I have two UnRAID servers, one was running v7.2.4 and the other something really old, like maybe 10-14 versions older (can't remember since the recent upgrade).

I upgraded both to v7.3.0. One server lost access to ALL drives, the other humming along like nothing changed.

The BAD: SuperMicro X11SPH-nCTPF with onboard LSI3008 with IT-mode FW. Upgraded to 7.3.0 and 7.3.1 and both versions all drives no longer show up anywhere after the upgrade, not even in the logs. Downgrade back to 7.2.4 and all happy again.

The GOOD: SuperMicro X10DRH-iT with PCIe SAS card (I think either an LSI 9211-8i or IBM M1015). All drives happy as can be under 7.3.0 and 7.3.1

Both MB's utilize SAS expanders of the chassis' they're installed into (see sig)

Before I go a hollerin' to LT and report these unacceptable BUGs, is there perhaps something I'm missing, a certain setting or something else?

I'm thinking there has to be some sort of incompatibility with the built-in SAS controller of the first system.

TIA!

UPDATE: I upgraded the first server to 7.2.7 and all's well. So it appears to be the jump from 7.2.x to 7.3.x.

Edited May 30May 30 by Auggie

Please post the diagnostics after the upgrade.

Perhaps 7.3.1 as Jorge mentioned might have a regression for LSI 3xxx cards? I had to revert to 7.3.0 to fix my install as well.

Ref: https://forums.unraid.net/topic/199098-unraid-730-731-array-startperformance-related-issues/

Skip

7 hours ago, JorgeB said:

Please post the diagnostics after the upgrade.

Will have to wait 3 days as one of my drives was disabled so have to do the ol' rebuild first...

18 hours ago, Skipdog said:

LSI 3xxx cards?

It's not all 3xxx, since 9300-8i works fine, at least for me, but it could be with the -16i and -24i models.

I have -16i shows as two 8port controllers

IOMMU group 26:				[1000:0097] 09:00.0 Serial Attached SCSI controller: Broadcom / LSI SAS3008 PCI-Express Fusion-MPT SAS-3 (rev 02)	8/8 GT/s	x8/x8(Gen:3)
IOMMU group 27:				[1000:0097] 0b:00.0 Serial Attached SCSI controller: Broadcom / LSI SAS3008 PCI-Express Fusion-MPT SAS-3 (rev 02)	8/8 GT/s	x8/x8(Gen:3)

7 minutes ago, SimonF said:

I have -16i shows as two 8port controllers

That is a 9300-16i right?

I think the issue may be just 9305-16i and 24i, those use a SAS32xx chip instead of one or two SAS30xx.

2 hours ago, JorgeB said:

That is a 9300-16i right?

I think the issue may be just 9305-16i and 24i, those use a SAS32xx chip instead of one or two SAS30xx.

Yeah 9300-16i

On 5/29/2026 at 11:43 PM, JorgeB said:

Please post the diagnostics after the upgrade.

Attached are 7.3.1 diagnostics, 7.3.1 syslog and 7.2.7 syslog.

None of the drives attached to the LSI3008 show up in the 7.3.1 syslog but there are no errors. Drives attached to SATA and USB ports on the MB are unaffected.

I reviewed the change logs for both 7.3.0 and 7.3.1 and there is no specific mention of changes involving HBA cards/chipsets, though are quite a few files that are updated and I have no idea what each of the individual files affect which hardware component(s).

unraid-diagnostics-20260602-0634.zip syslog-7.3.1.txt syslog-7.2.7.txt

Edited June 2Jun 2 by Auggie

On 5/30/2026 at 6:26 AM, Skipdog said:

Perhaps 7.3.1 as Jorge mentioned might have a regression for LSI 3xxx cards? I had to revert to 7.3.0 to fix my install as well.

Ref: https://forums.unraid.net/topic/199098-unraid-730-731-array-startperformance-related-issues/

Skip

Hmm... I'd rather not have to buy a compatible PCIe card to get around this issue, but because I just bought a complete 36-bay SuperMicro server as an eventual replacement for my Chenbro setup (this chassis barely fits my XRack Pro cabinet, can't find any slide rails for it, the drive caddy release buttons are all breaking off due to heat and age which Chenbro will NOT sell replacements direct to customers [which the chassis has now been discontinued], and it's just plain loud even inside the noise reducing enclosure, and I really thought the MB SAS ports may have entirely failed when I discovered the drives no longer showing up after the UnRAID upgrade), it came with an LSI 9361-8i MegaRAID card so I need to replace it, which will probably be the BroadCom 9400-16i.

SO, if this issue doesn't get fixed any time soon (if ever), I may just go ahead and get the 9400-16i, which uses the SAS3416 chip; I hope this is not affected by the 7.3.x "upgrade."

56 minutes ago, Auggie said:

Hmm... I'd rather not have to buy a compatible PCIe card to get around this issue

LSI 9300-8i (SAS3008) is working fine for other users, problem is the 9305-16i and -24i, so yours is a different issue than the ones mentioned on this thread.

Jun 2 06:30:49 UnRAID kernel: mpt3sas_cm0: port enable: FAILED with timeout (timeout=300s)

The HBA is failing to initialize; try a different PCI slot or look for a BIOS update.

21 hours ago, JorgeB said:

Jun 2 06:30:49 UnRAID kernel: mpt3sas_cm0: port enable: FAILED with timeout (timeout=300s)

The HBA is failing to initialize; try a different PCI slot or look for a BIOS update.

The LSI3008 chip is embedded into the MB so can't switch slots. I'm loathe to update BIOS AGAIN on this SuperMicro board since upgrading it from its original BIOS introduced PAID or SUBSCRIPTION requirements to access features that I previously could access for free.

But the question I should really ask: Did UnRAID make a change that requires updating the BIOS of a MB? Again, I didn't see any mention in the upgrade logs that would hint a BIOS update may be necessary.

Or are did you mean do a FW update to the LSI3008 chip, which I haven't updated since I believe 2018 when I first got this MB and flashed it to IT mode FW.

10 minutes ago, Auggie said:

But the question I should really ask: Did UnRAID make a change that requires updating the BIOS of a MB? Again, I didn't see any mention in the upgrade logs that would hint a BIOS update may be necessary.

A kernel change can sometimes affect things like that, where in some boards or hardware it worked with the previous kernel, and it doesn't with the current one. If that's the case, a future kernel update may also resolve the issue.

I will mull over a BIOS update. Not really inclined to do one given how SuperMicro clearly went for money-grab when they arbitrarily imposed a fee in a BIOS update to access previously free features. Then again, the damage is already done unless they've restricted even more features for cash.

I'm performing an overdue drive replacement but will probably do the BIOS update next week as I don't want to be left permanently stuck at UnRAID 7.2.7.

Now, if the BIOS update does not resolve the issue, what further steps can be done? I think a bug report would be in order at a minimum...

You can create a Linux kernel bug report.

FYI, instead of updating BIOS, I went ahead and splurged on a new Broadcom 9400-16i HBA, upgraded UnRAID to v7.3.1 and all is well. I have to eventually use this board anyways if and when I switch out to a new SuperMicro 36-bay server waiting patiently in a ginormous box in storage. I may look at virtualizing my pair of UnRAID iron and consolidate into a single machine while swapping out the too-deep Chenbro 48-bay to a shorter WD Data60.

Someday I'll try the BIOS and revert back to the embedded LSI 3008 and report my findings, but for now, I'm marching forward with new HBA.