HP P400 RAID controller reporting p-disks incorrectly
Harout S. Hedeshian
reporting the physical disks incorrectly? Whenever I do a show config
detail from the hpacucli (and hpaducli), it reports the three drives
as all being 0:0 with box 0 bay 0 and is reporting the model
information for the disk repeated. This is problematic as I can't
perform operations on the disks individually. The disks are SATA3G
(1.5G limited by controller). I am running the card on an Intel
S5000XSL motherboard in a supermicro chassis. I don't expect the card
to report bay/box given the non-hp backplane, but why all drives on
channel 0:0 with the first drive repeated? I can't figure it out. Of
course, I have the latest firmware and hpacucli.
After some googling, I found this forum thread on HP with no
resolution:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?admit=109447626+1245803945537+28353475&threadId=1296865
What's also inconsistent is the drive show NCQ capable and enabled by
the controller config shows false...
Thanks for the input,
Harout
HP Array Configuration Utility CLI 8.28-13.0
...
=> ctrl slot=3 show config detail
Smart Array P400 in Slot 3
Bus Interface: PCI
Slot: 3
Serial Number: PA5360BBFT4NKZ
Cache Serial Number: PA82C0CBFSY6UC
RAID 6 (ADG) Status: Enabled
Controller Status: OK
Chassis Slot:
Hardware Revision: Rev B
Firmware Version: 5.26
Rebuild Priority: High
Expand Priority: Medium
Surface Scan Delay: 15 secs
Post Prompt Timeout: 0 secs
Cache Board Present: True
Cache Status: OK
Accelerator Ratio: 25% Read / 75% Write
Drive Write Cache: Disabled
Total Cache Size: 256 MB
No-Battery Write Cache: Disabled
Cache Backup Power Source: Batteries
Battery/Capacitor Count: 1
Battery/Capacitor Status: OK
SATA NCQ Supported: False
Array: A
Interface Type: SATA
Unused Space: 0 MB
Status: OK
MultiDomain Status: OK
Logical Drive: 1
Size: 596.1 GB
Fault Tolerance: RAID 5
Heads: 255
Sectors Per Track: 32
Cylinders: 65535
Stripe Size: 256 KB
Status: OK
MultiDomain Status: OK
Array Accelerator: Enabled
Parity Initialization Status: Initialization Completed
Unique Identifier: 600508B1001042424654344E4B5A0008
Disk Name: /dev/cciss/c0d0
Mount Points: /volumes/c0d0p1 596.1 GB
Logical Drive Label: A021625DPA5360BBFT4NKZD3C7
physicaldrive 0:0
Box: 0
Bay: 0
Status: OK
Drive Type: Data Drive
Interface Type: SATA
Size: 320.0 GB
Firmware Revision: 3.AAE
Serial Number: 6QF1SA0A
Model: ATA ST3320620AS
SATA NCQ Capable: True
SATA NCQ Enabled: True
PHY Count: 1
PHY Transfer Rate: 1.5GBPS
physicaldrive 0:0
Box: 0
Bay: 0
Status: OK
Drive Type: Data Drive
Interface Type: SATA
Size: 320.0 GB
Firmware Revision: 3.AAE
Serial Number: 6QF1SA0A
Model: ATA ST3320620AS
SATA NCQ Capable: True
SATA NCQ Enabled: True
PHY Count: 1
PHY Transfer Rate: 1.5GBPS
physicaldrive 0:0
Box: 0
Bay: 0
Status: OK
Drive Type: Data Drive
Interface Type: SATA
Size: 320.0 GB
Firmware Revision: 3.AAE
Serial Number: 6QF1SA0A
Model: ATA ST3320620AS
SATA NCQ Capable: True
SATA NCQ Enabled: True
PHY Count: 1
PHY Transfer Rate: 1.5GBPS
Benjamin Gawert
> Has anyone using the P400 controller experienced the ACU CLI utility
> reporting the physical disks incorrectly? Whenever I do a show config
> detail from the hpacucli (and hpaducli), it reports the three drives
> as all being 0:0 with box 0 bay 0 and is reporting the model
> information for the disk repeated. This is problematic as I can't
> perform operations on the disks individually. The disks are SATA3G
> (1.5G limited by controller). I am running the card on an Intel
> S5000XSL motherboard in a supermicro chassis. I don't expect the card
> to report bay/box given the non-hp backplane, but why all drives on
> channel 0:0 with the first drive repeated?
Because the HP utilities are not able to resolve the drive location due
to the lack of a HP backplane. If you look at a HP backplane you'll
notice that there is more on it than just SATA/SAS connectors and power
connectors, it contains some logic which tells the controller the slot
numbers and the type of enclosure. Standard backplanes don't have this
logic, so the HP utilities report all disks as 0:0.
I have a ProLiant ML350G5 with the original 6-slot (4+2 slots) backplane
and a non-HP (Icy Box) 4-slot backplane (the Icy Box BP and the 4 slots
of teh HP backplane are connected to a Smart Array P400, the 2 slots of
the HP backplane are connected to a Smart Array E200), and ADU/ACU both
show the same 0:0 for the disks in the Icy Box backplane.
> What's also inconsistent is the drive show NCQ capable and enabled by
> the controller config shows false...
Sure, since the Smart Array P400 only supports SATA 1.5Gbps (SATA1)
which while not being a performance bottleneck doesn't support NCQ
(which requires SATA2).
Benjamin
Harout S. Hedeshian
> to the lack of a HP backplane. If you look at a HP backplane you'll
> notice that there is more on it than just SATA/SAS connectors and power
> connectors, it contains some logic which tells the controller the slot
> numbers and the type of enclosure. Standard backplanes don't have this
> logic, so the HP utilities report all disks as 0:0.
I was afraid of this. As far as I can tell, there doesn't seem to be
a workaround. It seems kind of odd that retailers (like Newegg) will
sell the card as a drop-in solution if you need to install it in HP
hardware to use the management tools? The breakout cables I bought
have 6-pin connectors for the SES logic. It looks like my backplane
(Supermicro SATA-743) doesn't have the headers. Is there a way to
override the management tool's faulty thinking? The card doesn't seem
to have a problem on bootup. This is quite crippling.
Benjamin Gawert
> I was afraid of this. As far as I can tell, there doesn't seem to be
> a workaround. It seems kind of odd that retailers (like Newegg) will
> sell the card as a drop-in solution if you need to install it in HP
> hardware to use the management tools?
You don't need to install it in a HP server just to use the management
tools.
> The breakout cables I bought
> have 6-pin connectors for the SES logic. It looks like my backplane
> (Supermicro SATA-743) doesn't have the headers. Is there a way to
> override the management tool's faulty thinking? The card doesn't seem
> to have a problem on bootup. This is quite crippling.
There is no way to change the management tools display (which isn't
faulty btw but fully correct), but there also is no need to. The
management tools still work perfectly fine, and if your backplane
supports activity/fault display (my Icy Box does) then if a drive fails
it will show up via the fault LED of the affected drive.
Also, you should not forget that these controllers are made for use in
HP ProLiant and Integrity servers and nowhere else. HP doesn't sell them
for use in other machines, and they also don't support that. However,
aside from the (more cosmetical) problem that the HP tools are unable to
detect a drive's location without the right backplane, the Smart Array
controllers work perfectly fine in most non-HP machines. Which is good
since the Smart Arrays are one of the best hardware RAID controllers you
can get.
Benjamin
Harout S. Hedeshian
I disagree. I can't add/change/assign hot spares, create new arrays
etc... because all drives are 0:0. That makes it BROKEN. Unless, of
course, your definition of broken is the same as Microsoft's.
I can see array status just fine, and migrate stripe size just fine. I
unplugged one drive and the ACUCLI showed all drives as failed
(although the array was clearly functional). I then reinserted the
drives and the controller rebuilt it no problem (and the ACU showed
correct status/progress). But seriously, no being able to manipulate
the p-disks is a showstopper because ORCA is useless.
I ordered the SAS backplane (CSE-SAS-743TQ) for my chassis and the
associated cables (CBL-0175L-01). The backplane supports SES-II
management through 8-pin sideband connectors. It supports I2C and
SGPIO. It will be interesting to see if it works.
Benjamin Gawert
>>> The management tools still work perfectly fine.
> I disagree. I can't add/change/assign hot spares, create new arrays
> etc... because all drives are 0:0.
Strange, as I can add/change/assign hot spares, create new arrays etc.
even with the drives that are listed as 0:0.
> That makes it BROKEN. Unless, of
> course, your definition of broken is the same as Microsoft's.
What should that be, another stupid anti-MS rant? I can give you a long
list of things that are still broken in Linux while working fine in
Windows for years, but that has nothing to do with the topic of this
thread or this newsgroup. So please keep your stupid OS flamewars out of
here.
And just a little reality check for you: it is *you* who put together a
bunch of unsupported hardware to build a server, by doing so it should
have been obvious that there are potential problems waiting to arise
(which is one reason why for anything that needs at least some form of
reliability it is just silly to build a server from standard hardware,
especially when considering that brand name servers like HP ProLiants
aren't necessarily more expensive).
The HP tools are *not* broken because they work for what they have been
designed for (managing HP disk drives in HP servers and drive enclosures
using HP controllers). It is hardly HP's fault if you use the wrong
hardware.
> I can see array status just fine, and migrate stripe size just fine. I
> unplugged one drive and the ACUCLI showed all drives as failed
> (although the array was clearly functional). I then reinserted the
> drives and the controller rebuilt it no problem (and the ACU showed
> correct status/progress). But seriously, no being able to manipulate
> the p-disks is a showstopper because ORCA is useless.
This sounds to me as if your backplane is a really primitive one which
only has SATA and power connectors. If this is the case then the
controller has no way to communicate with it.
> I ordered the SAS backplane (CSE-SAS-743TQ) for my chassis and the
> associated cables (CBL-0175L-01). The backplane supports SES-II
> management through 8-pin sideband connectors. It supports I2C and
> SGPIO. It will be interesting to see if it works.
It should, but the problem of ACU/ADU showing 0:0 for all drives will
probably persist.
Benjamin
Cydrome Leader
> * Harout S. Hedeshian:
>>>> The management tools still work perfectly fine.
>> I disagree. I can't add/change/assign hot spares, create new arrays
>> etc... because all drives are 0:0.
>
> Strange, as I can add/change/assign hot spares, create new arrays etc.
> even with the drives that are listed as 0:0.
>
>> That makes it BROKEN. Unless, of
>> course, your definition of broken is the same as Microsoft's.
>
> What should that be, another stupid anti-MS rant? I can give you a long
> list of things that are still broken in Linux while working fine in
> Windows for years, but that has nothing to do with the topic of this
> thread or this newsgroup. So please keep your stupid OS flamewars out of
> here.
>
> And just a little reality check for you: it is *you* who put together a
> bunch of unsupported hardware to build a server, by doing so it should
> have been obvious that there are potential problems waiting to arise
> (which is one reason why for anything that needs at least some form of
> reliability it is just silly to build a server from standard hardware,
> especially when considering that brand name servers like HP ProLiants
> aren't necessarily more expensive).
Ha.
Benjamin is correct here. It's just not possibly to make a storage system
better than HP does yourself. HP's PC server lineup (Proliants) even make
larger storage systems from places like Sun look childish and stupid as
well.