6

This is a question related to this one:

Third-party SSD solutions in ProLiant Gen8 servers

but not covered by the question or the answers.

I have an OCZ 120 gig SSD as system drive on a Proliant ML310e Gen8 server, sitting on SATA port 6. In order to use that port (on the motherboard) for the SSD, I must set the SATA controller to SMART Array controller. This gives me all 4 bays for a RAID set of HDD drives.

So far, so good. I've also installed Win Server 2008 R2 on the system drive and all works well.

BUT - when I chack the SMART Array, it say "predicted failure of drive 0 (SSD)". I have checked and verified the actual SMART settings for the SSD, and the drive is 100% OK. It's brand new, and the SMART settings have been verified as 100% OK by OCZ support.

I cannot clear this error on the HP diagnostic side, and at OCZ's suggestion I have been trying to figure out how to turn off SMART diagnostics on the HP, to no avail.

SO - how can I either clear the HP SMART Array (incorrect) diagnostic reading of the OCZ SSD SMART data, or else how can I turn off the HP reading the SMART stats off the SSD.

I realize this is entirely my own fault for not wanting to spent $1000+ on an HP branded SSD, but the OCZ one is working perfectly fine. It's just the OCZ version of SMART is not the same as HP's idea of SMART for SSD drive stats, and I'm really tired of the yellow blinking "system degraded" front panel light.

And yes, I've already been told "put black electrical tape over it". :-)

But I'd really like to find a better solution. This SMART stuff was supposed to be quasi-universal.

Huntrods
  • 227
  • 2
  • 10

2 Answers2

5

But the OCZ SSD isn't working fine... It's probably not reporting its temperature properly to the controller.

Your HP ProLiant ML310e Gen8 server has an embedded "Dynamic" Smart Array B120i controller on the motherboard. This is a "softraid" device. Is this the only disk in the server? It's not clear from your question. If not, where are the other disks connected?

There are so many limitations with this setup, that I'm not sure what to tell you.

  • This is a software-assisted RAID controller.
  • Only 2 of its 6 ports are 6Gbps. That's why I'm asking what else is inside the server.
  • It only supports SATA devices.
  • Many situations require disabling RAID functionality.
  • No cache.

Your specific issue may just be the OCZ drive (they're bankrupt anyway). Try with another SSD if you can.

Edit:

I would skip using an non-RAID SSD in this server and just rely on your (4?) internal disk bays.

In the HP ProLiant world, you can create multiple logical drives on a physical array. So, take your group of disks, using the Array Configuration Utility to carve out a 120GB logical drive for your OS, then allocate the rest of the space for the data. The Smart Array B120i supports 8 logical drives, so that's the smoothest way to handle this... unless you need an SSD. But it doesn't seem to be necessary for this server.

ewwhite
  • 194,921
  • 91
  • 434
  • 799
  • The consumer OCZ drives do not have a temperature sensor, so the temp reported is 30/30 at all times. – Huntrods Dec 18 '13 at 19:11
  • This is the only SSD in the system. It's in SATA slot 6, which is not part of the 4-slot Smart RAID array. Like slot 5, which is used for the optical drive, slot 6 is designed for a single system disk. That's what I used it for. – Huntrods Dec 18 '13 at 19:12
  • Slot 6 was not designed for that. But your options are to try a different SSD or just create logical drives on your regular array. No need for an SSD here. – ewwhite Dec 18 '13 at 19:47
  • @ewwhite: can you please check [my question](https://serverfault.com/q/855029/93133) ? – el.severo Jun 10 '17 at 14:28
1

I'm running an ML310e Gen8 v2 with SATA SSDs on the built in b120i port 5/6 (no arrays/jbod - just using controller to make port 6 bootable) and have found a solution to stop the SSD drives showing as failed on reboot! This is homelab running linux/zfs, but I would certainly use in production.

  1. Open Smart Storage Administrator (SSA) - I did this on POST using F5 once controller was detected. You can also do through Intelligent Provisioning (F10).

  2. Go to Modify Controller Settings and disable "Surface Scan Analysis Priority". This will stop the controller from trying to surface scan your SSDs which stops them being marked as failed.

  3. Enjoy 3rd party SATA SSDs without them being disabled as failed on boot.

The HP Smart Storage Administrator User Guide had the hint at the solution. It appears the surfacescanmode causes the issues with SSDs being marked failed. This manual shows other ways of accessing SSA - and I believe surfacescanmode can be turned off per slot using the ssacli tool (maybe via serial/iLo also?) if you really want to use this on other drives (however I didn't try this since I don't use any of the array controller functionality...use smart monitoring instead)

Per the SSA user guide, this setting should work with Gen9 also. Also with Gen6, Gen7 (access SSA through download image) - I suspect Gen5 would work also since it supports ssacli.

deeess
  • 51
  • 1