Lenovo 430-16i flash and cache logging issues

TrueNAS General
1

I flashed my 430-16i last night to the latest LSI 9400-16i firmware and bios I could find on the Broadcom site:

  • FW Package Build: 24.00.00.00
  • FW Version: 24.00.00.00
  • BIOS Version: 09.47.00.00_24.00.00.00

Since reinstalling in my server, I’m seeing a LOT of these via dmesg:

mpt3sas_cm1: log_info(0x31200205): originator(PL), code(0x20), sub_code(0x0205)

Looking up the sub_code, and I found:

0x0205 Information “%s Cache Flush aborted by user” Logged when the user aborts the Cache Flush operation

Anyone know what this is about? I also had the server hard lock overnight, so I’m trying to investigate whether that is a related issue or not.

Note that I’m on the latest version of SCALE.

2

So, I ended up fixing this hopefully. I managed to find the latest driver package on Lenovo’s support site, after much pain. I then flashed to that using the same process. dmesg is nice and quiet now. If it helps anyone, I just did up a quick gist as I’ll likely need to do this again myself.

lenovo_430-x_hba_firmware_BIOS_update.md (github.com)

3

Did you have to do the flashing on Windows? Or just download and extract the files somehow? Just wondering whether I have to remove the 430-16i HBA from my TrueNAS box and flash it on my Windows PC.

4

Hi,

It seems you’ve flashed it in the proper way though EFI either the official Broadcom firmware: why do you have those cache log error ?

internet is full of references that cross flashing in this unit is more than ok, and even “art of server” sells this HBA leonovo 430 16.

Maybe Crossflashing is not actually supported and everyone has your same errors without noticing it ?

5

It’s difficult to parse what you are trying to ask here, but my interpretation of buswedg’s posts was that they first tried the 24.00 firmware from Broadcom and got message spam in the system log. Then then switched to 24.05 firmware from Lenovo, after which their issue with the message spam was fixed.

Is something stopping you from doing the same?

6

I tough I was clear, sorry. I try to state it in different way:

if Lenovo 430-16i is 1:1 compatible with Broadcom 9400-16i, how is possible that generates errors with the Broadcom firmware ?

Even worst that updating it with the Lenovo firmware those errors disappear, its additional prove to some strange incompatibility with official Broadcom FW.

7

The Broadcom and Lenovo firmwares were different versions, 24.00 vs 24.05.
That may, or may not, be the reason for the message spam.

As far as as I can tell, the available information doesn’t make that clear. You are jumping to conclusions.

8

So your assumption is that even the official Broadcom with FW 24 throws the same error logs in the kernel?

yes that could be an option: quite strange given the fact that this HBA was one of the suggested and I couldn’t find any other references to this error online except for Lenovo 430 version… but yes I couldn’t deny it.

9

No. I am not assuming anything. I am surmising that there isn’t enough information to draw conclusions one way or the other, at this point.

Do note that I am also not agreeing or disagreeing with your statement that the Lenovo card is 1:1 compatible. I haven’t assimilated enough information to comment on that. It’s not relevant to me so I will not spend more time researching it.

10

And even if the cards were different, what would be the reason not to flash a Lenovo 430 with Lenovo’s firmware?

11

That show cache log error as said before ?

You want to make Work your super uber 2k euro over an HBA board that throws random cache error for no reason?

No thanks.

12

Errors apparently came from cross-flashing LSI firmware on a Lenovo card, and were fixed by flashing Lenovo firmware.
So LSI on LSI and Lenovo on Lenovo: What’s the problem with that?

13

That compatibility has been given and proved with 9400-16i board…I don’t find any historical usage data for the Lenovo 430 that could be reliable with truenas.