Hi Petter, My setup is quite easy. I installed debian-edu 6 on a HP Proliant gl380 G7 server with 12GB memory and two four kernel procs on 3.47Ghz so I guess this server should be enough. On serverside things seems to work.
Now I try to make the clients work properly. Not with much success though... The first computer I tested was an old Pentium 2 computer with, as you say, unknown hardware. That one works both as a thin client and as a diskless workstation. I have a small book sized pc (Lex systems). Quite rare, but even that one works nicely. Then I tried the HP t5530 thin client. It boots nicely, but when trying to log in it tries to start a kde session, but fails. So after about 10 seconds it pushes you back to the login screen again. I tried to ssh into it, and even hit ctrl-alt+F1 that gives me a console screen. But I cant login either via ssh or on the console (F1). It says wrong credentials.. But logging in on the welcome screen (F7) works when it comes to credentials, it just dont start the kde session. Strange thing, this HP t5530 worked with debian-edu 5... I also have a HP t5630, as you mentioned, but that one works also. It is the t5530 that I cant login to. Im now trying to figure out how to see what is missing and what to do about it. Im not used to this chrooted environment, thus all the questions. Before I have used the Thinstation project to netboot the clients and there I know how to add extra drivers and such... (and to see where it goes wrong.. ) But its always good to learn new things ;) In syslog I get this when trying to login to the malfunctioning client (t5530) (from the moment I hit login till Im back on the login screen again.) Mar 16 12:52:55 tjener ldminfod[30698]: connect from 192.168.0.20 (192.168.0.20) Mar 16 12:52:56 ltsp20 acpid: client 3163[0:0] has disconnected Mar 16 12:52:56 ltsp20 acpid: client 3163[0:0] has disconnected Mar 16 12:52:56 ltsp20 acpid: client connected from 3895[0:0] Mar 16 12:52:56 ltsp20 acpid: client connected from 3895[0:0] Mar 16 12:52:56 ltsp20 acpid: 1 client rule loaded Mar 16 12:52:56 ltsp20 acpid: 1 client rule loaded Mar 16 12:52:56 ltsp20 kernel: [ 840.474414] agpgart-via 0000:00:00.0: AGP 3.5 bridge Mar 16 12:52:56 ltsp20 kernel: [ 840.474414] agpgart-via 0000:00:00.0: AGP 3.5 bridge Mar 16 12:52:56 ltsp20 kernel: [ 840.474472] agpgart-via 0000:00:00.0: putting AGP V3 device into 8x mode Mar 16 12:52:56 ltsp20 kernel: [ 840.474472] agpgart-via 0000:00:00.0: putting AGP V3 device into 8x mode Mar 16 12:52:56 ltsp20 kernel: [ 840.474581] pci 0000:01:00.0: putting AGP V3 device into 8x mode Mar 16 12:52:56 ltsp20 kernel: [ 840.474581] pci 0000:01:00.0: putting AGP V3 device into 8x mode and after this you are just back to the welcome screen again. Im trying to fix this one now, since it worked on debian-edu 5, but problem, since I cant login on a console I cant see the logs on the client.. Shouldnt it use the same credentials when you hit ctrl-alt+F1 as it does when logging in via the welcome screen? Regards /George --- Den fre 2012-03-16 skrev Petter Reinholdtsen <[email protected]>: Från: Petter Reinholdtsen <[email protected]> Ämne: Re: Login problems squeeze Till: "George" <[email protected]>, [email protected] Datum: fredag 16 mars 2012 11:44 [George] > Hi, > > I checked syslog. I have now checked the others you mentioned also.No > it was not on the same machine. It is a HP t5630 thinclient that I > netboot. It worked nicely with debian-edu 5 yesterday ;) but of course > something might have changed in version 6. I am not quite sure I understand your setup. Let me know if I got it right or wrong. You have at least three computers, one installed as a main server + thin client server (unknown hardware), one used as a diskless workstation (HP t5630) and one used as thin client (unknown hardware). Both the thin client and the diskles workstation machines work fine when booted as thin clients, but the diskless workstation crash when the first user try to log in after boot. I have not seen such problem myself, and have no experience nor recommondation. :( Could the machine run out of memory? Is there any information in /var/log/syslog on the main server about the client when the client crash? -- Happy hacking Petter Reinholdtsen
