array shows disks error that disappear on reboot

[mo679]

hello all,

i'm trying to figure out how to manage various issues on my unraid setup

Hp Ml350 gen 9, dell perc 310 in it mode, connected to lower half of the first backplane, upper half of the first backplane is connected to motherboard A sas port, B sas port on the motherboard is connected to parity drive and one other 3tb drive.

i removed the Hp smart array because it was non passing the smart and temperature of disks (though in HBA mode), the SAS host on the motherboard is in ahci mode. 

the array is working but i keep on getting disk errors on the same 4 disks connected to the upper portion of the backplane on sas port A.. these errors reset after reboot and then start slowly to build up again.

the disks are older 1tb sata that came with the machine. the errors are the same on the drives

I thought i was a bad cabling issue and switched the sas cable, but errors keep on coming. 

until now i did not get drives lockout, but i'm worried if that could happen.

Could I move drives around to see if it's a backplane issue? can I connect them to a sas to sata cable bypassing the backplane without messing up the array?

Log files shows a bunch of errors, maybe someone can help.

syslog.txt

[JorgeB]

Please post the diagnostics: Tools -> Diagnostics

[mo679]

thanks sir for taking a look, I changed cable again today, just in case the two I tried were faulty.

homeserver-diagnostics-20210412-2116.zip

[JorgeB]

Kind of a strange issue, you're getting errors in multiple disks, all connected to the onboard SATA ports, and they are reported as a disk problem, try replacing those cables and/or connect those disks to the LSI instead.

[mo679]

I replaced the cables yesterday, started parity check  and this morning I got errors again, just in these four disks and on the exact same sectors, the disks are fine and pass the smart test, 

can I connect the cable to the lSi adapter without messing up the array? In my synology I Can’t ove disks around.

by the way, i remember getting error also with the hp smart array mounted (I thought about an incompatibility with unraid as many other suggested. The disks at the time were located in the lower bays and therefore connected to B port on the integrated sas controller.

 

[JorgeB]

44 minutes ago, mo679 said:

can I connect the cable to the lSi adapter without messing up the array?

Yes.

 

Could be some compatibility issues with those disks, what are they, some white label?

[mo679]

well actually they are original branded Hp sata disks which came with the server, they were hooked to the smart array controller in raid 5. I precleared the disks after installing unraid, but, initially I was not getting smart and temp data. so I set the smart array controller into HBA mode. no luck. I removed the smart array and hooked the mini sas cables directly on the motherboard mini sas imput. drives showed up fine, temperature and smart data are passed trough. but these random errors keep on showing up. 

I'll try connecting the cable to LSI adapter, though I really needed to add 4 more drives to that free connector...

[mo679]

Ok switched cable for a brand new mini sas AND moved the cable to the LSI adapter. I checked today, but the error appeared again, same error on the four disks. Might it be, since the disks have some kind of hp firmware that makes them incompatible?

[JorgeB]

56 minutes ago, mo679 said:

since the disks have some kind of hp firmware that makes them incompatible?

It's possible.

[mo679]

Ok thanks for the reply, I’ll try to connect the drives using a mini sas to 4 sata cable and a proper sata power supply to bypass the backplane, do you think I’m going to break the array doing so? I think there is some sort of talking between the backplane and the drives that could generate the issue

[JorgeB]

Just now, mo679 said:

do you think I’m going to break the array doing so?

Nope.

[mo679]

hello again, 

I tried to move one disk around and connected it to a sas to sata cable so to bypass the backplane. same result. the disk produces errors which look identical on the other 3 branded Hp Disks. this, I think, rules out an hardware issue (cabling controller backplane) since I switched pretty much everything. the problem persists and It seems a software/firmware incompatibility issue.

other Ideas? is it possible to maybe flash the disk firmware to a generic one or to avoid Unraid misinterpreting these "errors".

homeserver-diagnostics-20210419-1045.zip

[JorgeB]

That suggests it's probably a disk issue with Linux, try different disks if that's an option.