[Ltsp-discuss] Gigabit switch and booting Image

ben-Nabiy Derush

2016-08-03 15:17:42 UTC

On any machine which I have experienced this, it has been the cabling
in all but one instance. Probably you have a short in the cable
somewhere, which when moved just slightly, it works, and when moved a
little the other way, it doesn't, or it has enough interference that at
GB rates, it is hindered. Replace the cabling (especially if you are
successful when connected direct to your switch.)
--
On Wed, 3 Aug 2016 12:25:27 +0200

Post by LTSP-PNP Adam
I'm using xUbuntu 16.04 and ltsp-pnp. Everithing works perfectly when
I connect server and terminals to Linksys 100Mbit switch.
But, when I connect to Tp-Link (managed or unamanged) or Netgear
(unmanaged) 1000Mbit (giga) switch, the boot process is broken.
PXE boots correctly, image is loading to terminal and starts to boot
up. After 3 second, when running /scripts/init-premount, there is a
r8169 enp2s0: link down
IPv6 ADDRCONF(NETDEV_UP): enp2s0: link is not ready
clocksource: Switched to clocksource tsc
No network interfaces are up
.. after few seconds BusyBox is starting and that is all....
But - this happens only, when I connect terminal to cables that are on
the wall. When I connect directly to switch with patchcord 6 category,
terminal starts ok.
So I think the problem is witch old cables on wall - but I have no
idea, what it can be and how to solve the issue. I need giga-switch
for performance. I know of course, that good solution is to put a new
cabling system category 6 - but school has no money for it. I think
that I will check if there is eia/tia-586 correctly on patchpanel and
outlets on wall, but nevertheless - I would ask if someone had the
sema issue.
Serwer is Dell T130, terminals are MSI Cubi PC's with Realtek PCIGBe
onboard
(https://us.msi.com/product/barebone/Cubi-Mini-PC-Kit.html#hero-specification)
Best regards,
Adam