Cannot create pool: I/O error

m...@saudette.net

unread,

Feb 24, 2014, 1:09:07 PM2/24/14

to zfs-...@googlegroups.com

Summary:

I'm trying to create a pool with my iStoragePro enclosure with 8x2TB connected to an ATTO ExpressSAS R380 RAID card, but it always fails with an I/O error:

sudo zpool create -o ashift=12 Data raidz2 disk4s2 disk5s2 disk6s2 disk7s2 disk8s2 disk9s2 disk10s2 spare disk11s2

"cannot create 'Data': I/O error"

Could be related to this old topic and this closed (invalid) bug.

Details:

I'm running MacZFS-74.3.3 on Mountain Lion 10.8.5.

I've created a few other pools on this system already using internal drives, so I know my install is good.

The hardware is a MacPro 2,1 8-core, 32GB RAM with an iStoragePro enclosure with 8x2TB connected to an ATTO ExpressSAS R380 RAID card. I've been running this setup as a RAID-6 for over a year without a problem.

I wanted to convert to a RAIDZ2, so I switched the enclosure to JBOD. I can see all 8 disk and format them individually with Disk Utility, so everything is working there, but if I try to create a pool I get "cannot create 'Data': I/O error"

sudo zpool create -o ashift=12 Data raidz2 disk4s2 disk5s2 disk6s2 disk7s2 disk8s2 disk9s2 disk10s2 spare disk11s2

"cannot create 'Data': I/O error"

I've tried the following, without success...

adding /dev/ before all the disk
creating a pool with just a single disk
creating a mirror with just two of the disks
not using the ashift parameter
picking a different pool name (tank, puddle, etc)

Anyone have any ideas?

Regards,

Steve

MacPro:~ bump$ diskutil list

/dev/disk0

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *120.0 GB disk0

1: EFI 209.7 MB disk0s1

2: Apple_HFS Macintosh HD 119.2 GB disk0s2

/dev/disk1

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *3.0 TB disk1

1: EFI 209.7 MB disk1s1

2: ZFS pool 3.0 TB disk1s2

/dev/disk2

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *3.0 TB disk2

1: EFI 209.7 MB disk2s1

2: ZFS pool 3.0 TB disk2s2

/dev/disk3

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *3.0 TB disk3

1: EFI 209.7 MB disk3s1

2: ZFS pool 3.0 TB disk3s2

/dev/disk4

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk4

1: EFI 314.6 MB disk4s1

2: ZFS 2.0 TB disk4s2

/dev/disk5

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk5

1: EFI 314.6 MB disk5s1

2: ZFS 2.0 TB disk5s2

/dev/disk6

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk6

1: EFI 314.6 MB disk6s1

2: ZFS 2.0 TB disk6s2

/dev/disk7

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk7

1: EFI 314.6 MB disk7s1

2: ZFS 2.0 TB disk7s2

/dev/disk8

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk8

1: EFI 314.6 MB disk8s1

2: ZFS 2.0 TB disk8s2

/dev/disk9

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk9

1: EFI 314.6 MB disk9s1

2: ZFS 2.0 TB disk9s2

/dev/disk10

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk10

1: EFI 314.6 MB disk10s1

2: ZFS 2.0 TB disk10s2

/dev/disk11

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *2.0 TB disk11

1: EFI 314.6 MB disk11s1

2: ZFS 2.0 TB disk11s2

MacPro:~ bump$ zpool status -v

pool: pool

state: ONLINE

scrub: none requested

config:

NAME STATE READ WRITE CKSUM

pool ONLINE 0 0 0

disk2s2 ONLINE 0 0 0

disk3s2 ONLINE 0 0 0

disk1s2 ONLINE 0 0 0

errors: No known data errors

Gregg Wonderly

unread,

Feb 24, 2014, 4:35:56 PM2/24/14

to zfs-...@googlegroups.com

What about doing it without the partitions? I thought that it was best to use the whole disk anyway.

sudo zpool create -o ashift=12 Data raidz2 disk4 disk5 disk6 disk7 disk8 disk9 disk10 spare disk11

Gregg

--

---
You received this message because you are subscribed to the Google Groups "zfs-macos" group.
To unsubscribe from this group and stop receiving emails from it, send an email to zfs-macos+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Bjoern Kahl

unread,

Feb 24, 2014, 5:45:09 PM2/24/14

to zfs-...@googlegroups.com

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 24.02.14 22:35, schrieb Gregg Wonderly:

> What about doing it without the partitions? I thought that it was
> best to use the whole disk anyway.
>
> sudo zpool create -o ashift=12 Data raidz2 disk4 disk5 disk6 disk7
> disk8 disk9 disk10 spare disk11

IIRC, the recommended way on Mac OSX is to use a partition, and to
create it using diskutil. That way one meets the expectations of Mac
OSX' disk management. One _can_ go the full disk path, but it is not
the default method for MacZFS-74.x

In any case, I doubt that using whole disk access will help.

@ m...@saudette.net :

Can you check if there is anything MacZFS related in
/var/log/kernel.log or /var/log/system.log around your attempt
to create the raidz2? You may need to try "syslog -k Sender kernel |
grep -5 -e zfs -e ZFS" if no /var/log/kernel.log exist on your system.

Björn

> On 2/24/2014 11:09 AM, m...@saudette.net wrote:

>> *Summary:* * * I'm trying to create a pool with my iStoragePro

>> enclosure with 8x2TB connected to an ATTO ExpressSAS R380 RAID
>> card, but it always fails with an I/O error:
>>
>> sudo zpool create -o ashift=12 Data raidz2 disk4s2 disk5s2
>> disk6s2 disk7s2 disk8s2 disk9s2 disk10s2 spare disk11s2 "cannot
>> create 'Data': I/O error"
>>
>> Could be related to this old topic

>> <https://groups.google.com/forum/#%21topic/zfs-macos/kZQTiICDZR0>
>> and this closed (invalid) bug
>> <https://code.google.com/p/maczfs/issues/detail?id=107>.
>>
>>
>> *Details:*

>>
>> I'm running MacZFS-74.3.3 on Mountain Lion 10.8.5. I've created a
>> few other pools on this system already using internal drives, so
>> I know my install is good.
>>
>> The hardware is a MacPro 2,1 8-core, 32GB RAM with an
>> iStoragePro enclosure with 8x2TB connected to an ATTO ExpressSAS
>> R380 RAID card. I've been running this setup as a RAID-6 for over
>> a year without a problem.
>>
>> I wanted to convert to a RAIDZ2, so I switched the enclosure to
>> JBOD. I can see all 8 disk and format them individually with Disk
>> Utility, so everything is working there, but if I try to create a
>> pool I get "cannot create 'Data': I/O error"
>>
>> sudo zpool create -o ashift=12 Data raidz2 disk4s2 disk5s2
>> disk6s2 disk7s2 disk8s2 disk9s2 disk10s2 spare disk11s2 "cannot
>> create 'Data': I/O error"
>>
>> I've tried the following, without success...
>>

>> * adding /dev/ before all the disk * creating a pool with just a
>> single disk * creating a mirror with just two of the disks * not
>> using the ashift parameter * picking a different pool name (tank,

| Bjoern Kahl +++ Siegburg +++ Germany |
| "googlelogin@-my-domain-" +++ www.bjoern-kahl.de |
| Languages: German, English, Ancient Latin (a bit :-)) |
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQCVAgUBUwvLdFsDv2ib9OLFAQJpAgP/Y/e773y8WAa/Rz08J8wXWIfzbJCxyK7b
BFXO2cbpQuD3yd6e3X0/gTQ9CBwSUAdHyjv4lTghCXBZgoHJEdwgiOYkVIacJrRN
f98Kk5JXDzoV0ETaOoZNS+g6YClzedkEbP8CNgvtU6GiUsaQ6hAxohH7O9viDU2Z
GIy87UIHPrE=
=11Vl
-----END PGP SIGNATURE-----

m...@saudette.net

unread,

Feb 24, 2014, 6:07:12 PM2/24/14

to zfs-...@googlegroups.com

Thanks, but I get the same error when trying to use the whole device. The Getting Started guide suggests using the slice 2 approach.

I have made some progress though. I found a setting in the ATTO ExpressSAS R380 RAID card when running in JBOD mode to treat the drives as 512byte block sizes instead of 4k. The zpool command now works, but overall performance is unacceptable.

I'm using the ATTO config utility to low-level format the drive to see if that allows me to access the drive with 4k block sizes. I'll update when I know more.

Regards,

Steve

m...@saudette.net

unread,

Feb 24, 2014, 7:58:45 PM2/24/14

to zfs-...@googlegroups.com

Björn, thank you for the command, I was curious why I didn't have a kernel.log. I found nothing in the logs related to zfs when running the failed zpool create command.

As a follow-up to my last post, the low-level format finished on the drive, but the zpool command still returns the I/O error when trying to create a pool for the drive with 4K block size.

I failed to mention previously that the drives in question are Hitachi HDS72202 2TB. Hitachi has a tool that can re-align the drives to 4K, but I haven't tried it yet.

So for anyone else running into this issue, you might check your device block size with the diskutil info command:

This one works:

MacPro:~ bump$ diskutil info /dev/disk4

Device Identifier: disk4

Device Node: /dev/disk4

Part of Whole: disk4

Device / Media Name: ATTO 100 Media

Volume Name: Not applicable (no file system)

Mounted: Not applicable (no file system)

File System: None

Content (IOContent): None

OS Can Be Installed: No

Media Type: Generic

Protocol: SAS

SMART Status: Not Supported

Total Size: 2.0 TB (2000348512256 Bytes) (exactly 3906930688 512-Byte-Blocks)

Volume Free Space: Not applicable (no file system)

Device Block Size: 512 Bytes

Read-Only Media: No

Read-Only Volume: Not applicable (no file system)

Ejectable: Yes

Whole: Yes

Internal: No

OS 9 Drivers: No

Low Level Format: Not supported

This one doesn't work:

MacPro:~ bump$ diskutil info /dev/disk5

Device Identifier: disk5

Device Node: /dev/disk5

Part of Whole: disk5

Device / Media Name: ATTO 200 Media

Volume Name: Not applicable (no file system)

Mounted: Not applicable (no file system)

File System: None

Content (IOContent): None

OS Can Be Installed: No

Media Type: Generic

Protocol: SAS

SMART Status: Not Supported

Total Size: 2.0 TB (2000348512256 Bytes) (exactly 3906930688 512-Byte-Blocks)

Volume Free Space: Not applicable (no file system)

Device Block Size: 4096 Bytes

Read-Only Media: No

Read-Only Volume: Not applicable (no file system)

Ejectable: Yes

Whole: Yes

Internal: No

OS 9 Drivers: No

Low Level Format: Not supported

Steve

Reply all

Reply to author

Forward