HP/ 3PAR Sandisk optimus "mainstream" sas ssd 920GB

[Biren78]

920GB/ $550 SAS2 SSDs, I think these are likely Optimus Ascends made special for 3PAR and HP SanDisk Optimus 920GB Enterprise SSD SAS 2 5" TXA2D2 | eBay

There isn't a ton of information but SanDisk says read and write operations: http://www.sandisk.com/assets/docs/how-to-order-hp-enterprise-sas-solid-state-drives-ssds.pdf so these aren't the cheap read only drives.

 

[OBasel]

BO $400 ea for 2.

 

[MiniKnight]

That's a rockn' deal actually. Under $0.50/GB and they are real drives. Any idea if they work in non 3PAR?

 

[Patriot]

Part of me always wonders just how hot the drives are...

 

[Patrick]

Part of me always wonders just how hot the drives are...

Yea, I was just thinking, these are really widely used SSD at major OEMs. I know there are lots of HP and IBM ones out there. Even if you figure 40-60% discount off of list, $400/ ea is wildly good. Very good drives that are really mainstream for enterprise.

$400 makes these are $0.43/GB which is less than a Crucial MX200.

Any idea if the 920GB is part of 3PAR's adaptive capacity? If these work then great deal.

 

[Patrick]

Two arrived today. Smart Storage branding. Will try to get them in a test rig in the next day or so.

 

[Patrick]

I/O device error on first attempt to initialize :-/

853.05GB on Windows Disk Management

 

[PigLover]

I/O device error on first attempt to initialize :-/

Did you ever get past this?

 

[azev]

I had some experience buying netapp relabeled hitachi spinners; those drive came with 520Byte sector, and I had to use some linux tools to reformat the sector to 512Byte before I can use it in a whitebox. Not sure if you could do something similar to this on an SSD.

 

[Mr. F]

I/O device error on first attempt to initialize :-/

853.05GB on Windows Disk Management

I had this happen when I pulled a pair of M550 1TB out of a mirror array on an HP P420 and connected them to a motherboard port. After reinitializing with Windows disk management and power cycling the drive I haven't seen any further issue.

 

[Patrick]

Did you ever get past this?

Not yet... grrr!

I had some experience buying netapp relabeled hitachi spinners; those drive came with 520Byte sector, and I had to use some linux tools to reformat the sector to 512Byte before I can use it in a whitebox. Not sure if you could do something similar to this on an SSD.

@azev - would you mind sharing a bit about how this was done?

I had this happen when I pulled a pair of M550 1TB out of a mirror array on an HP P420 and connected them to a motherboard port. After reinitializing with Windows disk management and power cycling the drive I haven't seen any further issue.

It will not initialize using Windows Disk Management. :-/

 

[azev]

@@azev - would you mind sharing a bit about how this was done?

Sure, here's the info I have in my documents:

Install RHEL6/CetOS6

If you look in /var/log/messages, you will see useful information:

Mar 9 08:08:54 vhc-carthage kernel: sd 6:0:7:0: [sdh] Attached SCSI disk
Mar 9 08:08:57 vhc-carthage kernel: ...ready
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] Unsupported sector size 520.
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] 0 512-byte logical blocks: (0 B/0 B)
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] 520-byte physical blocks
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] Write Protect is off
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] Write cache: disabled, read cache: enabled, supports DPO and FUA
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] Unsupported sector size 520.
Mar 9 08:08:57 vhc-carthage kernel: sd 6:0:8:0: [sdi] Attached SCSI disk

# yum install sg3_utils


# sg_scan -i
/dev/sg8: scsi6 channel=0 id=7 lun=0
NETAPP X287_S15K5288A15 NA00 [rmb=0 cmdq=1 pqual=0 pdev=0x0]
/dev/sg9: scsi6 channel=0 id=8 lun=0
NETAPP X287_S15K5288A15 NA00 [rmb=0 cmdq=1 pqual=0 pdev=0x0]


Now you should format the offending drive using the "sg_format" command.
[root@azev /]# sg_format --format --size=512 /dev/sg8

NETAPP X287_S15K5288A15 NA00 peripheral_type: disk [0x0]
Mode Sense (block descriptor) data, prior to changes:
Number of blocks=573653847 [0x22314357]
Block size=520 [0x208]

A FORMAT will commence in 10 seconds
ALL data on /dev/sg8 will be DESTROYED
Press control-C to abort

The above commands let me re-format my drive with drive sector of 512 and i was able to use it using LSI-9211 and LSI-9285 controllers.

Good Luck, wish this would work for you.

 

[Patrick]

Ok update ---- and it is surely a strange one.

I did the steps above on both drives in CentOS 6.5. Both were not able to be formatted. I went back to Windows Server 2012 R2. Neither formatted but one was giving a "device is write protected" error. I took that drive offline. Then online. Then initialized it and it worked!!!!

Here is the strange part: the second 920GB drive, which had the same commands run on it, did not work and still has the I/O device error. I could not even offline the drive in Windows Server 2012 R2.

I am now going to try this on the 100GB SLC Hitachi drive also but progress made!

 

[azev]

Interesting, did you see it in the linux log that the drive sector is other than 512 ?? most of drives destined for enterprise raid system usually comes with custom sector size. I think i've seen drive with 528 drive sector in the past.
Anyway, you can also use MSM to identify the drive sector size.

 

[PigLover]

Patrick - if the drive has a partition table in a strange format then windows will puke on it...try just doing a full disk wipe with it on the Centos machine. 'dd if=/dev/zero of={raw disk device}'.

 

[Patrick]

Wow! OK so did a reboot, sg_format both drives. The Hitachi worked right away after that, the SanDisk required an offline->online->format

So they require the CentOS step, but do work. Not the fastest drives in the world but really good price.

 

[Patrick]

Looks like he has 480GB drives too: SanDisk Optimus 480GB Enterprise SSD SAS 2 5" TXA2D2 | eBay

 

[Patrick]

Patrick - if the drive has a partition table in a strange format then windows will puke on it...try just doing a full disk wipe with it on the Centos machine. 'dd if=/dev/zero of={raw disk device}'.

@Patriot and I tried this on the Hitachis before. Looks like the sg_format was the answer! @azev I think this deserves its own post.

 

[azev]

I am glad this work. I've only done these steps on sas spinners, but now we know this also works on ssd

 

[Patrick]

Here is a very interesting note. The 920GB is due to 3PAR's adaptive OP. Basically, 3PAR is managing OP in software. These drives are 920GB but very little OP. Somewhat cool since you can then dial up the OP to 800GB for 10DWPD (I think they are Ascends) or keep as is and get lower endurance but more capacity.