12-21-2018 12:29 AM
We have detected the issue with BMC (IPMI) on our Lenovo RD640 servers.
The main issue is: when we try to reboot the server via IPMI or external IP-KVM module (Lantronix), BMC module loses network connection (port link become down for a moment), and Management Interface becomes unavailable, as well as ping to the BMC IP address, and we can not reach the server's management interface. The reboot can be performed via BIOS / BMC management interface in the Web or via Ctrl+Alt+Del in the console.
We tried to update BIOS and BMC Firmware to the latest version, but the issue is still persists. We also tried BIOS/BMC settings reset and BMC cold reset, but it did not help us.
It should be noted, that there is no difference, if the BMC works in 'Shared with Lan port' mode or 'Dedicated' mode. The result it the same. The standby voltage more then 5V, so seems like the power supply is okay.
Actualy, we have several RD640 servers, but the issue appears just on two servers.
May be somebody know if it is a power of configuration issue, and what we can do with that?
Solved! Go to Solution.
12-26-2018 12:27 PM
Welcome to the community!
How are you confirming the port goes down? Can you check the switch port and see what it is reporting? Does it stay down and recover after a reboot or does it go down and come back up?
12-27-2018 01:32 AM
01-02-2019 06:09 AM
Does a certain action trigger the disconnect? how often does the port bounce? Can you try replacing the network cable and/or try a different port on the switch? To see if the issue follows.
01-10-2019 06:48 PM
Suggest to check if BMC has restart when the link down/up happen. And suggest to check if there have network broadcast storm occurs. I have a experience that the network storm may cause BMC restart, and the management port will have a down/up action during BMC restart. Hope this can help you, thanks.
01-11-2019 04:59 AM
I have a RD640 and the same thing happens on the BMC, the connection will drop but only when rebooting the server. Not sure what the reason is for this, but it's exactly as the OP is seeing. Would like to know what I can check on the RD640 (settings wise) that is causing this, as it's clearly a problem on the unit.
01-20-2019 01:07 AM
Apologize for the silence.
Seems like the issue was in the network switch's port. We have changed it and the issue seems is resolved now.
I hope the second server's BMC will be okay too.
01-22-2019 08:12 AM
Glad to hear! I would check on the switch side to see if there are any CRC errors on the port. This usually indicates either a bad cable or a bad port/sfp and can cause some inconsistent behavior.
If you feel like this post has been resolved would you please mark it as complete.