Raptor Computing Systems Community Forums (BETA)
Raptor Computing Systems Hardware => Blackbird => Topic started by: dr.chinme on January 12, 2025, 09:15:20 pm
-
The BMC of by Blackbird has gotten stuck during its boot process a couple of times lately, and I'm hoping that folks here can suggest ways to get more information about what's causing the problem.
The first time it got stuck, I disconnected power from the PSU for about a minute and, after reconnecting power, the BMC booted normally. The second time it got stuck, however, I disconnected power overnight, but it still wouldn't boot the next day—the Status LEDs just kept looping a green-blue-green–green-blue-green pattern. I left it powered off for five days, and then it finally booted.
I cannot connect to the BMC via SSH or HTTP/S when it gets stuck like this, so that means I'll have to connect to the Blackbird's internal COM2 header, but I'd have to get the correct cable.
According to the Talos II wiki (https://wiki.raptorcs.com/wiki/Talos_II/Hardware_Compatibility_List#Serial_Adapters_for_J7701_Header), it requires an DTK/Intel cable. Does the Blackbird require this kind of cable as well? It probably does, but I'd like to check.
Are there any other troubleshooting steps that I can try?
-
Yes, it should be the same. See https://www.talospace.com/2020/04/what-to-do-when-bmc-wont-talk-to-you.html for some notes on using it over serial.
-
You could also check the PSU. (https://wiki.raptorcs.com/wiki/Troubleshooting/BMC_Power)
Surprisingly, the PSU was often the cause of such problems.
Edit: Here is also a checklist (https://wiki.raptorcs.com/wiki/Troubleshooting/Support_Request_Checklist).
-
Thanks to you both. I'll get one of those cables and try the i2cget BMC troubleshooting commands.
I suspect the PSU is the culprit as well. The one I have is 10 years old and may indeed be at the end of its life. However, I'd like to confirm before buying a replacement.
-
You don't need any cables. Just ssh into it.
-
Thanks to you both. I'll get one of those cables and try the i2cget BMC troubleshooting commands.
I suspect the PSU is the culprit as well. The one I have is 10 years old and may indeed be at the end of its life. However, I'd like to confirm before buying a replacement.
I would replace that decade-old PSU unless I had no other options. They get unpredictably flaky as their components degrade, and doing so will help a lot in your troubleshooting. Best wishes regardless.
-
You don't need any cables. Just ssh into it.
Unfortunately in this case, the BMC gets stuck before it acquires an IP address, so it's not available via SSH.
Thanks to you both. I'll get one of those cables and try the i2cget BMC troubleshooting commands.
I suspect the PSU is the culprit as well. The one I have is 10 years old and may indeed be at the end of its life. However, I'd like to confirm before buying a replacement.
I would replace that decade-old PSU unless I had no other options. They get unpredictably flaky as their components degrade, and doing so will help a lot in your troubleshooting. Best wishes regardless.
Yeah, I've been considering that.