Supermicro sas2 expander chassi firmware update

[jswede]

I recently bought some used enterprise equipment (supermicro storage servers) and cannot update the sas2 expander firmware. Updating to latest firmware for all other stuff went well without problems (Motherboards, HBAs).

For example, one of the servers is a SMC storage server 6027R-E1R12L, with a X9DRD-7LN4F-JBOD motherboard and a BPN-SAS2-826EL1 expander backplane, and a on-board LSI 2308 HBA flashed for IT mode. There are also 3 HBA cards, LSI 9207-8e for the jbod chassis. All HBAs runs in IT-mode

I have tried using the supermicro ExpanderXtools but it cannot find any expanders
ftp://ftp.supermicro.com/utility/ExpanderXtools_Lite/
Since the tool comes for Linux and windows, I have tried booting from centos, ubuntu 1804, 1204 (since I though perhaps an old OS would help) , and now also windows 10 (I borrowed my kids windows 10 hard drive since I seldom use windows
I have tried connecting the onboard HBA and the separate pci-x HBAs to the expander back plane. I have tried connecting to the backplanes of the SMC jbod chassis I have as well (the have the SAS2-EL2 backplanes, e.g. BPN-SAS2-846EL2)

But the result has always been the same "expander not found"

It seems i have tried all combinations now it seems (well not windows 7).
If anyone here have some tips or info I would be grateful!

Thanks!
Jim from Sweden

(one thing popped up in my head: Perhaps I need to downgrade one of the HBAs to an older firmware? Could this help you think?)

 

[BlueFox]

Why do you need to upgrade the firmware on it? You're probably best off leaving it alone.

 

[jswede]

Why do you need to upgrade the firmware on it? You're probably best off leaving it alone.

Fair question, should have written some text about that.

I see some strange things that perhaps could relate to SAS LSI expander firmware being mismatched with the SAS LSI HBA firmware.
For example, disk ordering (presentation) in freenas seems "buggy"; the boot pool disk is shown as "unused" and another disk among the storage disks are shown as boot pool (and this storage disk is thus missing when you create new pools/vdevs, at least when using the gui)

A question back: Are the firmware in the expanders typically updated when there are updates for the HBAs? (The LSI SAS HBAs have seen quite a few firmware updates over the years, but I have no record of expander firmware, I am new to expanders)

Thanks

 

[raiderj]

I've updated firmware on my HBAs several times over the years, but never on a backplane. My current expander is a BPN-SAS2-826EL1, would be interested to learn if there's a test/update I should perform on it.

 

[BeTeP]

LSI expander firmware can contain bugs but there is no such thing as

LSI expander firmware being mismatched with the SAS LSI HBA firmware

The latest update for SAS2 generation of LSI SAS expanders is dated circa 2013 IIRC.

 

[jswede]

I've updated firmware on my HBAs several times over the years, but never on a backplane. My current expander is a BPN-SAS2-826EL1, would be interested to learn if there's a test/update I should perform on it.

Ok, thanks for replying, good to know!

 

[jswede]

LSI expander firmware can contain bugs but there is no such thing as


The latest update for SAS2 generation of LSI SAS expanders is dated circa 2013 IIRC.

Ok no possibility for mismatch!
That is interesting, care to elaborate? (I guess it has to do with the expanders being much "simpler" than HBAs, and more easy/"robust" to implement?)

Latest sas2 expander firmware was 2013, good to know - thanks!

 

[BeTeP]

I do not really enjoy explaining very basic things too much but here it is:

You are probably thinking about potential (and quite real) mismatch between LSI IOC driver and LSI IOC firmware versions. Which is different from the "mismatch" between LSI SAS expander firmware and LSI IOC firmware versions you invented.

The difference is that in the first case a single vendor (i.e. LSI) controls both sides (driver and IOC) so they are free to use a proprietary communication protocol and make any changes to it when they see fit. Obviously both sides have to support the same version of the protocol. Hence the potential mismatch.

In the second case the sides (SAS IOC and SAS expander) communicate using industry standard protocols. No single vendor controls it. If both sides are compliant - everything works as it should. If something is not working - the cause of the malfunction is called a "bug", not a "mismatch".

In other words, the mismatch is always user's fault. Vendors introduced the change and released the requirements for the upgrade. The user did not follow through.

The software bugs are vendor's fault. They released a non-compliant product.

 

[jswede]

I do not really enjoy explaining very basic things too much but here it is:

You are probably thinking about potential (and quite real) mismatch between LSI IOC driver and LSI IOC firmware versions. Which is different from the "mismatch" between LSI SAS expander firmware and LSI IOC firmware versions you invented.

The difference is that in the first case a single vendor (i.e. LSI) controls both sides (driver and IOC) so they are free to use a proprietary communication protocol and make any changes to it when they see fit. Obviously both sides have to support the same version of the protocol. Hence the potential mismatch.

In the second case the sides (SAS IOC and SAS expander) communicate using industry standard protocols. No single vendor controls it. If both sides are compliant - everything works as it should. If something is not working - the cause of the malfunction is called a "bug", not a "mismatch".

In other words, the mismatch is always user's fault. Vendors introduced the change and released the requirements for the upgrade. The user did not follow through.

The software bugs are vendor's fault. They released a non-compliant product.

OK got most of it, thanks for explaining! (I am a more or less a newbie with this stuff, but interested to learn, so be nice )
So, a follow up question, if I may; is it so that the expander requires no driver? (the expander is just "relaying" data according to the old SAS2 standard protocol? that would explain the old firmware I guess? no need to update since no OS/driver dependency?)

 

[jswede]

Update: I finally got the Supermicro ExpanderXtools_lite up and running (I used Centos 6.3 and you have to run smc and xflash as sudo). Contacted Supermicro support who kindly send me the latest firmwares with the correct mfg binaries for the different backplanes (Latest rev firmwares are 55.14.18.01 for BPN-SAS2-846-EL, and 55.14.18.0 for BPN-SAS2-847-EL and BPN-SAS2-826-EL, make sure you get the correct mfg binaries as well)

Installed freenas again and now the glitches I saw before are gone (for example, boot pool shows as "Unused").

Now I will need some some advice as how to configure pools for a good tradeoff between speed, robustness, etc. (98 drives in 3 jbods).
But I will post that question in the Freenas forum.

Thanks for replies/info!

 

[MisterE]

Contacted Supermicro also. Information of jswede is somewhat incomplete.
Latest firmware of 55.14.18.01 is for backplane with the revision ending with 'A'. For example, BPN-SAS2-846EL2 rev1.11A
Latest firmware of 55.14.18.0 is for backplane 1.10

 

[BackupProphet]

Could you post the firmware and process to update the backplane here. Thanks

We can also help with how to configure pools, a lot of the stuff I've read on FreeNAS forum is wrong and bad. So you can create a thread here too, there are many ZFS users here, and I've used ZFS for almost 15 years.

 

[MisterE]

Sure, included 55.14.18.0 for backplane 1.10.
Can not really help with the steps required. Missing some critical components for my new nas so i even do not know the current firmware.
But as this is a new setup i like to know if everything is up to date.

The next step is to verify the version using this steps:

FAQ Entry | Online Support | Support - Super Micro Computer, Inc.

Frequently Asked Questions

www.supermicro.com

 

[MisterE]

learned some new stuff. Will add it to CLIXTL results in "Error, device doesn't contain required data"

 

[Cipoint]

I have several BPN-SAS2-826EL1 and BPN-SAS2-826EL2 Backplanes with Rev. 1.01. MFG Ver. 02.29 and FW Ver. 55.07.23.00. So quite old.
Some of the backplanes are unstable when all HDDs are at full load. I'd like to try a FW update if possible. Anyone has the firmware for Rev. 1.01?

 

[i386]

be carefull when updating the firmware of sas expaders on backplanes, there is this thread https://forums.servethehome.com/ind...ane-firmware-update-problem.27149/post-250433 where a user bricked his

 

[Cipoint]

Thanks for the warning. I received the latest firmware from supermicro. They also warned me about bricking the backplanes if not done correctly. I'm looking for a tested and confirmed way to flash the firmware, probably witht the xflash tool.