Fusion-io ioDrive 2 1.2TB Reference Page

[acquacow]

Can anyone confirm the sustained write speeds on the 825Gb version of this card? I used to get 800MB/sec at least but for some reason performance of the card went down on the new system and I cannot figure out why. Now I am getting 400MB/sec and I am not sure if this is the expected speed for this card.

Should be 1.5Gb/sec read, 1.1GB/sec write for any ioDrive II over 785GB. This will be slightly slower with a filesystem on it.

Your drive is the same as the 785, but with less over-provisioning, so you have a tad less write "buffer" for sustained write before you start putting pressure on the groomer.

You can adjust this over-provisioning yourself with fio-format to reduce formatted size and increase free NAND for sustained writes.

 

[rootpeer]

Can anyone confirm the sustained write speeds on the 825Gb version of this card? I used to get 800MB/sec at least but for some reason performance of the card went down on the new system and I cannot figure out why. Now I am getting 400MB/sec and I am not sure if this is the expected speed for this card.

Are you sure you are using a PCIe slot that is not PCIe 2.0 x8 or PCIe 3.0 x4 or otherwise limited in speed like going through the chipset? 400MB/s sounds like it is close at the 500MB/s theoretical limit for said slot configurations.

Edit: I got my calculations wrong, you need to be running at PCIE 2.0 x1 for 500MB/s. Still might be limited if it is plugged into a chipset slot.

 

[Bert]

Thank you very much everyone for confirming the issue. It seems like I am also having a heat issue. The card shut down after reaching to temp limit.

I remember I was able to hit 800MB/sec on the same card. I am using it in a HP Z840 workstation, not an ideal set up.

 

[acquacow]

Thank you very much everyone for confirming the issue. It seems like I am also having a heat issue. The card shut down after reaching to temp limit.

I remember I was able to hit 800MB/sec on the same card. I am using it in a HP Z840 workstation, not an ideal set up.

Stick a 60mm noctua on your IOH heatsink aimed at the ioDrive and you'll be fine.

 

[Bert]

For now, I just increased the idle fan speed of hp 840. Quite noisy but lowered the temps to 55C.

Thanks for the tip though. How am I going to keep it stable there and power the fan?

 

[acquacow]

For now, I just increased the idle fan speed of hp 840. Quite noisy but lowered the temps to 55C.

Thanks for the tip though. How am I going to keep it stable there and power the fan?

Hmm, doesn't look like you have as much room as I did in my Z420

http://imgur.com/0UecnjL

I was able to just stick a high-flowing 120mm fan there, ziptied to the drive cage above it. Ran like this for years running ESX.

If you can re-route some of the wires on the bottom, or can build a stand-off from the bottom of the case, you should be able to attach a fan closer to the pci-e cards no problem...

 

[naimc]

Suggestions on fio-attach failed.

I have about a dozen of these cards (6 TB SX350 SanDisk and Cisco branded ) running on ProLiant DL380p Gen8 win Server 2016. a few times a year one of these server will BSOD, and typical the fio-attach fails. We have to power off the server let it site for a few min and on power on the fio-attach will work.

This time that did not work for us.

device manager shows the same errors when we manually try to attach the card or the auto-attach as system start.

(Windows event logs events in SYSTEM)
fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18448 Unable to attach - load metadata (-22).
fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 16384 Internal software error (data)
fiowrn Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): Unable to save metadata. Reattach will require rebuild.
fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18455 Unable to attach device (-22).
fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18460 Unable to attach - failed to create device: Invalid argument (-22).

The "Reattach will require rebuild" hints at another command ? We tried the fio-attach -r ( Force a metadata rescan.)

Does rebuild only mean fio-sure-erase.exe followed by a fio-format.exe ?

Could not find other command that would do this rebuild.


Info about the card :

fct1 Detached
ioMemory Adapter Controller, Product NumberFIOS64002, SN:1710D00F2
ioMemory Adapter Controller, PN:5491-72338-6400G
Microcode Versions: App:0.0.14.0
Powerloss protection: protected
PCI:07:00.0, Slot Number:2
Vendor:1aed, Device:3001, Sub vendor:1137, Sub device:19d
Firmware v8.9.9, rev 20170222 Public
6400.00 GBytes device size
Format: v501, 1562500000 sectors of 4096 bytes
PCIe slot available power: unavailable
PCIe negotiated link: 8 lanes at 5.0 Gt/sec each, 4000.00 MBytes/sec total
Internal temperature: 37.90 degC, max 52.17 degC
Internal voltage: avg 1.01V, max 1.01V
Aux voltage: avg 1.78V, max 1.80V
Reserve space status: Healthy; Reserves: 100.00%, warn at 10.00%
Active media: 100.00%
Rated PBW: 22.00 PB, 99.01% remaining
Lifetime data volumes:
Physical bytes written: 218,389,506,209,040
Physical bytes read : 589,605,049,511,936
RAM usage:
Current: 36,132,032 bytes
Peak : 388,767,936 bytes

Thank you for any suggestions.

 

[acquacow]

Haven't run into that before. Have you tried -R to do a slow rescan?

 

[naimc]

thanks for the suggestion, Might be Linux thing ? , the -R

On windows :

C:\Windows\system32>fio-attach -R /dev/fct1
Unknown option -- R
Error: Invalid option R (0x52)

Lower case only on windows .

C:\Windows\system32>fio-attach -r /dev/fct1
Attaching: [========== ] ( 50%) /

C:\Windows\system32>fio-attach -r /dev/fct1
Attaching: [====================] (100%)
Error: Device '/dev/fct1' had an error while attaching. Invalid input.
Additional details about the error may be found in system

Same result on this drive.

 

[acquacow]

Yeah, I saw it when I ran 'strings' against fio-attach on my linux box that is running VSL 4.3.7

Can you try loading it up on a linux instance to see if it attaches as a block device?

 

[naimc]

Linux OS , not availed for us, We decided to secure erase the card, Initial that worked it got the card re-attached but a few hours later it failed again.

We powered the server off /on a few times now the fio-sure-erase is failing everytime :

5/15/2023 5:17:08 PM fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 06161 ioDrive failed to attach, errorcode-2.
5/15/2023 5:17:08 PM fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18460 Unable to attach - failed to create device: No such file or directory (-2).
5/15/2023 5:17:08 PM fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18455 Unable to attach device (-2).
5/15/2023 5:17:08 PM fioerr Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): 18433 Unable to attach - ebo (-2).
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): ***************************************************
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): *** Please contact Customer Support. ***
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): *** No media format found! ***
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): ***************************************************
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): Attempting to load ioDrive.
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): probed fct1
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): Init NV data
5/15/2023 5:17:08 PM Cisco UCS 6400GB SanDisk ioMemory SX350 (7,2): Unable to recover management NV data (-2).


We are now in un-charted waters ! fear the card is dead)

Any point in trying to re-flash the drive ? It's hard to understand why is failed given the NAND values still look good :

Rated PBW: 22.00 PB, 100.00% remaining
Lifetime data volumes:
Physical bytes written: 1,900,544
Physical bytes read : 579,411,968

 

[Whaaat]

It's hard to understand why is failed given the NAND values still look good :

Rated PBW: 22.00 PB, 100.00% remaining
Lifetime data volumes:
Physical bytes written: 1,900,544
Physical bytes read : 579,411,968

Two posts above it was another card with 99.01% life remaining

 

[bucovaina89]

Hi I'm new here, I just got an HPe QK762A Mezzanine card for the BladeSystem for free. I'm trying to make it work under Debian 12 but I'm hitting against the stddef.h problem which was also mentioned by @Bert . I read the posts but not really sure what I'm supposed to do now. I got this script from the proxmox forums. I guess this will probably work on Debian 11:

root@blade01:~# cat iodrive2.sh
#apt update && apt install --assume-yes pve-headers pve-headers-`uname -r` zip unzip gcc fakeroot build-essential debhelper rsync dkms && apt upgrade && apt autoremove --assume-yes && \
apt update && apt install --assume-yes linux-headers-$(uname -r) zip unzip gcc fakeroot build-essential debhelper rsync dkms && apt upgrade && apt autoremove --assume-yes && \
mkdir /home/temp && cd /home/temp && \
wget -O iomemory-vsl.zip https://www.dropbox.com/s/ktj2ive9elah04n/iomemory-vsl-4.20.1.zip?dl=1 && \
wget -O fio-common_3.2.16.1731-1.0_amd64.deb https://www.dropbox.com/s/pd2ohfaufhwqc34/fio-common_3.2.16.1731-1.0_amd64.deb?dl=1 && \
wget -O fio-firmware-fusion_3.2.16.20180821-1_all.deb https://www.dropbox.com/s/kcn5agi6lyikicf/fio-firmware-fusion_3.2.16.20180821-1_all.deb?dl=1 && \
wget -O fio-sysvinit_3.2.16.1731-1.0_all.deb https://www.dropbox.com/s/g39l6lg9of6eqze/fio-sysvinit_3.2.16.1731-1.0_all.deb?dl=1 && \
wget -O fio-util_3.2.16.1731-1.0_amd64.deb https://www.dropbox.com/s/57huby17mteg6wp/fio-util_3.2.16.1731-1.0_amd64.deb?dl=1 && \
unzip iomemory-vsl.zip && cd /home/temp/iomemory-vsl && \
cp -r /home/temp/iomemory-vsl/root/usr/src/iomemory-vsl-3.2.16 /usr/src/ && \
mkdir -p /var/lib/dkms/iomemory-vsl/3.2.16/build && \
ln -s /usr/src/iomemory-vsl-3.2.16 /var/lib/dkms/iomemory-vsl/3.2.16/source && \
dkms build -m iomemory-vsl -v 3.2.16 && \
dkms install -m iomemory-vsl -v 3.2.16 && \
modprobe iomemory-vsl && \
cd /home/temp && \
dpkg -i fio-firmware-fusion_3.2.16.20180821-1_all.deb fio-util_3.2.16.1731-1.0_amd64.deb fio-sysvinit_3.2.16.1731-1.0_all.deb fio-common_3.2.16.1731-1.0_amd64.deb

When DKMS rebuilds the module, I get this:

compilation terminated.
make[2]: *** [/usr/src/linux-headers-6.1.0-10-common/scripts/Makefile.build:255: /var/lib/dkms/iomemory-vsl/3.2.16/build/driver_init.o] Error 1
make[2]: *** [/usr/src/linux-headers-6.1.0-10-common/scripts/Makefile.build:255: /var/lib/dkms/iomemory-vsl/3.2.16/build/kcpu.o] Error 1
In file included from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/linux/ktypes.h:48,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/ktypes.h:76,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/kfio.h:37,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/fio-port.h:32,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/pci.c:38:
/var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/common-linux/commontypes.h:37:10: fatal error: stddef.h: No such file or directory
   37 | #include <stddef.h>
      |          ^~~~~~~~~~
compilation terminated.
In file included from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/linux/ktypes.h:48,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/ktypes.h:76,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/kfio.h:37,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/dbgset.h:33,
                 from /var/lib/dkms/iomemory-vsl/3.2.16/build/kfio.c:55:
/var/lib/dkms/iomemory-vsl/3.2.16/build/include/fio/port/common-linux/commontypes.h:37:10: fatal error: stddef.h: No such file or directory
   37 | #include <stddef.h>
      |          ^~~~~~~~~~
compilation terminated.
make[2]: *** [/usr/src/linux-headers-6.1.0-10-common/scripts/Makefile.build:255: /var/lib/dkms/iomemory-vsl/3.2.16/build/pci.o] Error 1
make[2]: *** [/usr/src/linux-headers-6.1.0-10-common/scripts/Makefile.build:255: /var/lib/dkms/iomemory-vsl/3.2.16/build/kfio.o] Error 1
make[1]: *** [/usr/src/linux-headers-6.1.0-10-common/Makefile:2037: /var/lib/dkms/iomemory-vsl/3.2.16/build] Error 2
make[1]: Leaving directory '/usr/src/linux-headers-6.1.0-10-amd64'
make: *** [Makefile:111: modules] Error 2

Do I need to switch branches?

root@blade01:~# uname -r
6.1.0-10-amd64

 

[COROVICD]

Please help. I'm trying to install Fusion-io ioScale2 1.65TB PCIe SSD F11-003-1T65-CS-0001 on Windows 10 on Dell T5400 Workstation. Driver installation worked OK and I can see the device in the Device Manager, but there is nothing to be seen in the Disk Manager.

Deriver version is 3.2.15 and firmware version is 7.1.17. Is that OK?

This is the report I get when I do fio-status command:


C:\WINDOWS\system32>fio-status -a

Found 1 ioMemory device in this system
Driver version: 3.2.15 build 1699

Adapter: ioMono
Fusion-io 1.65TB ioScale2, Product Number:F11-003-1T65-CS-0001, SN:1412G0444, FIO SN:__________
ioDrive2 Adapter Controller, PNA005004003
External Power: NOT connected
PCIe Power limit threshold: 24.75W
PCIe slot available power: unavailable
Connected ioMemory modules:
fct0: Product Number:F11-003-1T65-CS-0001, SN:__________

fct0 Status unknown: Driver is in MINIMAL MODE:
General channel initialization failure
ioDrive2 Adapter Controller, Product Number:F11-003-1T65-CS-0001, SN:__________
!! ---> There are active errors or warnings on this device! Read below for details.
ioDrive2 Adapter Controller, PNA005004003
SMP(AVR) Versions: App Version: 1.0.19.0, Boot Version: 1.0.5.1
Powerloss protection: not available
PCI:06:00.0
Vendor:1aed, Device:2001, Sub vendor:1aed, Sub device:2001
Firmware v7.1.17, rev 116786 Public
Geometry and capacity information not available.
Format: not low-level formatted
PCIe slot available power: 25.00W
PCIe negotiated link: 1 lanes at 2.5 Gt/sec each, 250.00 MBytes/sec total
Internal temperature: 73.82 degC, max 73.82 degC
Internal voltage: avg 1.01V, max 1.01V
Aux voltage: avg 2.49V, max 2.49V
Rated PBW: 8.00 PB
Lifetime data volumes:
Physical bytes written: 0
Physical bytes read : 0
RAM usage:
Current: 0 bytes
Peak : 0 bytes

ACTIVE WARNINGS:
The bandwidth of the PCI slot is not optimal for the ioMemory.
The ioMemory is currently running in a minimal state.


. . . as well, when I run fio-pci-check I get this:

C:\WINDOWS\system32>fio-pci-check
This utility directly probes the pci configuration ports.
There are possible system stability risks in doing this.
Do you wish to continue [y/n]? y

Root Bridge PCIe 10250 MB/sec needed max
Memory Controller

Bridge 0000:00:09.00 (03-07) (8086:4029)
PCI-to-PCI Bridge
* Needed 2000 MB/sec Avail 1000 MB/sec

Bridge 0000:03:00.00 (04-06) (8086:3500)
PCI-to-PCI Bridge
* Needed 2000 MB/sec Avail 1000 MB/sec
* Non-Fatal Error(s): Detected
* Unsupported Request(s): Detected
Clearing Errors

Bridge 0000:04:01.00 (06-06) (8086:3514)
PCI-to-PCI Bridge
* Needed 2000 MB/sec Avail 250 MB/sec

ioDrive 0000:06:00.0 (2001) Firmware 116786
Fusion ioMemory VSL Device
Slot Power limit: 25.0W (25000mw)
WARNING: All PCIe lanes are not in service.

 

[acquacow]

Move it to a x4 slot and try again. It isn't going to work in a x1

 

[COROVICD]

Move it to a x4 slot and try again. It isn't going to work in a x1

Thanks for the answer but I have a problem deciding which slot is x4, in my Dell Precision T5400 Workstation.

According to the User Guide and markings on the system board PCI-e Slot #1, is the one on the far left and it is x4 Electrical / x8 Mechanical slot. I already put card in that slot and its not working there. As you can see, there are no other slots that are marked x4. So, I'm going to try putting the Fusion-io ioScale2 ioDrive2 card into PCI-e #4, which is actually x16 Gen-2, although that slot has many more connections than Fusion-io card needs.

Is that a good idea? Is there a risk of causing magic smoke?

 

[COROVICD]

yeah, I moved Fusion-io ioScale2 1.65TB PCIe SSD F11-003-1T65-CS-0001 from PCI-E slot #1 (x4/x8 PCI-E slot) to PCI-E slot #4 (x16 PCI-E 2.0 slot) and nothing happened. Still the same error " The bandwidth of the PCI slot is not optimal for the ioMemory" and nothing is showing in the Disk Manager for Windows 10.

Actually, problem got worst. Fusion-io device used to be "fct0", but now its changed to "fct1", but Fusion-io driver thinks that device is in PCI-E slot #1, while device is really physically in the slot #4.

Pity its so hard to get this storage drive working. Its really fantastic device, when it works.

 

[michel333]

I think you need to uninstall and reinstall the driver, if you change pcie slot

 

[michel333]

And also there is a command for fusion driver, so you override 25 w power to 75 w power if your pcie slot supports it

 

[COROVICD]

I've done all of that without any luck.

My PC is very old Dell T5400 Workstation built back in 2009, although system board has two PCI-E 2.0 slots. My Fusion-io ioDrive2 was built in 2012. maybe that's a problem?