This wasn't a Sun Ray problem. It was due to my little knowlegde of CDE desktop confirguration. The accounts had been set-up with dual head config. To solve the problem for all the accounts I simply copied ~/.dt/sessions/dtwnfp.session1 to dtwmfp.session[2-5], and copied ~/.dt/types/fp_dynamic1 to fp_dynamic[2-5]
Copying the dtwnfp.session and fp_dynamic instead of the *1 didn't work.

S

LOEWENTHAL Simon wrote:


Checked our errorlog and it these I/O errors . Such examples for the 6 head policy are: (Don't know if this is related)


Thu Mar 09 09:51:53 2006
Workspace Manager: I/O error on display:: :2.0

Wed Mar 08 21:47:21 2006
Workspace Manager: I/O error on display:: :2.0

Wed Mar 08 13:14:35 2006
Workspace Manager: I/O error on display:: :5.0

Wed Mar  8 14:13:42 2006 (/usr/dt/bin/dtexec) /home/ssa/bin/tsd

LD_LIBRARY_PATH: Undefined variable

Wed Mar  8 15:29:12 2006 (/usr/dt/bin/dtexec) /usr/local/bin/color-xterm

Warning: locale not supported by C library, locale unchanged


LOEWENTHAL Simon wrote:

I deleted the group again and tried with the list of DTUs in the reverse order, but now only one screen is usable (the primary) and the rest have the Sun logo on the top left hand side. When the user logs in he can use only one DTU. Commands used were:

# utmhadm -d mkoSSA6x2
# utmhadm -a mkoSSA6x2 -g3x2 -p IEEE802.0003baffabd5 -l IEEE802.0003badc2cca,IEEE802.0003baff9826,IEEE802.0003badc2f6e,IEEE802.0003badc2a12,IEEE802.0003badc2a33,IEEE802.0003baffabd5

# utmhadm -d mkoSSA6x2
# utmhadm -a mkoSSA6x2 -g3x2 -p IEEE802.0003baffabd5 -l IEEE802.0003baffabd5,IEEE802.0003badc2a33,IEEE802.0003badc2a12,IEEE802.0003badc2f6e,IEEE802.0003baff9826,IEEE802.0003badc2cca

Very bizarre.  Tried rebooting the server but no change.



LOEWENTHAL Simon wrote:

I deleted the group and created it again with the the primary DTU position changed as you said, and still the same problem. DTUs 1 and 2 are OK (IEEE802.0003baffabd5,IEEE802.0003badc2a33), but the remainder are not. So, no change :(

# utmhadm -a mkoSSA6x2 -g3x2 -p IEEE802.0003baffabd5 -l IEEE802.0003baffabd5,IEEE802.0003badc2a33,IEEE802.0003badc2a12,IEEE802.0003badc2f6e,IEEE802.0003baff9826,IEEE802.0003badc2cca
1 Group for multihead added.

[EMAIL PROTECTED] utmhadm

Multihead Group Geometry CIDs ------------------ ------------------ --------------------------------
mkoSSA6x2          geometry=3x2       IEEE802.0003baffabd5 (P)
                                     IEEE802.0003badc2a33
                                     IEEE802.0003badc2a12
                                     IEEE802.0003badc2f6e
                                     IEEE802.0003baff9826
                                     IEEE802.0003badc2cca
                                  1 group total.

Craig Bender wrote:

Should the first server in the -l group be the same as the primary?

LOEWENTHAL Simon wrote:

More info:
Here are the DTUs that have the correct workspace:
IEEE802.0003baffabd5
IEEE802.0003badc2a33

The others appear with the default.

The command that I used to create the policy was:

# /opt/SUNWut/sbin/utmhadm -a mkoSSA6x2 -g3x2 -p IEEE802.0003baff9826 -l IEEE802.0003baffabd5,IEEE802.0003badc2a33,IEEE802.0003badc2a12,IEEE802.0003badc2f6e,IEEE802.0003baff9826,IEEE802.0003badc2cca

# utmhadm

Multihead Group Geometry CIDs ------------------ ------------------ --------------------------------
mkoSSA6x2          geometry=3x2       IEEE802.0003baffabd5
                                    IEEE802.0003badc2a33
                                    IEEE802.0003badc2a12
                                    IEEE802.0003badc2f6e
                                    IEEE802.0003baff9826 (P)
                                    IEEE802.0003badc2cca
                                 1 group tota_l.
_

LOEWENTHAL Simon wrote:

Thanks to everybody for their suggestions: I disabled Ximeria and the load dropped back down to normal. I removed the shmsys:shminfo_shmmax from the /etc/system. Ximera and 8-bit displays might have been too much for the 440.

Another problem:
After I turned off Ximera something stange happened with CDE. The user's CDE workspace appeared correctly on two of the displays, but on the others the default CDE workspace was shown, without any of the custom menus we had configured. Has anyone else seem this?
The CDE errorlog gives little information.

Simon.



Craig Bender wrote:

Good call.  Didn't see it at first, Now I see he mentioned JDS.

Rob Giltrap wrote:

Are you running Solaris 10? if so I believe this parameter has been deprecated.

Refer: http://docs.sun.com/app/docs/doc/817-0404/6mg74vsb0?a=view

LOEWENTHAL Simon wrote:


So for our 6 heads running at 1280x1024 then the setting would be:

6*1280*1024*4 = 31457280

Is this the correct calculation?

Craig Bender wrote:

Not sure it it's related to this issue, but there is a tip in the admin guide regarding shmsys:shminfo



Tip - Because XINERAMA consumes a lot of CPU, memory and network bandwidth, please set the shmsys:shminfo_shmmax parameter in the /etc/system file to at least LARGEST_NUMBER_OF_HEADS * width * height * 4 for reasonable performance.

LOEWENTHAL Simon wrote:


Cheers Shivu.  I shall reconfigure the policy.

On a related, yet different note:

With a 6 head policy Xsun is running at about 22% when the user's CDE starts up, and hits 60% when the user launches firefox. Once firefox appears Xsun's CPU activity drops back to normal. We have tried this with 3 different user accounts. While this happens the mouse judders around the screen, when one moves it.

I tried this with JDS, but gave up after ten mins...

Is this a problem with using high multihead policies? The server is Netra 440 (SunFire 440) 2 x CPU UltraSPARC III and 16Gb of memory? There is only SunRay s/w and one user using the system.

Ideas?

S


Shivu Vibhuti wrote:



LOEWENTHAL Simon wrote:

Thank-you everybody who relied.

The command works:
utmhadm -a mkoSSA6x2 -g3x2 -p IEEE802.0003baff9826 -l IEEE802.0003baffabd5,IEEE802.0003badc2a33,IEEE802.0003badc2a12,IEEE802.0003badc2f6e,IEEE802.0003baff9826,IEEE802.0003badc2cca


Interestingly enough, the keyboard/mouse is attached to the primary IEEE802.0003baff9826, but dtlogin is displayed on IEEE802.0003baffabd5.

The CID list provided with the "-l" options determines the screen order (left to right), In your case IEEE802.0003baffabd5 is the screen 0, where dtlogin GUI appears.

-Shivu

Also, when you log out not all the DTUs are updated, even if you log in again. The update from displaying black only happens when one moves the mouse onto the DTU's display area. All DTUs are the same (Sun 170s).

Any clues about this?

Many thanks, Simon.

Shivu Vibhuti wrote:



Shivu Vibhuti wrote:

You don't











you don't  need

Smart Cards to setup Multihead Policy, use utpolicy CLI or Admin GUI to enable multihead session capability

And to create the Multihead Groups, you can either use AdminGUI or utmhadm CLI

#/opt/SUNWut/sbin/utmhadm -a <groupname> -g <COLSxROWS> -p <primaryCID> -l <CID list>

CID is in the form of IEEE802.<DTU MAC addr> -Shivu


LOEWENTHAL Simon wrote:

Hi there,

We've arrived at a remote site (on a mountain top at 4300metres) to set-up some DTUs, but someone forgot the smartcards. Does anyone know how to configure a multihead policy on the command line?

We cannot get anyone else to come up here an delivery some cards until tomorrow :(

Cheers, S.
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users















_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users











_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users









_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users









_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users







_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users








_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users







_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users






_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users






_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users





_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users




_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users




_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users



_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users



_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users

_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users

Reply via email to