LTSP Client specific boot trouble.
On a newly installed Debian-Edu 9 with LXDE as destop, some of my
newer(faster) diskless clients fail to boot properly.
This seems to be a sort of race condition. Sometimes it happens early in
the boot. That is rare though, so i dont have logs for that yet, but i
remeber something bout "SQUASHFS". Most of the time the client gets a
black screen when ldm starts up. When that happens, I can alt+ctrl+F2 to
get another session, log in an look at logs. Syslog allways ends with:
"block nbd9: NBD_DISCONNECT"
"block nbd9: shutting down socket"
Doing: cat syslog | grep error | less
I get some failed driver messages for wifi and bluetooth that i think is
irrelevant. Then i get:
"blk_update_request: I/O error, dev nbd9, sector 7002672"
"Buffer I/O error on dev ndb9, logical block 3501336, async page read"
ldm.log contains 2 entries:
"INFO: started on client with IP address 10.0.10.1"
"INFO: calling rc.d init scripts"
I found these 2 webentries that i think talk about the "2"
raceconditions i have:
I realy hope someone can help with this, cus i think it will affect
everyone. I have replicated exact same result on 3 different
installations( 3 different servers, 2 new and one old) and with a lot of
dirrerent clients. Slower clients seems less affected than newer faster
Helge Tore Høyland