JorgeB

It has been used successfully before by other Unraid users.

The other devices were initially detected but then stopped responding and were dropped after several attemps, if I had to guess it looks like some compatibility problem between those Firecuda drives and the LSI (or the Dell and the LSI but that seems not so likely since the other model disk and SSD work).

The OP had a cable problem, likely not the same as you, please post the diagnostics: Tools -> Diagnostics

Cheapest would be something like this, using an existing ATX case: You can get an used Intel SAS2 expander for around $60.

Check that cache use is still set to "yes" for that share, you can also try toggling it.

It's on the flash drive, mounted at /boot so all are under: /boot/config/shares

Likely the same reason as most or all your shares exist with capital an no capital first letter, no idea how that happened though.

Is this for the photos share? Looks like it exists in the flash as photos and Photos, and only Photos has the correct settings, delete config\shares\photos.cfg and then rename Photos.cfg to photos.cfg.

There aren't very good 4 port controller options. Marvell controllers are not recommended for some time, they might not even work with newest Unraid v6.7, if 2 ports are enough you can get an Asmedia controller, for more than that an 8 port LSI HBA is the best option, they can be found used on ebay starting at around $40/50.

It should work, but not really recommended, for a variety of reasons, like bad performance, possibly no SMART, and it might no even work if the devices have same or similar identifiers, if you really need external use SAS enclosures, more expensive but very reliable.

Not that I recall.

There are additional reads on disk2, to see if speed is normal or not you'd need to stop all other array access.

No, not possible to rebuild and add drives at the same time, you can rebuild and and parity2 and when done add the remaining drives, or vice-versa. Grab diags during current check.

You can check the serial number with Broadcom to see if it's genuine, though likely only after buying it, I usually buy used server pulls, less chance of getting a fake.

Yes, and NVMe adapters should work with any board as long as used in an appropriate slot, they are transparent to the OS.

Unraid doesn't currently support Infiniband, most/all Mellanox Infinand NICs can be set to Ethernet mode, and Unraid does support the Connectx-X 10GbE.

@ezhikfound that downgrading LSI 2008/2308 firmware to p16 restores the trim function with current driver, so the trim issue is caused by the driver and the firmware and the filesystem, and while I personally wouldn't like to be running an older firmware it might be worth considering for users without a better option.

I would prefer using onboard SATA2 than downgraded LSI, though it can be a good option for those users that don't have any onboard SATA ports, like some Dell servers.

That's interesting, so trim working depends on the driver version and the firmware version and the filesystem used, still I wouldn't downgrade firmware because of trim, you might miss on important fixes, especially when used with the newer driver, just use the onboard SATA ports.

Try updating your bios, since you're not on the latest, it might help with the IRQ16 getting disabled, and if nothing else it should help with this: Feb 26 20:17:19 Tower kernel: L1TF CPU bug present and SMT on, data leak possible. See CVE-2018-3646 and https://www.kernel.org/doc/html/latest/admin-guide/l1tf.html for details.

No disks are being detect by Linux, check the board or HBA bios, depending where the disks are connected, they most likely won't be detected there either, and until they do none will appear on Unraid. If you're using the HBA make sure you're using the right cables, you need forward breakout cables, reverse breakout look the same but won't work.

It's a problem with both HP RAID controllers: Feb 25 23:04:33 TheNewdaleBeast kernel: hpsa 0000:81:00.0: Controller lockup detected: 0x00130000 after 30 Feb 25 23:04:33 TheNewdaleBeast kernel: hpsa 0000:81:00.0: controller lockup detected: LUN:0000000000800601 CDB:01030000000000000000000000000000 Feb 25 23:04:33 TheNewdaleBeast kernel: hpsa 0000:81:00.0: Controller lockup detected during reset wait Feb 25 23:06:03 TheNewdaleBeast kernel: hpsa 0000:03:00.0: Controller lockup detected: 0x00130001 after 30 Feb 25 23:06:03 TheNewdaleBeast kernel: hpsa 0000:03:00.0: controller lockup detected: LUN:0000000000800401 CDB:01030000000000000000000000000000 Feb 25 23:06:03 TheNewdaleBeast kernel: hpsa 0000:03:00.0: Controller lockup detected during reset wait Maybe a driver/kernel issue?

Yes it is, and it dropped offline before rebooting since there are millions of errors: Feb 25 16:34:41 Tower kernel: BTRFS info (device sdk1): bdev (null) errs: wr 3718232, rd 2992936, flush 94483, corrupt 0, gen 0 See here on how to better monitor your cache pool: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=700582

That's not correct, trim isn't working in any of the them, the only one that would be trimmed by the LSI HBA is the Intel since it's only one that supports read zeros after trim, hence the different error, but it also doesn't work with latest Linux drivers, see here for more info: https://forums.unraid.net/topic/74493-samsung-ssd-hba-trim/?do=findComment&comment=720627 https://forums.unraid.net/topic/74493-samsung-ssd-hba-trim/?do=findComment&comment=720637

This came up before, apparently that happens when the mover runs on a schedule: