HDIO_GET_IDENTITY ioctl error: -22

[Shonky]

Hi guys,

 

I've been evaluating various forms of NAS and keep coming back to unRAID. I've looked at FreeNAS, Openfiler, Windows Home Server, EON and a couple of others with varying success. Windows Home Server probably came the closest but I didn't like the redundancy feature effectively halving the available space.

 

However, I'm doing it with a bit of twist though trying to run it under VMware Server on a Windows 7 x64 host. I realise I'll take a bit of a performance hit but I really need Windows as a host to use the machine as a desktop. I'm getting a reasonable 25-30MB/sec transfer speed across the network between the host and guest so I'm happy with that.

 

Unfortunately VMware server doesn't seem to support any form of disk larger than exactly 1024GB. See my unanswered post on the VMware forums if you want more exact details:

 

http://communities.vmware.com/message/1448042#1448042

 

I'm trying to use direct physical access to 3 x 1.5TB SATA drives at this stage dedicated to unRAID. As IDE drives, I have irq timeout problems etc as discussed above. This happens with virtual disks as well as real disks and on other linux distributions like CentOS AND Windows Home Server so it's VMware issue. If I limit the direct disk access to <1024GB, unRAID and other OSes work just fine on IDE controllers.

 

So I looked at SCSI devices instead of IDE. The LSI Logic SAS controller works just fine in unRAID in terms of the driver support, disk size etc using direct disk access. unRAID detects them no problem and likewise I've created the direct disk access files all no problem and they work as devices.

 

Unfortunately unRAID 4.5 falls over when trying to add them to the system as drives. I get errors like this when I select the drives in the unRAID web GUI.

 

Jan  7 19:30:54 Mars kernel: md: import disk2: HDIO_GET_IDENTITY ioctl error: -22
Jan  7 19:30:54 Mars kernel: md: disk2 missing

 

Likewise when I execute "hdparm -i /dev/sda" I get:

 

/dev/sda
HDIO_GET_IDENTITY failed: Invalid argument

 

unRAID does appear to correctly detect the SCSI drive generally in the boot process e.g.

 

Jan  7 19:26:48 Mars kernel: scsi0 : ioc0: LSISAS1068 B0, FwRev=00000000h, Ports=1, MaxQ=128, IRQ=18
Jan  7 19:26:48 Mars kernel: mptsas: ioc0: attaching ssp device: fw_channel 0, fw_id 0, phy 0, sas_addr 0x5000c298abcffdad
Jan  7 19:26:48 Mars kernel: scsi 0:0:0:0: Direct-Access     VMware,  VMware Virtual S 1.0  PQ: 0 ANSI: 2
Jan  7 19:26:48 Mars kernel: mptsas: ioc0: attaching ssp device: fw_channel 0, fw_id 1, phy 1, sas_addr 0x5000c297d4ee4c2f
Jan  7 19:26:48 Mars kernel: scsi 0:0:1:0: Direct-Access     VMware,  VMware Virtual S 1.0  PQ: 0 ANSI: 2
Jan  7 19:26:48 Mars kernel: mptsas: ioc0: attaching ssp device: fw_channel 0, fw_id 2, phy 2, sas_addr 0x5000c2989218cbc1
Jan  7 19:26:48 Mars kernel: scsi 0:0:2:0: Direct-Access     VMware,  VMware Virtual S 1.0  PQ: 0 ANSI: 2
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] 2000000000 512-byte logical blocks: (1.02 TB/953 GiB)
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Write Protect is off
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Mode Sense: 5d 00 00 00
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:0:0: [sda] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel:  sda:
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] 2000000000 512-byte logical blocks: (1.02 TB/953 GiB)
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Write Protect is off
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Mode Sense: 5d 00 00 00
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:1:0: [sdb] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel:  sdb:
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] 2000000000 512-byte logical blocks: (1.02 TB/953 GiB)
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Write Protect is off
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Mode Sense: 5d 00 00 00
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Cache data unavailable
Jan  7 19:26:48 Mars kernel: sd 0:0:2:0: [sdc] Assuming drive cache: write through
Jan  7 19:26:48 Mars kernel:  sdc:

 

 

Sooo... is this more a driver issue that might be solved with a new kernel compile or something like that? Could unRAID be forced to work around this issue?

 

Any suggestions? I found one post asking about if this may have been fixed in an updated unRAID release but no response from there.

 

(PS what's with limiting posts to about 10 lines. The editor keeps jumping back to the top whlst typing - running IE8 - very annoying, so ended up typing it in notepad)

 

Thanks

Christian

[erikatcuse]

Take  look at my post http://lime-technology.com/forum/index.php?topic=3109.0 to see a history of LSI SAS related issues.  Bottom line is if you want to use your card you need to email Tom and ask if you can try 4.5.1sastest version of unraid.  I'm using it with no major problems right now and I'm sure he wouldn't mind anther tester.

 

Erik

[Shonky]

Thanks. I did search but I think I may have only been search in sub forums or something. I am a little confused though. I go to the top of this page and use the search box for "HDIO_GET_IDENTITY" minus quotes and it only shows my post. Anyway....

 

That was exactly a year ago and 4.5.1beta was still the next thing. 4.5 only just came out last month... Ah I see it *was* and old post and you came back and updated it. Cool.

 

The thing is I'm not using a physical card at all. I'm using what VMware server emulates as a LSI Logic SAS card. It appears to be the same problem though with the driver so hopefully the same fix will work.

 

I'll see if I can get on to Tom. Obviously I'm pretty new around here . I assume that is "tomm" known as limetech on the forums?

 

Thanks again

Christian

[erikatcuse]

Yes it is.  From the main web page [email protected]

[Shonky]

I got a pretty quick reply from Tom and he pointed me to a newer version.

 

4.5.1-sastest4 is working fairly well for me so far. I do have some concerns about the virtual machine though based on his comments.

 

Tom did also mention that this version may not pass through hard drive information and serial numbers through correctly which could cause problems if the drives move onto different controllers etc.