I just went to replace one of my 2TB disks with a 3TB and encountered a problem. When I tried to restart the array I got a message saying that the parity disk is no longer the largest disk, and the start array button was greyed out.

 

My parity disk is a WD30EZRX and the new disk I was adding is also a WD30EZRX, all other disks are WD20's. At the moment I have put the original 2TB back in and it is currently doing a data rebuild on that disk.

 

What do I do?? Is is complaining over a few bytes difference in size?? Should I add the new disk as the parity disk and use the parity to upgrade the 2TB?

 

Running version 5.0 rc 10.

 

Thanks.

I just went to replace one of my 2TB disks with a 3TB and encountered a problem. When I tried to restart the array I got a message saying that the parity disk is no longer the largest disk, and the start array button was greyed out.

 

My parity disk is a WD30EZRX and the new disk I was adding is also a WD30EZRX, all other disks are WD20's. At the moment I have put the original 2TB back in and it is currently doing a data rebuild on that disk.

 

What do I do?? Is is complaining over a few bytes difference in size?? Should I add the new disk as the parity disk and use the parity to upgrade the 2TB?

 

Running version 5.0 rc 10.

 

Thanks.

post a syslog.  You may have an HPA.  (look it up  in the wiki)  If not, you can do as you said, add the new disk as the parity disk, and then use the old parity disk to upgrade the 2TB AFTER waiting a few days and after performing a parity CHECK to make sure the initial parity calc can be read.

Syslog is attached (syslog is pretty big since the server has been up for weeks, the serial for the new drive is WCC1T0620677 - that shoiuld make it easier to find in the syslog)

 

The thing is I just precleared this disk in the same machine I have used to preclear all my disks in the past, and neither my server nor the preclear machine have gigabyte boards. Why whould a HPA suddenly be being created?

 

The original 2TB disk is back in for now and has been rebuilt, the server is up and running fine. I will wait for advice before doing anything with the new disk again.

syslog-20130509-122209.zip

Your parity drive is reporting HPA

Apr 13 20:11:27 NAS kernel: ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300)

Apr 13 20:11:27 NAS kernel: ata4.00: HPA detected: current 5860531055, native 5860533168

Apr 13 20:11:27 NAS kernel: ata4.00: ATA-9: WDC WD30EZRX-00DC0B0, 80.00A80, max UDMA/133

Apr 13 20:11:27 NAS kernel: ata4.00: 5860531055 sectors, multi 0: LBA48 NCQ (depth 31/32), AA

Apr 13 20:11:27 NAS kernel: ata4.00: configured for UDMA/133

Apr 13 20:11:27 NAS kernel: scsi 3:0:0:0: Direct-Access    ATA      WDC WD30EZRX-00D 80.0 PQ: 0 ANSI: 5

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: [sde] 5860531055 512-byte logical blocks: (3.00 TB/2.72 TiB)

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: Attached scsi generic sg4 type 0

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: [sde] 4096-byte physical blocks

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: [sde] Write Protect is off

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: [sde] Mode Sense: 00 3a 00 00

Apr 13 20:11:27 NAS kernel: sd 3:0:0:0: [sde] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

Apr 13 20:11:27 NAS kernel:  sde: sde1

 

I just put the disk in another computer and ran HDAT2 and it reports that there is no HPA, the entire disk is available.

 

Do I need to put the disk in the server and run HDAT2 from there?? Or do I need to try hdparm -N instead?

It's not clear HOW you got the HPA ... but it's clearly there.

 

You have a couple of options:

 

(a)  Simply use the new disk (without the HPA) as your parity drive, and use the other drive in the array (it will work fine -- you'll just lose the few bytes in the HPA)

 

(b)  Remove the HPA with HDat2 ... just use MaxSize to adjust the drive to its full capacity

(5860533168)

 

Alright I powered down my server, installed the HPA disk into the server and booted HDAT2 from the usb stick.

 

This time HDAT2 did show a HPA on that disk so I wiped the HPA. Once that was done I rebooted into Unraid, Unraid has decided it wants to do a parity check so I am letting that run now, I will install the new disk once the parity check has finished.

 

The most interesting thing about this sequence of events is that HDat2 originally said there was no HPA.    I've found HDat2 very reliable -- I have to wonder just what happened (although we'll never know).    Are you CERTAIN that you were "looking" at the correct disk when you checked it earlier?  [i.e. were there any other 3TB drives on that same system that you could have inadvertently selected?]

 

Absolutely certain, It was pluuged into my desktop which only has a Kingston 64GB SSD and a 500GB Western Digital, I was definitely looking at a 3TB disk and it was the only one in the system.

 

The thing that has me more puzzled is how the HPA came to be there in the first place. Like I said the machine I used to clear the disk is the one I have to clear all my disks in the past, and none of them have HPAs created.

Agree it's a very strange set of events -- both how an HPA came to be; and why HDat2 didn't show it when you first looked !!

 

But I suppose "all's well that ends well" is the bottom line 

was the disk that had HPA new when you got it? if no, then the prev owner might have put the HPA on it.

 

do you have a gigabyte motherboard? if yes, you need to go into the BIOS and disable anything that says SAVE BIOS TO DISK, or something similar. then for good measure ensure your cache drive is hooked to PORT0 so if the HPA gets written again it will likely go to the cache drive which won't affect anything.

Yep disk was bought brand new a couple of days ago, opened the anti static bag myself.

 

Neither the preclear machine or the server have Gigabyte boards (the preclear machine is a 10 yr old thing with a single core P4 and an Albatron motherboard, and the server is a Supermicro motherboard.

As I noted above, "... all's well that ends well ..." 

 

I had the same thought initially r.e. a Gigabyte motherboard, but noted you had already excluded that (in post #2).

 

Bottom line:  We'll never know what happened -- but it's nice to know it's now resolved !!

 

 

Yep disk was bought brand new a couple of days ago, opened the anti static bag myself.

 

Neither the preclear machine or the server have Gigabyte boards (the preclear machine is a 10 yr old thing with a single core P4 and an Albatron motherboard, and the server is a Supermicro motherboard.

Just beware.  What you order as new may not be.  What I'm about to say doesn't happen very often to me but has on multiple occasions - even from newegg.  I got a MB that I bought as new that was packed in it's anti static bag.  When I opened it to look at the board there was heat sink paste stuck to the edges of the load plate plastic cover.  If the original user had been more diligent in cleaning up after his install I never would have seen that I got a USED MB instead of a new one. 

I had the same thought r.e. "new".  But about the only way to really know is to always do a S.M.A.R.T. test as the FIRST thing you do with a new drive -- to confirm the "power on hours" is a small number (it won't be 0, since even new drives are tested).

 

... but even that wouldn't catch a drive that someone had bought, installed (in a Gigabyte board), and then for some reason decided not to use [e.g. perhaps discovered they had a board that didn't support > 2TB dirives].

 

But it's useless to speculate ... and frankly irrelevant.    The drive pre-cleared okay, so it's fine, and it's now working as desired.    Just write the events of this thread off as one of life's little mysteries 

Well some news.

 

Unraid wanted to do a data rebuild on the original 2TB after I attempted to replace it with the 3TB and failed. So I let that run and it completed successfully. I thought it was wise to do a parity check before attempting to upgrade with the now HPA free 3TB, so that has been running while I was at work. I came home to find it had basically stopped at 56% and was throwing out a whole bunch of handle stripe read errors.

 

So it looks like the new 3TB (that had the HPA) is now going to replace my existing 3TB parity disk, which I will now have to preclear and pull the smart reports to see if it needs to be RMA'd.

 

So much for a simple disk upgrade, but atleast all my data is intact and I have a spare 3TB here to replace the parity with.

 

BTW - regarding the HPA drive, I did check the power on hours from the smart reports from the preclear and it has 36 hrs which is the same time the preclear took, and the manufacturing date was only 6 weeks ago, so I doubt it was returned. I have no idea how the HPA happened but as garycase said it doesn't matter its fixed now.

 

Thanks for the update.    Always nice to have a bit of status r.e. ongoing issues.

 

...so I doubt it was returned.

Sometimes I wonder if that crap only happens to me :'(

Sometimes I wonder if that crap only happens to me :'(

 

Fear not, you have company 

... I've had more than one instance of receiving a clearly-used item when I bought it NEW !

 

More news.

 

I installed the new (ex HPA) disk in the server to replace the parity disk that decided to have a tantrum. Parity has just finished rebuilding onto that disk and the array is up and running as normal.

 

I started a preclear on the old parity disk and everything was going well, when I woke up this morning I found it has slowed to a crawl and hadn't even finished step 1 yet. So I pulled the disk and connected it to my desktop intending to run WD Lifeguard and see what that said, as soon as I connected the disk to the desktop I got a BSOD. I rebooted but the desktop got caught in an endless rebooting loop (it would start booting but then something would cause problems and it would automatically reboot.) I removed the disk and the desktop booted fine.

 

I booted my preclear machine into XP and connected the drive to it intending again to run WD Lifeguard, it actually detected the disk fine and didn't BSOD on me, but WD Lifeguard is showing some rediculous number of hours to complete what is usually a 6 hour scan.

 

So it looks like the drive has died and will have to be RMA'd (it's actually an RMA drive itself). Which is actually pretty typical of this server, it will run rock solid for months, but when one thing goes wrong it sets off a whole cascade of problem and usually ends up in me having to RMA discs. Which is why I now keep a couple of pre-cleared spares in the cupboard.

 

I think I'll go and get another 3TB this week and install it and then when the replacement drive comes in a few weeks swap them over and keep the new one as one of my spares.

 

Thanks everyone for the help.

You may not want to spend the extra $$, but personally I'd recommend you stop buying the EZRX drives and buy the WD Reds (WD30EFRX) => longer warranty; run cooler; use less power; and (in my experience) definitely more reliable than the Green drives.

 

Pricing obviously varies, but at this moment they're $11 more than the EZRX's at Newegg

($150.99 vs. $139.99)

 

Don't add a lot more storage than you need, however => WD is slated to release a 5TB Red series drive later this year    [supposedly there will be a 4TB unit in October;  a 5TB unit by the end of the year]

 

I did consider the Reds before I bought the green but the price difference stopped me (I'm not in the US so I can't buy from Newegg.)

 

At the moment the prices locally are $129 for the green vs $156 (on special) for the reds.

I did consider the Reds before I bought the green but the price difference stopped me (I'm not in the US so I can't buy from Newegg.)

 

At the moment the prices locally are $129 for the green vs $156 (on special) for the reds.

 

21% more does cause a bit of thought ... although with the VERY good results I've had with the Reds I've bought, I'd still go with Reds.  Better performance; 25% less power (and heat);  50% more warranty; etc.    In addition, I've had ZERO defective drives of the 15 or so I've bought (i.e. they passed my very-thorough initial testing); whereas I've had quite a few green drives that failed my initial testing (about 1 in 5 or so).    I will say that both models have been VERY reliable once they passed the initial tests.