I did open a support ticket with Supermicro and somewhat surprisingly they did respond and send me the firmware update I needed to perform the firmware update.
Folks, I have some server blades that have the AOC-MTG-i4S NIC on them. I'm having LLDP problems and wanted to upgrade the firmware. Supermicro's page for the card has firmware available (9.0 and 9.4) but they don't perform the upgrade. Both require version 7.x as the starting point and all...
I did want to follow up here just in case - I have a third sled as I mentioned before which is configured basically the same, uses the same OS install, but doesn't have a PM951 SSD in it.
I got hold of a pair of Xeon 6234s. Pulled the 4114's and replaced them. Did nothing special. Booted the...
Thanks all, I think (hope) this is now fixed.
TrevorH - Great suggestion. I was convinced the temp alerts were bogus but "prove it" is certainly a valid test. I didn't feel like swapping the SSD so to run the test on BladeB I booted off the Alma Live USB - and the throttling errors happened...
Nope. This machine is barely loaded. The heaviest thing it does is run a couple of small VMs. So I disabled autostart on the VMs, I set the fan mode to Heavy IO to beef up the cooling and power cycled the blade.
It's sitting here right now throwing throttling error events every 2 seconds only...
So I've done a lot here, let me back up and explain how I got here as it may help. This chassis has 4 blades in it. 2 are configured identically from a hardware and software perspective. They run Almalinux 9.6 and are used for infrastructure services. A third blade is also running Almalinux 9.6...
Well this is super weird. I just started trying random things to see if I could figure this out.
I didn't try reflashing the BIOS, so I gave that a shot, making sure to deselect all the "preserve" options to wipe everything out. Rebooted the machine and now the hardware inventory in the IPMI...
The blade's been out a dozen times at least, so if a cold boot is what's supposed to do it, it hasn't.
I tried this also, since the board does have an on-board SAS RAID backplane controller. Took everything out and booted up to the EFI firmware shell. No change.
So I got annoyed and put the...
Can you elaborate on what a normal boot consists of? I think this might be a clue as to what's going on, because I had my friend check his machine that was previously doing the same thing, and now he says the BMC reports the correct CPU installed and he didn't do anything.
The blades in...
Hi folks, have a blade chassis with X11DPFR-SN motherboards. I upgraded the procs in 2 of them from 4114 to 4214. After I did that, the heath logs filled up with constant CPU1 overtemp errors even though the sensors show everything is normal. When I check the BMC's web interface and look at the...
Hi all, can anyone confirm what the latest version of EOS is for a 7050QX-32? From what I can tell the device is locked down to use 2GB images even if your flash is bigger.
Actually it looks like there's no way to unlock the 2GB limit (is that memory size?) So I think I'll just ditch this...
And I guess the final note here, I swapped out the 9300 series HBA with a 9400 series and the tape library works now. Though not through the expander. Something is still weird with Windows and that SAS expander. Directly connected it works fine, and I can live with that.
This is interesting if anyone's curious. I found a second-gen LSI SAS card, A 9207-8e. I plugged that in and cabled it up, and the tape library and drives appear as they should. Seems like a driver problem with the 3008 which is unfortunate as these are very commonly embedded on server boards...
So after some more debugging- Drivers are "current" - the SAS3008 drivers haven't been updated since the P16 set in 2020. The drivers MountainBOFH is pointing to are for newer chips than the 3008.
The missing drive turned out to be a cabling problem apparently. I recabled everything and that...
Here's what I learned. Arista will give you a key to override the module detection if you have a valid support contract, or maybe ask them nicely I guess. The format of this config command is:
service unsupported-transceiver name key
so it would look something like
service...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.