INT. SATA drives keep disappearing

[Tsunamijhoe]

*UPDATE*

OK, something is going on,and i do not know what. I did a flexraid verify, it got to 67% and did an invalid EOF state error, which is the same error when drives disappear, but this time,all the drive letters were in "my computer".

However, pc wouldnt shut down,a sure sign one of the drives couldnt be accessed,had to hard reset,and it wouldnt reboot, hung on classpnp.sys on boot up,another hard reset,and it booted fine.

So,somehow these drives gets dismounted/locked/loose power/whatever at random intervals of time,and cant be accessed unless pc gets shutdown completely,wait 2 minutes boot up again,then they are seemingly fine again.

I am soon at the point, where i take this as a loss,and throws everything the f**k out,and start over once i get enough money.

What i dont get is, even though i have tried changing cables,drives,controllers,mix and match drive slots, it still happens,so it must be an error that occur on specific parameters only,which i cannot manually reproduce,and such an error is impossible to fix.

I have one last stand i can try,and that is to replace the drives that keep doing this for a completely different type,ALL my drives are WD greens, and i know 1 drive is bad (not one of the drives that keeps dismounting) and will be replaced by WD 3TB red's,but i cant afford this until 2 months from now,so what was your last idea?

 

[Lady Fitzgerald]

First, a question. Were the drives that dropped off being fed by the cable that is also feeding both of the case fans? If so, try moving one fan to another cable.

Since it is the same drives that keep dropping out, try running a diagnostic program on them. WD has a free one for its own drives here (scroll down until you see the download box); I use it to check all my WD HDDs. Another one with a free version that I see recommended a lot is HD Tune. I've used it only once but it seems to work well. It might not hurt to use both on the odd chance WD's utility is a wee bit biased. Testing the drives takes a long time. My 2TB WD Greens take around 20 hours each in a USB 2.0 dock (being the anal old coward I am, I always test new drives).

If the drives pass inspection, try swapping out the troublesome drives with other ones in your rig that are behaving. The purpose of this is to see if the cables might be the problem. If the troublesome drives behave in the new location and the ones replacing the troublesome ones start misbehaving, we will have narrowed the problem down to somewhere in the cables.

The WD Reds have been getting worse reliability reviews than the Greens. The Blacks get good reviews but are power hungry (not to mention a wee bit expensive).

 

[Tsunamijhoe]

First, a question. Were the drives that dropped off being fed by the cable that is also feeding both of the case fans? If so, try moving one fan to another cable.

Since it is the same drives that keep dropping out, try running a diagnostic program on them. WD has a free one for its own drives here (scroll down until you see the download box); I use it to check all my WD HDDs. Another one with a free version that I see recommended a lot is HD Tune. I've used it only once but it seems to work well. It might not hurt to use both on the odd chance WD's utility is a wee bit biased. Testing the drives takes a long time. My 2TB WD Greens take around 20 hours each in a USB 2.0 dock (being the anal old coward I am, I always test new drives).

If the drives pass inspection, try swapping out the troublesome drives with other ones in your rig that are behaving. The purpose of this is to see if the cables might be the problem. If the troublesome drives behave in the new location and the ones replacing the troublesome ones start misbehaving, we will have narrowed the problem down to somewhere in the cables.

The WD Reds have been getting worse reliability reviews than the Greens. The Blacks get good reviews but are power hungry (not to mention a wee bit expensive).

I already did all the above suggestions. In fact,i bought HD Sentinel specfically for that purpose,as that is the only program that can fetch smart info through the ibm controller.

WD's own program sucks,and cannot see the drives on the ibm controller,crystaldisk info worked on the old controller,and all drives are fine,except 1 which i am going to replace,but this drive is NOT one of the ones that keep disappearing..

Also, i have tried SO many times changing cables, changing connector position, changing drive bay nothing have worked.

Also, no the drives that are dropping off was NOT on the power cable shared with the case fans.

I checked the windows log,and it seems to have ALOT of atapi errors,140 in the last 24hrs,specifically on ide port 8,11,17 and 19,and also 40 errors from my sas2 controller in the last 24hrs. Now the errors stopped around 3pm this afternoon,maybe due to me changing the cables around yet again,but i am afraid this is probably because 1. windows drive spin down has been disabled yet again, and 2. i have stopped ALL form of automatic s.m.a.r.t. reporting.

Clearly i have conflict somewhere,and i am betting my left nut it is related to the green drives power save feature and/or windows power save,but ofcourse i am not sure,as everytime i think i have seen the light,well the light sorta goes out again,if u catch my drift , but then again,why would it only hit 1 of 8 WD20EARX drives,and 1 of 6 WD20EARS Drives,see my predicament? Nothing logically makes sense

Anyways, if the system runs stable with power save/smart features turned completely off, i can manage until i can afford to exchange the drives, if problem persists,i give up,and will disassemble the server,and sell off the hardware,and buy new stuff when i can afford it.

*UPDATE*

System did NOT Run stable, drive D disappeared now,i had to pull it out of drive cage,wait 20 secs,put it back in,and drive was back..

 

[Lady Fitzgerald]

Ok, I did some more digging. Something started rattling around in the back of the empty space where my brain should be about greens not being suited for servers. I checked the 1 egg reviews for the 2TB green drive and there were a couple of vague mentions about the greens not being recommended for NAS. I saw a NewEgg video about the reds saying that they were designed for use in servers so, sadly, that may be your answer (although I'm getting a little queasy thinking about how much replacing twenty 3TB greens with reds is going to cost you; btw, are you aware that WD has an upgrade program?).

The fact that, even though your machine was locking up on you, all your HDDs were still showing up in My Computer suggests that, although not the total solution, voltage drop from overloading a single feed was a factor contributing to your problems. Beyond that, I'm stumped.

 

[Tsunamijhoe]

Ok, I did some more digging. Something started rattling around in the back of the empty space where my brain should be about greens not being suited for servers. I checked the 1 egg reviews for the 2TB green drive and there were a couple of vague mentions about the greens not being recommended for NAS. I saw a NewEgg video about the reds saying that they were designed for use in servers so, sadly, that may be your answer (although I'm getting a little queasy thinking about how much replacing twenty 3TB greens with reds is going to cost you; btw, are you aware that WD has an upgrade program?).

The fact that, even though your machine was locking up on you, all your HDDs were still showing up in My Computer suggests that, although not the total solution, voltage drop from overloading a single feed was a factor contributing to your problems. Beyond that, I'm stumped.

Yeah, well, i was already in the process of replacing all my WD20EARS drives with WD20EARX drives,as they were reported to have the idle timer fix,but there are 2 types of WD20EARX the old type has the same idle problem as the EARS drives,and as far as i can gather, it seems the only drives that crap out is EARS and the old EARX drive.

Now,i was about to purchase 3x3TB reds to replace my parity drive,a faulty EARS and a suspicious EARX. I knew the greens were not suited for server use,but this isnt exactly a server,it is a windows 7 with A LOT of drives ..

But,maybe the combination of consumer mainboards and drives paired with server controllers is a bad idea. What i dont get is,a lot of people have no problems whatsoever,but i suspect they might be using server mainboards also,not some cheap asrock like me..

I am gonna have to live with this until i can afford to exchange my drives and look into maybe replacing my mainboard for a server rated one that is built for this stuff.

Since i havent sold my old supermicro controller, do you think my problems would go away if i bought this Supermicro | Products | Motherboards | Xeon® Boards | X9SCL-F ? This mb will accept all my current hardware except my ram,but ram is cheap,my core i3 and will accomodate both my ibm serveraid card AND the supermicro, and it would be cheaper to buy this mb instead of a cheapo mainboard and another ibm controller. Also, the drive dropout problems on the supermicro would probably go awya using a mb that is built for their sas controllers..

I could then shelf my 4 asrock sata controllers,and i think that might be a good thing..

 

[Lady Fitzgerald]

Supermicro | Products | Motherboards | Xeon® Boards | X9SCL-F[/URL] ? This mb will accept all my current hardware except my ram,but ram is cheap,my core i3 and will accomodate both my ibm serveraid card AND the supermicro, and it would be cheaper to buy this mb instead of a cheapo mainboard and another ibm controller. Also, the drive dropout problems on the supermicro would probably go awya using a mb that is built for their sas controllers..

I could then shelf my 4 asrock sata controllers,and i think that might be a good thing..

That I can't really help you with. Hopefully, someone who knows more than I do (which is most people here) will chime in.

 

[Tsunamijhoe]

Supermicro | Products | Motherboards | Xeon® Boards | X9SCL-F[/URL] ? This mb will accept all my current hardware except my ram,but ram is cheap,my core i3 and will accomodate both my ibm serveraid card AND the supermicro, and it would be cheaper to buy this mb instead of a cheapo mainboard and another ibm controller. Also, the drive dropout problems on the supermicro would probably go awya using a mb that is built for their sas controllers..

I could then shelf my 4 asrock sata controllers,and i think that might be a good thing..

That I can't really help you with. Hopefully, someone who knows more than I do (which is most people here) will chime in.

Hmm,ok..The lets let this thread die,as we are no closer to a solution,and i wont spend anymore time on it,i am gonna replace the mainboard for a dedicated server one,which i should have bought to begin with anyhow,i just didnt think it would make that much of a difference to be honest,but reading the unraid forums,noone there uses "generic" mainboards for this many drives,so it must be a compatability issue of some kind or limitations in hardware that causes the drive dropouts.

Anyways, thank you very much for taking the time to try and help me,i really appreciate it.