policykit is not even installed in 9.10. I'm not sure what
 authentication mechanism is used now, I just resorted to the shell.

Hmm. This (in german)
http://wiki.ubuntuusers.de/PolicyKit
says that the policykit GUI went away with Karmic, but that it's still 
used and accessible via CLI.
The Karmic laptop in front of me has it installed, but that may be 
because it's upgraded from Jauny.

What does
$ dpkg --get-selections | grep policykit
give?



Jakob

--
SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW
http://p.sf.net/sfu/solaris-dev2dev
_
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

Re: [Ltsp-discuss] 64 bit server

[Jakob Unterwurzacher]

Am 2010-02-11 22:55, schrieb Ken Walker:
 I am about to order a server which will be used as a file server for a
 mixed windows/linux lan of about 8 stations and as an ltsp server for a
 couple of client terminals running on older pc's with plans to end up
 with a half dozen client terminals.  I was about to order a unit from
 dell for the server when I ran across a comment somewhere that 64 bit is
 the preferred architecture for the server.  I have stayed away from 64
 bit pc's for linux until now.  I expect that my small lan would not
 overload either as far as processing capability is concerned, but  I am
 wondering, should I use 64 bit?  Is there any difficulty using a 64 bit
 server with 32 bit clients?  I plan to use ubuntu.  The terminals will
 be used for word processing, email, web browsing and not much else.


http://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg37686.html

In short: Go 64 bit.


Jakob

--
SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW
http://p.sf.net/sfu/solaris-dev2dev
_
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

Re: [Ltsp-discuss] EEE PC with LTSP on Fedora 11

[Jakob Unterwurzacher]

Am 2010-01-29 01:17, schrieb P. Scott DeVos:
 Here is a little more information.  I am getting the following in my
 messages log on the server:

 Jan 28 18:15:38 www mountd[2622]: refused mount request from
 192.168.1.82 for /opt/ltsp/i386 (/opt/ltsp/i386): unmatched host


 Any help with this additional info?

So you are using NFS.
Looks like you need to allow the IP adresses of the clients in 
/etc/exports .
Sorry, can't give any details on what to put there as i use NBD instead 
of NFS.

Jakob



 On Wed, 2010-01-27 at 20:21 -0600, P. Scott DeVos wrote:
 I am trying to set up LTSP at my office.  I have a number of EEE PC 700
 and 900s available.  I get the same problem for each one:  The client
 begins to boot but then hangs shortly after switching to the frame
 buffer device.  Here is what I get on the client:

 console: switching to colour frame buffer device 100x30
 [drm] LVDS-8: set mode 800x480 13
 FB0: inteldrmfb frame buffer decice
 registered panic notifier
 [drm] Initialized i915 1.6.0 20080730 for :00:02.0 on minor 0
 Athereos(R) L2 Ethernet Driver - version 2.2.3
 copyright (c) 2007 Atheros Corporation.
 atl2 :03:00.0: PCI INT A -  GSI 17 (level, low) -  IRQ 17
 RPC: Registered udp transport module.
 RPC: Registered tcp transport module.
 Slow work thread pool: Starting up
 Slow work thread pool: Ready
 FS-Cache: Loaded
 FS-Cache: Netfs 'nfs' registered for caching
 nbd: registered device at major 43
 Synaptics Touchpad, model: 1, fw: 6.3, id: 0x180b1, caps:
0xa04753/0x20
 squashfs: version 4.0 (2009/01/31) Phillip Lougher
 input: SynPS/2 Synaptics TouchPad
   as /devices/platform/i8042/serio1/input/input7
 atl2: eth0 NIC Link is UP100 Mbps Full Duplex
 _


 Any help would be appreciated.





--
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_
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

Re: [Ltsp-discuss] LTSP 5, Gentoo, Dell Optiplex

[Jakob Unterwurzacher]

Am 2010-01-26 12:20, schrieb Michael George:
 I'm trying to move to LTSP 5 on my Gentoo system.  I have the client
 system built, the TFTP (tftp-hpa) and DHCP (dhcp) working to provide the
 necessary files, but the terminal hits a kernel panic when it tries to
 mount / over NFS.  It says that there's no network available.

 I know the connection is good, because it pulled the kernel and
 initramfs.  It looks like it picks the tg3 NIC driver to use for network-
 ing, but the system uses the 3C950 chipset and would need the 3C95x driver.

 I'm wondering if anyone here uses LTSP much and might be successfully
 using old Dell Optiplex GX1's and could perhaps give me some insight.


I think i use one, and that it needed IPAPPEND 3 to boot because it 
couldn't get an IP after the PXE boot
(IPAPPEND explained here:
https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP#Adjusting ).

If this does not help, posting a screenshot of the kernel messages 
wouldn't be bad.

Jakob

--
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_
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

Re: [Ltsp-discuss] LTSP 5, Gentoo, Dell Optiplex

[Jakob Unterwurzacher]

Am 2010-01-28 14:57, schrieb Michael George:
 On Thu, Jan 28, 2010 at 10:39:13AM +0100, Jakob Unterwurzacher wrote:
 Am 2010-01-26 12:20, schrieb Michael George:
 I'm trying to move to LTSP 5 on my Gentoo system.  I have the client
 system built, the TFTP (tftp-hpa) and DHCP (dhcp) working to provide the
 necessary files, but the terminal hits a kernel panic when it tries to
 mount / over NFS.  It says that there's no network available.

 I know the connection is good, because it pulled the kernel and
 initramfs.  It looks like it picks the tg3 NIC driver to use for network-
 ing, but the system uses the 3C950 chipset and would need the 3C95x driver.

 I'm wondering if anyone here uses LTSP much and might be successfully
 using old Dell Optiplex GX1's and could perhaps give me some insight.

 I think i use one, and that it needed IPAPPEND 3 to boot because it
 couldn't get an IP after the PXE boot
 (IPAPPEND explained here:
 https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP#Adjusting ).

Uh, the full Link is
 https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP#Adjusting 
 pxelinux.cfg/default
but i'm afraid it's killed by the space in the anchor.


 If this does not help, posting a screenshot of the kernel messages
 wouldn't be bad.

 Did you need this even without a proxy DHCP server?

Well, my setup is a little strange. I have the LTSP DHCP server (not 
dnsmasq, the regular dhcpd) and also one in a router that i cannot turn 
off. They hand out different ranges so this should be fine. PXE worked 
well from the start, but the kernel (or the dhcp client) on some 
machines failed to get a IP adress afterwards and paniced.

IPAPPEND 3 fixed that, everything works fine now.


 Most of the docs I find on setting up LTSP 5 use a different DHCP server
 (dnsmasq) and TFTP server (atftp) than I am using (dhcpd and tftp-hpa).
 However, they should function similarly.  I have LTSP 4 working and I
 am hoping to avoid making lots of changes in trying to go to 5 -- in
 case I have to fall back to 4 for the summer.

 I'll try the IPAPPEND 3 option and see what that does for me.  Then I
 think I'll try different hardware and see if it behaves the same way.

Good luck!

--
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
_
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

Re: [Ltsp-discuss] remote desktop to the actual ltsp server

[Jakob Unterwurzacher]

 ~$ vncviewer 10.20.3.13:0
 Connected to RFB server, using protocol version 3.7
 
 and then nothing happens. i've definitely disabled the confirm each
 access option so it shouldnt be popping up with an authorise access box
 on the server.

Hmm. When you use ssh -X something this will only spawn one window but
not the whole desktop. I'm not sure what vnc could display then. Anyway,
you don't need this to work on the server.

Evan Ingram schrieb:
 On Wed, 2010-01-13 at 17:23 +0530, CyberOrg wrote:
 When you work on any LTSP client you are actually working
 _on_the_server_ already.

Yes!

 
 how come things are different if i log in physically on the server
 though? such as users and groups.
 
 if i log in as admin on the box itself i can authorise and add users and
 groups. if i log in via ltsp i goto user settings and the unlock icon is
 greyed out.

It is because those buttons are clever. They notice that you are not
logged in directly but over a thin client.

There is a bug about this here:
https://bugs.launchpad.net/ubuntu/+source/policykit/+bug/221363

Anyway, the workaround is

 I was able to get access via VNC tunneled through SSH by changing the 
 following settings in policykit. You can do it locally via Authorizations, or 
 you can do it remotely using sudo ck-launch-session 
 polkit-gnome-authorization in a terminal window in your tunneled VNC 
 session. This worked on Ubuntu 9.04 Server RC running xubuntu-desktop, so as 
 always YMMV.
 
 For system configuration, change all implicit authorizations under org - 
 freedesktop - systemtoolsbackends - Manage System Configuration 
 (org.freedesktop.systemtoolsbackends.set) to Admin Authentication.
 
 For user management, change all implicit authorizations under org - 
 freedesktop - systemtoolsbackends - self - Change User Configuration 
 (org.freedesktop.systemtoolsbackends.self.set) to Authentication.
 
 Reset gdm by rebooting or running sudo /etc/init.d/gdm restart from a 
 terminal window, and you should be able to unlock the user settings control 
 panel and other similarly useful things through your tunneled VNC session.


Jakob

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] Help with ThinClient everythin upside down (gnome)

[Jakob Unterwurzacher]

Jose Rivera Merla schrieb:
 *Hi:
 *
 *Problem:
 *
 * *After the login screen, gnome appears with all the charactersm
 includin menus, etc. upside down.
 
 This means that the letter P of Places appears like a letter b.
 
 I don't know where to fix this, *I tried KDE and it appears OK, so the
 problem is with gnome.*
 
 I'd like to keep gnome because it's the one I'm used to work with it, so
 it would help me to  keep it
 in the Thin Clients too.
 
 *Configuration:*
 I'm using Ubuntu 9.10 installed with the ltsp option.
 
 After following some help in the ubuntu pages about dhcp, networking,
 and setting up the two NIC (ethernet network cards), the thin client works.
 
 *Thanks in advance for your help.*
 
 *Regards,
 Jose*

Wow. Would you send a screenshot?
Then, what is the graphics card on the thin client?
Have you tried creating a brand new user?

Jakob

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] Multiple netcards in one LTSP server, how...

[Jakob Unterwurzacher]

Verner Kjærsgaard schrieb:
 
 Alkis Georgopoulos skrev:
 Στις 12-01-2010, ημέρα Τρι, και ώρα 12:30 -0700, ο/η David Burgess
 έγραψε:
 On Tue, Jan 12, 2010 at 12:18 PM, Alkis Georgopoulos alk...@gmail.com 
 wrote:

 See also https://help.ubuntu.com/community/UbuntuLTSP/FlowControl for a
 possible problem with mixed-speed networks that may gravely affect
 perfomance.
 Flow control is a problem when you have mixed media speeds sharing a
 server. and without a gbit switch port for the server, flow control is
 not likely to be an issue, and each subnet will cruise along at up to
 (!)100 mbps. 
 I think the poster said that he has a mixed mode network:

 - I've got 4 24-port switches that connects to all the clients. 100Mbit
 ports to the clients, 1000Mbit to the server.


 
 Hi again,
 - a few more facts.
 
 Well, I've not got 100 clients but someting like 70 or so. They don't
 run at the same time, perhaps 20 - 20 are active at the same time. The
 swithes in use are 24 port swithes, hence the need for distributing
 properly.
 
 I think the solution is to opt for the only real correct one...get new
 switches that are NOT 1000/100 switches, but rather 1000/1000 switches...
 
 This way the switches may be uplinked together properly. I'm quite
 convinced that the servers 1000Mbit netcard (just one) will handle the
 traffic just fine.

Those switches really have 24 100Mbit ports and only one 1000Mbit port?
Sick.
But then an additional 5 port (cheap!) 1000Mbit switch would solve the
problem, too.

 
 This whole exercise was meant to (possibly) avoid buying new switches
 and solve the issue by software/configuration/bonding/bridging/whateever :-)

As Alkis pointed out, this would work, too.

 Still, this thread was very helpful!
 
 - a perfect solution would be a powerfull managed switch which accepts
 bonding. Then use TWO 1000Mbit cables from the servers two 1000Mbit
 ports to this ONE master switch (that supports bonding). This way, one
 would make absolutely sure that this first superswitch would get ample
 suplly from the bit-bucket :-)
 
 Right?

Yes. There even have been reports on this list that a normal unmanaged
switch will work well when used with linux bonding (doubles the server
send bandwidth, while receive stays the same as the switch sends only to
one port at a time).


Jakob


--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] Red Hat Enterprise Virtualization for Desktops vs LTSP?

[Jakob Unterwurzacher]

Jordan Erickson schrieb:
 Ondrej Valousek wrote:
 *snip*
 
 Also, LTSP/Linux kernel developers should pay more attention to the 
 resource control as that's the biggest gap I am seeing at the moment
 
 This is a very valid question that I've run up against.. is there no
 software for Linux that throttles seemingly out-of-control processes? I
 can see how it would be very difficult for it to be determined as an
 actual runaway process and not something that's just CPU intensive for
 the task it's handling...but per-user/global nice monitor perhaps? Hmm,
 probably man nice would help. Actually, no..it doesn't. ;) But still, hmm.

I think cgroups can do what you want. cgroups exist in Linux since
kernel 2.6.24, so Ubuntu Karmic has it. It can do memory, cpu and i/o
limiting for groups of processes (i.e. one user can be made one group).
I have never used it myself, though.

More info:
http://kernelnewbies.org/Linux_2_6_25#head-450b26e12955b8035a05cf07b3f31c501ee4bfab
and
http://broadcast.oreilly.com/2009/06/manage-your-performance-with-cgroups-and-projects.html

Jakob

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] pxes bootloader on compact flash memory card

[Jakob Unterwurzacher]

Joshua Higgins schrieb:
 title Fedora Core (2.6.12-1.1398_FC4)
 root (hd0,0)
 kernel /boot/vmlinuz-2.6.12-1.1398_FC4 ro root=LABEL=/ rhgb
 quiet
 initrd /boot/initrd-2.6.12-1.1398_FC4.img
 This is useless too, remove.

 
 How big is the CF card? It might have Fedora Core installed on it...

Probably. But it's not needed for booting LTSP so it can be removed to
simplify things.

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] Slowness (mainly with Firefox) on Jaunty

[Jakob Unterwurzacher]

Matthaus Litteken schrieb:
 Hey everybody,
 (Apologies to the moderator for the first post in HTML, won't happen again)
 
 I had LTSP running on 8.04 for a long time, and just recently upgraded
 the server to 9.04. Most everything is running better, and xorg is now
 reading the monitors and setting display resolutions properly, which
 is a nice bonus.
 The downside is that certain actions are fairly slow that weren't
 before. Mostly it is in select dropdowns within firefox, but dropdown
 menus anywhere in Ubuntu are slow to respond. I've read a bit that

Sound like you are seeing this:
https://bugs.launchpad.net/ubuntu/+source/libxcb/+bug/277069

It says Fix Released for Karmic, for Jaunty you will have to use
unofficial packages that include the fix (see comment 34 on launchpad).

Jakob

--
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_
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

Re: [Ltsp-discuss] slow install

[Jakob Unterwurzacher]

Gavin McCullagh schrieb:
 Hi,
 
 On Mon, 14 Dec 2009, David Burgess wrote:
 
 Ubuntu 32-bit has been known to be slow on 4GB of RAM. I would
 strongly encourage you to run a 64-bit OS on the server unless you
 have a very good reason not to.
 
 One of the standard good reasons to use 32-bit over 64-bit was the lack of
 a 64-bit flash player.  However, it would seem that that issue is finally
 getting fixed:
 
   http://www.cyberciti.biz/tips/install-flash-10-ubuntu-linux-64bit.html
   
 http://blogs.computerworld.com/64_bit_linux_adobe_flash_player_surprisingly_good
 
 Gavin

I think this works for some time now with 32 bit flash using
nspluginwrapper. I also tried 64 bit flash but it was too unstable to be
usable.

Anyway, i'd just

apt-get install flashplugin-installer

and it will do the right thing and make flash *just work*, on 32 and on
64 bit. You get updates for flash then, too.


Jakob

--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_
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

Re: [Ltsp-discuss] slow install

[Jakob Unterwurzacher]

joe auerbach schrieb:
 I have not.  I think now that I've heard a bit about it, my plan is to 
 come in this weekend and reinstall the server.  I've still got their old 
 one running and switching them over to it is as simple as rebooting and 
 starting dhcpd on the other system.  Hopefully, I can get the 64 bit 
 version up and running.

64 bit is the way to go, esp. with that amount of ram.

Download the 64 bit alternate CD, just follow this tutorial:
https://help.ubuntu.com/community/UbuntuLTSP/LTSPQuickInstall

I guess your clients will be mostly 32 bit. I'm not sure how clever the
installer is, but you will probably need an additional

sudo ltsp-build-client --arch i386

to make your 32 bit clients work.


Jakob

--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_
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

Re: [Ltsp-discuss] slow install

[Jakob Unterwurzacher]

joe auerbach schrieb:
 I've just started moving people here over to my new LTSP server.  the 
 only one was a graduated desktop machine.  this new one is a legit 
 server with 16 gigs of ran and a 2.33 quad core processor.
 
 So you'd think it would run faster than the old machine, but that is not 
 the case. and I'm not sure why.
 
 I talked a bit about this in the irc forum the other day.  it's running 
 on a fresh karmic install (with all updates).
 
 It just seems to crawl and I'm not sure why.  I don't seem to have 
 tracker installed (that was my first thought).  I am running single nic 
 (though i have two installed on the machine, only one is configured).
 
 What would help are some tests that I can run to help track down the 
 source of this issue.  Is there any such thing?
 
 iotop looks normal.  lots of init, firefox, and squid at the top, but 
 they're not sucking up much processing power.  Is Karmic just slow?  I 
 installed via apt, is that not the best method?

apt is definitely okay.

 
 I haven't installed a version of ltsp since ubuntu 8.04, so I'm a bit 
 out of the loop.

You meant that on the thin clients it's slow?
Is it fast when you login directly at the server?
Is the network saturated (iftop)?


Jakob

--
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_
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

Re: [Ltsp-discuss] booting process

[Jakob Unterwurzacher]

Husain Nagri schrieb:
 
 Now the Problem is  :
 
 a) i Dowloaded the DVD K12LTSP_v5EL-dvd from site given.this is image file.
 b) i burn the image file on dvd
 c) When i run this dvd to setup lstp..this dvd doesnt starts the setup/.
 
 now what should i do to make lstp sever.. even i try downloading all cd
 version ( 7 cds) this cds also doesnt book n start lstp installation..
 
 even when i press enter to take dvd to book..its skips
 
 pls help

Can you boot from any cd (say, a windows cd)? If not, the BIOS settings
are wrong or the DVD drive is broken.
If yes, something is wrong with the dvd/cd burning. Try a different
burner, burning application.

Jakob

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_
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

Re: [Ltsp-discuss] logoff very slow/hangs when using NIC Bonding

[Jakob Unterwurzacher]

john schrieb:
 Hi Jakob et all,
 
 I have half figured out my issue. My bonding related logoff problems
 were caused by my /etc/resolv.conf file being overwritten with
 information handed out by my ltsp client facing dhcp server. The
 second interface that I added to the bond, was getting dhcp
 information meant for thin clients via dhcp and overwriting
 /etc/resolv.conf. The bond still came up, but I had misconfigured
 /etc/resolv.conf file.
 
 This is all happening on Ubuntu 8.04.3/Hardy, btw
 
 What is the right way to keep an interface from being a dhcp client
 without interfering with the bonding process? Google hasn't answered
 this for me yet.
 
 Should I simply add a line like:
 
 auto eth3
 iface eth3 inet static
 
 to /etc/network/interfaces
 
 should I remove dhcdbd from /etc/rc2.d ?
 
 Any advice would be appreciated. Below is my /etc/network/interfaces file

Does adding

iface eth4 inet manual
iface eth5 inet manual

to /etc/network/interfaces help?


Jakob

 
 Thanks!
 
 John
 
 ---
 
 # The loopback network interface
 auto lo
 iface lo inet loopback
 
 # The primary network interface
 auto eth3
 iface eth3 inet static
   address 10.110.0.105
   netmask 255.255.0.0
   network 10.110.0.0
   broadcast 10.110.255.255
   gateway 10.110.0.1
 
 
 auto bond0
 iface bond0 inet static
address 192.168.0.254
netmask 255.255.255.0
network 192.168.0.0
broadcast 192.168.0.255
post-up ifenslave bond0 eth4 eth5


--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_
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

Re: [Ltsp-discuss] logoff very slow/hangs when using NIC Bonding

[Jakob Unterwurzacher]

john schrieb:
 Hi all,
 
 I set up nic bonding today following
 http://www.howtoforge.com/network_bonding_ubuntu_6.10
 
 The bond came up and appears to function correctly, however it
 apparently interferes with log-off on the thin clients.
 I am running Ubuntu Hardy LTS with the default Gnome desktop.
 
 Without bonding I can click the red power icon in the upper right
 corner or go to system quit and immediately be presented with
 the icon which allows me to log off, lock screen, swith user.
 
 With bonding enabled this screen takes more than 30 second to appear
 after being clicked. Clicking log off after the screen appears
 allows me
 to log off immediately. I had a similar issue once when my
 /etc/nsswitch.conf file had the wrong file permissions and DNS wasn't
 being handled correctly, but I
 don't know that nic bonding would matter in this case.
 
 My bond used mode0, I haven't tried the other modes yet.
 
 I need the extra bandwidth but I can't have a 30 second wait before
 students are able to log off.
 
 Does anyone have any ideas about how to proceed?

Get a wireshark dump of the logout process, this will probably reveal
what's going on. Post it if you want me to take a look.

Jakob

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_
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

Re: [Ltsp-discuss] nic bonding vs multiple dhcp scopes on separate nics

[Jakob Unterwurzacher]

Hi!

I will try to answer a few questions in regard to linux' bonding and
switches that don't support it explicitely (i.e. unmanaged, dumb, cheap).

 kind of relying on the switch to do the round robin for you? I guess I
 am confused because( in my tiny brain) mode =0 will result in the
 server presenting the same mac address (e.g. bond0) on two different
 switch ports. Then I guess the switch would add an entry for that mac
 on each port in it's CAM table, and sort of send stuff in both
 directions, depending on how well it can pull this off (being a mere
 dumb switch) things may or may not go well. Am I misunderstanding the
 details?

The server will present the same MAC address on both ports. Each time
the server sends out a packet, the switch will *replace* the table entry
for this MAC address with the current port. So where the switch thinks
the MAC address is will jump around all the time, but it will send the
packets only to one port at a time.

I have been worried that the rapid port-jumping could crash the switch
or something (because it is buggy or poorly engineered), but others have
reported in this thread it works well for them.

 So that's transmit, but I guess I wonder what about receiving traffic?
 Aren't we really relying on the switch to do the right thing?

Receiving traffic will not gain anything here. It will be limited to the
bandwidth of one port.
For LTSP fortunately, this is mostly irrelevant. Traffic is mostly
server - client, so what we really need is transmit bandwidth.


Jakob

--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_
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

Re: [Ltsp-discuss] Pulseaudio dies on thin clients

[Jakob Unterwurzacher]

Steve Cayford schrieb:
 Jakob Unterwurzacher wrote:
 [...]
 Sorry, i missed that pulseaudio actually gives you a helpful error message:

 Oct 29 16:09:02 ltsp-client pulseaudio[3071]: rtpoll.c: Assertion 'usec =
((pa_usec_t) 100ULL)*60ULL*60ULL' failed at pulsecore/rtpoll.c:516,
function pa_rtpoll_set_timer_relative(). Aborting.
 I don't think a backtrace will give additional info, then.
 
 Did you start a bug report anywhere? Not sure if this should go with ubuntu 
 or pulseaudio.

No, didn't. Creating an ubuntu bug is definitely correct as it affects
ubuntu, and reading Red Hat's bug tracker it should be filed with the
linux kernel package as the bug is in the driver.


Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Pulseaudio dies on thin clients

[Jakob Unterwurzacher]

Steve Cayford schrieb:
 I've been losing the audio on thin clients lately. I'm not sure when this 
 started, either when I upgraded to Jaunty or sometime since then. Audio was 
 definitely working with earlier versions.
 
 After initially booting up a client I can usually log in and audio will 
 work for a little while. I'm testing this with the Beep media player and I 
 may get a minute or so of an mp3. Then the pulseaudio server on the client 
 appears to die. I have to reboot the client to get audio again.
 
 Server is Ubuntu 9.04 32 bit. Client is an HP ePC40.
 
 [...]
 
 Any suggestions are welcome.
 
 -Steve

FYI: The problem is common and has been discussed here:
http://marc.info/?l=ltsp-discussm=125319976031611w=2

I still did not have the time to do it myself, but my plan is to obtain
a backtrace ( https://wiki.ubuntu.com/DebuggingProgramCrash - has to be
done on the client, see
https://help.ubuntu.com/community/UbuntuLTSP/UnlockChrootRootAcct ),
report it to the pulseaudio developers and see what they think.

Jakob


--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Pulseaudio dies on thin clients

[Jakob Unterwurzacher]

Steve Cayford schrieb:
 Jakob Unterwurzacher wrote:
 FYI: The problem is common and has been discussed here:
 http://marc.info/?l=ltsp-discussm=125319976031611w=2

 I still did not have the time to do it myself, but my plan is to obtain
 a backtrace ( https://wiki.ubuntu.com/DebuggingProgramCrash - has to be
 done on the client, see
 https://help.ubuntu.com/community/UbuntuLTSP/UnlockChrootRootAcct ),
 report it to the pulseaudio developers and see what they think.

 
 Sorry, I should have seen that thread. I'll try to get a backtrace as well.
 
 -Steve
 

Sorry, i missed that pulseaudio actually gives you a helpful error message:

 Oct 29 16:09:02 ltsp-client pulseaudio[3071]: rtpoll.c: Assertion 'usec =
((pa_usec_t) 100ULL)*60ULL*60ULL' failed at pulsecore/rtpoll.c:516,
function pa_rtpoll_set_timer_relative(). Aborting.

I don't think a backtrace will give additional info, then.


Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Ubuntu + LTSP + HP TC + DELL 17 inch FP

[Jakob Unterwurzacher]

Job Cacka schrieb:
 To All,
 
   I am running Ubuntu Hardy Heron Desktop with LTSP installed via
 synaptic. We are using an HP Thin Client T5545, and a Dell 17 monitor. The
 Thin Client PXE boots to the Ubuntu system, and we then connect to the Dell
 monitor via the DVI port on the TC. The problem is that the video appears
 squashed, like it is displaying at a wide screen resolution instead of the
 standard 1280x1024 or 1024x768 which we normally would use with this setup.
 We have noticed by accident that when we use the VGA port  cable to connect
 to the same Monitor the display uses a standard resolution. 
 
 Why? Different driver or resolution detections in the image file for DVI vs
 VGA ports?

Can't you just switch to the correct resolution in the display settings
control panel (Ubuntu Menu: System - Preferences - Display after
logging in)?

 Also, The thin client appears to react faster to mouse clicks when using the
 VGA port. Is this just our imagination, or is there a way for us to test
 this?

What you see is probably the upscaling delay when you use a non-native
resolution on DVI. It should go away when you set the correct (native)
resolution.

Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Updated: Automated LTSP client shutdown via cron

[Jakob Unterwurzacher]

Jordan Erickson schrieb:
 Hi fellow LTSPers,
 
 I updated a couple of Ubuntu LTSP wiki pages (I'm sure this is fairly
 portable with a few minor changes to other distros) describing how to
 implement automated shutdown for LTSP clients. I've added content on
 implementing a script using Zenity to warn users of the pending
 shutdown, and give them the option to disable it for that day. I hope
 it's of some use to you! Let me know if you find any errors, I *just*
 completed testing myself, though, and all seems pretty good.
 
 https://help.ubuntu.com/community/UbuntuLTSP/AutomatedTCShutodwn
 
 (I know, typo in the URL - blah)
 
 
 Cheers,
 Jordan/Lns

Nice work!

One thing:

I think the script should include

#!/bin/bash

(or whatever interpreter it's supposed to run on, looks like sh or bash
to me)
as the first line.

Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Testing a new switch

[Jakob Unterwurzacher]

Gavin McCullagh schrieb:
 iPerf is a good tool for testing capacity and stressing things a little.
 You can install it on two machines and put the switch in between them.  On
 machine1 run:
 
   iperf -s 
 
 to setup a server and on machine2 run 
 
   iperf -c machine1
 
 to get a client to connect to the server.  By default it'll do a 10 second
 TCP test.  You might want to use '-t 60' in the latter command to extend
 the test to 60 seconds.  If you want to see what's going on, add '-i 1' to
 get ongoing output every second.
 
 You can also do bidirectional tests (probably worthwhile, you should be
 able to get GigE both ways at once) and you can also do UDP tests where you
 specify the bandwidth that should iperf should send at (whereas TCP will
 send as fast as it can without losing packets). You can also run several
 streams simultaneously. Look at the man page for more.

That's exacty what i would do.
Connect a few clients, then run a bidirectional TCP iperf between each
of them for a few hours (overnight?)

On one end
 iperf -s
On the other
 iperf -c IPADDRESS --dualtest -t 36 -i 10 -w 512k -l 512k

Take a look at the troughput you get (should be ~ 1GBit in both
directions when the client is fast enough), if it's the same after a few
hours i would consider the switch good (transmission errors / packet
loss would automatically lower the troughput)

Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] kiosk

[Jakob Unterwurzacher]

David Burgess schrieb:
 Is there a preferred or more-developed method of running a web kiosk?
 I have some thin clients that are running a full desktop, but really
 only need a browser.
 
 I'm thinking about the possibility of automatically launching them
 straight into a fullscreen browser. Looks like Internet Explorer and
 Firefox both have kiosk or fullscreen modes available, but it also
 looks like both are easily escaped using alt-F4 or other means. I'm
 not particularly worried about security at this point (we have other
 measures in place), but I'm concerned that if some user unwittingly
 kills the browser and has no desktop environment to fall back on, the

A wrapper script is an easy solution for this problem (the browser could
crash, so this is a good idea in any case)

#!/bin/bash
while true
do
firefox
sleep 1s
done

( note that this will open new firefox windows indefinitely if you
already have firefox open )

Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Problem with Pulseaudio

[Jakob Unterwurzacher]

Marius Flage schrieb:

 Made any progress?
 
 None whatsoever. I have really no idea where to continue looking. It happens 
 no matter what kind of hardware I'm using, so it has to be some sort of 
 server issue.
 
 I had your setup earlier, with a 64-bit system with 32-bit clients, but 
 stopped using that due to problems with flash. So you may want to check out 
 that. Try playing audio by using mplayer or similar and see what happens 
 then. For me it plays for a while (generally around 1-2 minutes) and then 
 just dies.

I have a 64bit Ubuntu Jaunty Server + 32bit Jaunty clients and i have
the same problem with pulseaudio. Flash works quite well BTW (using
ndiswrapper).

We should obtain backtraces of the crashes
( https://wiki.ubuntu.com/Backtrace#Already%20running%20programs )
then report it to the pulseaudio devs.

There is one open ticket about replay crashes on thin clients, but
that's related to skype ( http://www.pulseaudio.org/ticket/47 ), we
should create a new one i think.

I will try to obtain a backtrace next week, but having several from
different users (platforms? distros?) will probably be more demonstrative.

Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Problem with Pulseaudio

[Jakob Unterwurzacher]

Marius Flage schrieb:
 Jakob Unterwurzacher wrote:
 
 I have a 64bit Ubuntu Jaunty Server + 32bit Jaunty clients and i have
 the same problem with pulseaudio. Flash works quite well BTW (using
 ndiswrapper).

It's *nspluginwrapper* i'm using, not ndiswrapper of course...
( 32bit flash on 64bit server )

 Does audio work for flash as well? And flash works quite well? Try 
 https://pgsc.udir.no/kursweb/pgsUser?marketplaceId=624075languageId=1method=previewTestcontentItemId=731428
  and see if you're able to go through that example test. That was the issue I 
 had with flash that prompted me to switch from 64-bit server to 32-bit.

I'm not sure right now whether flash audio works in the first minute.
Will test at the next opportunity.
Flash generally works quite well - even youtube ist okay on a fast thin
client.
The flash game above works.

 You have the same problem as in you can playback audio for a random while and 
 then suddenly it crashes and you're not able to start playback until 
 rebooting the thin client?

Yes.

 We should obtain backtraces of the crashes
 ( https://wiki.ubuntu.com/Backtrace#Already%20running%20programs )
 then report it to the pulseaudio devs.
 
 Do you run this on the server or as a localapp?

There has to be a pulseaudio server on the client AND on the server - we
don't know yet which one crashes. I'd do it on both. Note that you need
gdb and pulseaudio debug symbols in the client chroot AND on the server.


Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] Slow FireFox Solutions?

[Jakob Unterwurzacher]

 HELP, if I cannot improve performance in the next few weeks, the teacher's
 protests will force me to abandon all LTSP at our district!
 
 You can try to install Firefox 2...

Could work, but Firefox 2 is really old...

I would install Opera ( http://www.opera.com/ ) to make the users happy
for the moment, then try to find a solution to the Firefox problem ( see
Jordan Erickson's post ).



Jakob

--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_
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

Re: [Ltsp-discuss] No login screen with 64 bits

[Jakob Unterwurzacher]

Asmo Koskinen schrieb:
 Mandi Nanga kirjoitti:
 
 Installation was OK. When booting the clients (also 64 bits), dhcp is
 ok (client is getting IP address), tftp is also ok (image is
 downloaded), but after I've a blank screen: no loggin window !!!
 
 Are you sure there is no other dhcp-server, any kind (wlan box), on your
 network?


That's exactly what troubled some of the clients here, with the same
symptopms.

Adding IPAPPEND 3 to pxelinux.cfg solved the issue, see
https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP#Adjusting%20pxelinux.cfg/default
(the page is about ProxyDHCP, but the solution is the same).

 
 Note that I choose 64 bits because 32 bits version didn't reconize 4G
 of RAM.
 
 That's not true. 32G memory and i386 LTSP server work very well
 together, seen that one. And I have two i386 servers, both have 6G.

But any individual process will be limited to 4GB (or 2GB?).
I always wondered wether this also applies to the kernel and the
filesystem cache. Does the fs cache use all availible memory at your
servers?

Regards,
Jakob

--
Come build with us! The BlackBerryreg; Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9#45;12, 2009. Register now#33;
http://p.sf.net/sfu/devconf
_
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

Re: [Ltsp-discuss] Reading lts.conf settings on client

[Jakob Unterwurzacher]

Steve Cayford schrieb:
 This is probably one of those doh questions.
 
 If I set up a custom rc script to run on the client at boot time is there 
 an easy way to read the settings from dhcp and the lts.conf file? e.g. Say 
 I want to get the list of server IP's. It seems like there must be some 
 sort of API or environment settings, but I don't know what they are.
 
 Thanks.

For reading lts.conf there is getltscfg .
For the server IP /usr/share/ltsp/ltsp_config looks promising (this will
also read lts.conf into environment variables AFAIS).

Regards,
Jakob

--
Come build with us! The BlackBerryreg; Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9#45;12, 2009. Register now#33;
http://p.sf.net/sfu/devconf
_
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