Help with Fusion/corestorage repair

My Fusion volume metadata got corrupted and I'm trying to repair it. I didn't have it backed up because a recent update to OS X broke support for my eSATA ExpressCard.

A botched attempt to use it by installing ubuntu seems to have corrupted it.

This is what it looked like when it was working:


[14-07-01@13:04][elvey@Computer-of-ME ~]$ diskutil corestorage list

CoreStorage logical volume groups (1 found)

|

+-- Logical Volume Group 147B7B2D-25FE-471F-BC2B-F40D401E5C00

=========================================================

Name: Fusion

Status: Online

Size: 620999999488 B (621.0 GB)

Free Space: 20889600 B (20.9 MB)

|

+-< Physical Volume B40D6368-C887-4918-87FC-B1EAC13C462A

| ----------------------------------------------------

| Index: 0

| Disk: disk1s4

| Status: Online

| Size: 492999999488 B (493.0 GB)

|

+-< Physical Volume 1452797C-9D64-49FB-9222-ED2B24DFDD1B

| ----------------------------------------------------

| Index: 1

| Disk: disk0s2

| Status: Online

| Size: 128000000000 B (128.0 GB)

|

+-> Logical Volume Family E2BC6303-B351-4DC5-8E44-BF127D15B1D6

----------------------------------------------------------

Encryption Status: Unlocked

Encryption Type: None

Conversion Status: NoConversion

Conversion Direction: -none-

Has Encrypted Extents: No

Fully Secure: No

Passphrase Required: No

|

+-> Logical Volume 5CC94C41-889F-4C93-9916-7A1B633CBBFA

---------------------------------------------------

Disk: disk2

Status: Online

Size (Total): 615074922496 B (615.1 GB)

Size (Converted): -none-

Revertible: No

LV Name: Macintosh HD

Volume Name: Macintosh HD

Content Hint: Apple_HFS


This is what it looks like now:


Computer-of-ME:~ newadmin$ diskutil corestorage list

CoreStorage logical volume groups (1 found)

|

+-- Logical Volume Group 147B7B2D-25FE-471F-BC2B-F40D401E5C00

=========================================================

Name: Fusion

Status: Offline

Size: 492484210688 B (492.5 GB)

Free Space: -none-

|

+-< Physical Volume B40D6368-C887-4918-87FC-B1EAC13C462A

| ----------------------------------------------------

| Index: 0

| Disk: disk1s4

| Status: Checking

| Size: 492484210688 B (492.5 GB)

|

+-< Physical Volume 1452797C-9D64-49FB-9222-ED2B24DFDD1B

----------------------------------------------------

(No properties)

Computer-of-ME:~ newadmin$ diskutil list

/dev/disk0

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *256.1 GB disk0

1: EFI 209.7 MB disk0s1

2: FFFFFFFF-FFFF-FFFF-FFFF-FFFFFFFFFFFF 128.0 GB disk0s2

3: Apple_HFS Boot OS X 134.2 MB disk0s3

4: Microsoft Basic Data 1.0 MB disk0s4

5: 0FC63DAF-8483-4772-8E79-3D69D8477DE4 121.3 GB disk0s5

6: Linux Swap 6.4 GB disk0s6

/dev/disk1

#: TYPE NAME SIZE IDENTIFIER

0: GUID_partition_scheme *1.0 TB disk1

1: EFI 209.7 MB disk1s1

2: Apple_HFS HD 400.0 GB disk1s2

3: Apple_HFS scratch 100.0 GB disk1s3

4: Apple_CoreStorage 492.5 GB disk1s4

5: Apple_Boot Recovery HD 650.0 MB disk1s5

6: Apple_HFS Install OS X Mavericks 6.5 GB disk1s6


So, it looks to me like the underlying data is still fine: /dev/disk0s2 is still there and 128 GB; it's type is wacky. /dev/disk1s4 is still looking fine.
Those are/were the two underlying PV. But I'm looking at the diskutil manpage and I'm at a loss with respect to how to undo the damage. create | createLVG is destructive, so a no-no. I've backed up the remaining valuable data - the contents disk1s2. I'm thinking of experimenting; in preparation, I've backed up disk0s2 to a .dmg and will back up disk1s4 to a .dmg too. I have a copy of the partition tables of disk0 and disk1 from before the damage occurred, in the form of dd dumps of the first 5 blocks of the disks, thanks to:

/bin/dd if=/dev/rdisk0 of=/HD/Backups/GPTs/rdisk0.raw.gpt bs=512 count=5 conv=noerror
/bin/dd if=/dev/rdisk1 of=/HD/Backups/GPTs/rdisk1.raw.gpt bs=512 count=5 conv=noerror


Where is corestorage metadata actually stored? I'm wondering if the (incomplete TM) backups I do have, have backed it up.

Maybe the corestorage metadata is actually fine and its just the partition table that needs repair...


Message was edited by: MrElvey Add final line.

MacBook Pro (15-inch Early 2008), OS X Mountain Lion (10.8.2), SSD, USB, FireWire, external drives

Posted on Aug 11, 2014 11:50 PM

Reply
Question marked as ⚠️ Top-ranking reply

Posted on Aug 12, 2014 4:38 PM

I guess I don't care if this this a destructive operation:
diskutil coreStorage create myLogicalVolGroup /dev/disk0s2 /dev/disk1s4

since my LVG still exists, and my PVs still exist (though the latter seems messed up); I do NOT need to run it.

I guess I DO need to run something like

diskutil coreStorage createLV 147B7B2D-25FE-471F-BC2B-F40D401E5C00 <type> <name> <size>

Where:

147B7B2D-25FE-471F-BC2B-F40D401E5C00 is my lvgUUID - parameter 1.

type is Journaled HFS+ ??? Journaled\ HFS+??? - parameter 2.

name is Phoenix - parameter3 - since it's rising from ashes!

size is 615074922496B - parameter 4 (better than "100%", I guess...


But first, I'd really like to know if createLV is a destructive operation!

5 replies
Sort By: 
Question marked as ⚠️ Top-ranking reply

Aug 12, 2014 4:38 PM in response to MrElvey

I guess I don't care if this this a destructive operation:
diskutil coreStorage create myLogicalVolGroup /dev/disk0s2 /dev/disk1s4

since my LVG still exists, and my PVs still exist (though the latter seems messed up); I do NOT need to run it.

I guess I DO need to run something like

diskutil coreStorage createLV 147B7B2D-25FE-471F-BC2B-F40D401E5C00 <type> <name> <size>

Where:

147B7B2D-25FE-471F-BC2B-F40D401E5C00 is my lvgUUID - parameter 1.

type is Journaled HFS+ ??? Journaled\ HFS+??? - parameter 2.

name is Phoenix - parameter3 - since it's rising from ashes!

size is 615074922496B - parameter 4 (better than "100%", I guess...


But first, I'd really like to know if createLV is a destructive operation!

Reply

Jun 18, 2015 10:18 PM in response to MrElvey

I had a similar issue, in that my Fusion drive wouldn't boot. It'd hang after the login screen. When I rebooted into recovery mode (cmd+R), and ran disk utils, I got scary messages about the "The Fusion Drive is Missing a Disk", and a warning that repairing it would erase the data on the following drives... with an empty list. I bailed out of that because I didn't want to lose data. Running diskutil on the cmd line things didn't look right there either.


What ended up fixing it:


I rebooted from the internet (alt+cmd+R).

wait while the world spins (4-5 minutes)

open the terminal and run "diskutil cs list".

I made note of the drives (disk0s2 and disk1s2), and also noticed the status of one of the volumes was "Locked" not "Online"

I ran "fsck_cs -p disk0s2" and "fsck_cs -p disk1s2" - no errors were reported!

After this, the aforementioned status became "Online" (from "Locked")

I rebooted, logged in (success!) and immediately started a TimeMachine back up.


I believe that this fault in the Fusion drive started either with Bootcamp misbehaving (malware?) or a bad power-down (my wife and kids apparently fought over the power button before this started).


Hope this helps anyone else landing here from a web search. I know I was freaking out.

Reply

This thread has been closed by the system or the community team. You may vote for any posts you find helpful, or search the Community for additional answers.

Help with Fusion/corestorage repair

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple Account.