root?I rebooted the computer and now I get -
sg_scan: Error opening /dev/sg0 : Permission denied
sg_scan: Error opening /dev/sg1 : Permission denied
/dev/sg2: scsi3 channel=0 id=0 lun=0 [em]
sg_scan: Error opening /dev/sg3 : Permission denied
I added a little more information on your TrueNAS post to clarify what the other poster was trying to tell you. Long story short there's no such thing as "SMART" or "vendor specific attributes" defined in the SCSI (SAS) standards. It won't help you to try other tools to get more data than smartctl is showing you, that data does not exist for SAS drives.According to replies on my post on Truenas forums the lack of the vendor specific attributes is common on sas drives.
Very useful insights, thanksI added a little more information on your TrueNAS post to clarify what the other poster was trying to tell you. Long story short there's no such thing as "SMART" or "vendor specific attributes" defined in the SCSI (SAS) standards. It won't help you to try other tools to get more data than smartctl is showing you, that data does not exist for SAS drives.
If Hitachi is storing this information on their SAS drives (and why wouldn't they?), then I suspect it's in what's called a "log page" in the SCSI standard. The problem is you don't know which log page (there are LOTS) and even if you knew which page, you don't know how the information is encoded on the page. You could reach out to Hitachi and ask them, and they could help you figure it out so you can query the page with smartctl and decode it manually. Alternatively, you can use smartctl to query log page 00h which should give you a list of all log pages supported by your drive. The vendor specific log pages are typically 30h-3Eh, so if the drive reports a page in that range it might have the data you seek and you could query those pages with smartctl. The problem is still that you wouldn't know how the data is encoded, but if you extract the data from those pages maybe something will stand out? More likely than not you're going to have to talk to Hitachi to figure this out. I create SMART tools and I often end up talking to the manufacturers, occasionally under NDA/PIA.
Tools like smartctl are great since they hide all this protocol messiness behind the scenes and try to give you an easy universal picture of how your drive is doing. At the same time they can be limited in what they understand outside the run of the mill attributes from the biggest manufacturers, since vendor specific information is by definition... vendor specific. Hope that helps!
Someone got a lot of them in one of the servers the seller is selling. He said the drives looked to be powered on for only a couple of days.Has anyone grabbed drives from the newest lot? It looks like they're sealed recertified drives now, so curious what that SMART data has to say.
So what were these used for? voice recording as in when you call into AT&T to bitch about their bad services, they recorded everything to these drives? lol..Btw, for those wanting to know where they got 10s of 1000s of these drives...
The seller also has 100s of 4U Cisco servers that holds 56x LFF HDDs for sale at one hell of a price. (Don't immediately jump in and buy that chassis without learning about this machine in that thread we have there. We're still trying to figure things out and the latest news is that the Cisco Management firmware for the chassis only will support specific hardware models in the HUU.)
One person got one of those chassis and it wasn't reset. It had AT&T registration information in the BMC, and seemed to bare-metal voice-recording machines according to some other text they found in the non-default configuration of the machine.
So the answer seems to be, "AT&T had 100s of 1000s of those machines, that all had these 10 TB SAS3 drives in them as it was standard sizing." Hence, why they have well over 50,000 drives to sale.
I guess the problem is more often with linux?Just a small update on my HUH* SAS drive adventures, i received the drives, 2 x 8TB HUH728080AL5205 bought from a big UK enterprise HW seller, about 80 EUR per drive, but with import duties and shipping costs ended up at about 15 EUR/TB. Drives looked nice, with PCB covered in a tamper proof seal etc..
Attached them to my LSI 3008 controller, and the problems began immediately. Debian 11 live threw all kinds of block errors at boot. Trying to run a smartctl short self-test was the weirdest thing ever, for a second i could hear the drive doing something, then the % remaining of test went from 100% down to 5% in a few seconds, and stayed there, nothing happening. Unable to create a filesystem on the drives, nothing but the weirdest stuff. Most weird was both drives had the exact same errors.
So tried them in another system, different mobo, different LSI 2008 controller, different cables, exact same thing. Other drives work fine in both systems.
Also noticed these had the same seagate code based 'media scan' crap that scan the drives for hours and days upon every damn boot. that scan never progressed etc etc.
Luckily the seller was very responsive and helpful, i returned the drives, and now i'm DONE with SAS drives. They've been nothing but pain for me. My only reason for playing with them was before the crypto markets, second hand SAS drives could be found much cheaper if you had a controller.
Anything seagate is poison to me, and i never imagined the HGST drives would be running the same seagate media scan crap, what a nightmare.
Now i have ripped my SAS controller out of my server, and will only be looking at SATA helium drives. Much less pain. As the crypto markets destroyed the only advantage of second hand SAS drives which was the price, i see no reason to bother with them anymore.
Well it MIGHT be, but everything else works just fine. I mean if the price of the drives would have been 10€/TB then i might have bothered to investigate further, but 15€/TB is nothing special. Between the cost of the sas cables, the controller producing extra heat and taking up a pcie slot, the seagate media scan CRAP, the smartmontools weirdness discussed in this thread, it's just not worth it.I guess the problem is more often with linux?
I actually like the sas controller. The breakout cables are so much smaller than each sata cable. and you can even get a controller that can handle 8 sas drives for around $25-30. That is not something you can find for a sata controller card.Well it MIGHT be, but everything else works just fine. I mean if the price of the drives would have been 10€/TB then i might have bothered to investigate further, but 15€/TB is nothing special. Between the cost of the sas cables, the controller producing extra heat and taking up a pcie slot, the seagate media scan CRAP, the smartmontools weirdness discussed in this thread, it's just not worth it.
It felt good ripping that controller out and be done with it. Now it's just a small pretty SATA cable between the drive and the mobo, the birds are singing again, and the sun just emerged from behind the clouds, gently warming my face.