dp 8000 not booting

We were updating hog 3 from 3.1.2 to 3.1.4.
Console software updated ok and working.
Dp2000 software updated ok and working.

Then the dp 8000 didn´t start to update.
So we tried forcing with pressing left button from dp8k when powerup. Nothing.
We ended up trying pressing all three buttons while powerup.
End result was that the dp8k isnt booting up any more at all.

If it is booted with all 3 buttons pressed the following happens:

First there is a lot of boot info runnning and eventually it goes to loop written below:

PXELINUX 3.71 3.71 Copyri
ght (C) 1994-2008 H. Peter
Anvin
Loading dp8000/vmlinuz-2.6
.26.5-dp.8k.....................
........................
Loading dp8000/intramfs-d
p87k_1.0_i386.cpio.gz........
.............................
.............................
....................
I/O Board Disconnected

At this point the Hog-Net Link light turns off and back on in approx 5 secs.


I/O Board Connected

And then dp goes back to beginnig of PXE... (loop above)


Then if is just turned on OR hard reseted (with center and right while poweron) a lot of info comes and everything stops to:

eth0: 10/100 speed: disabling TSO

and then nothing...

Havent had similar problems previously.
DP8k worked just fine before this.

Any ideas? Running out of buttons to press while powerup...
Parents
  • 0
    In most cases, when you upgrade the console, it will detect that there are nodes on the network with old software, ask if you want to update them, and install the new software. If your network is setup properly, it will work without any network changes.

    The procedure I'm describing should only be used when you have a DP with corrupt firmware, or when an upgrade fails irrecoverably. Even then, the failsafe recovery procedure should work across a network.

    The reason I recommend using a crossover cable when you need to recover from a failed upgrade is that it eliminates any external variables. I frequently see networks that have rogue DHCP servers, or other devices that the operator isn't aware of. Spanning tree related delays can prevent PXE from working properly and cause the netboot to timeout before the switch port has entered the forwarding state. Some network management features intercept PXE and don't pass it along to the console.

    By using a crossover cable, you eliminate all of these variables and it may help you get your unit back online more quickly after an interrupted upgrade.
Reply
  • 0
    In most cases, when you upgrade the console, it will detect that there are nodes on the network with old software, ask if you want to update them, and install the new software. If your network is setup properly, it will work without any network changes.

    The procedure I'm describing should only be used when you have a DP with corrupt firmware, or when an upgrade fails irrecoverably. Even then, the failsafe recovery procedure should work across a network.

    The reason I recommend using a crossover cable when you need to recover from a failed upgrade is that it eliminates any external variables. I frequently see networks that have rogue DHCP servers, or other devices that the operator isn't aware of. Spanning tree related delays can prevent PXE from working properly and cause the netboot to timeout before the switch port has entered the forwarding state. Some network management features intercept PXE and don't pass it along to the console.

    By using a crossover cable, you eliminate all of these variables and it may help you get your unit back online more quickly after an interrupted upgrade.
Children
No Data
Related

www.etcconnect.com