| 1 |
>>>> I have a USB stick with a crucial file on it (and only an old backup |
| 2 |
>>>> elsewhere). It's formatted NTFS because I wanted to be able to open |
| 3 |
>>>> the file on various Gentoo systems and my research indicated that |
| 4 |
>>NTFS |
| 5 |
>>>> was the best solution. |
| 6 |
>>>> |
| 7 |
>>>> I decided to copy a 10GB file from a USB hard disk directly to the |
| 8 |
>>USB |
| 9 |
>>>> stick this morning and I ran into errors so I canceled the operation |
| 10 |
>>>> and now the file manager (thunar) has been stuck for well over an |
| 11 |
>>hour |
| 12 |
>>>> and I'm getting errors like these over and over: |
| 13 |
>>>> |
| 14 |
>>>> [ 2794.535814] Buffer I/O error on dev sdc1, logical block 2134893, |
| 15 |
>>>> lost async page write |
| 16 |
>>>> [ 2794.535819] Buffer I/O error on dev sdc1, logical block 2134894, |
| 17 |
>>>> lost async page write |
| 18 |
>>>> [ 2794.535822] Buffer I/O error on dev sdc1, logical block 2134895, |
| 19 |
>>>> lost async page write |
| 20 |
>>>> [ 2794.535824] Buffer I/O error on dev sdc1, logical block 2134896, |
| 21 |
>>>> lost async page write |
| 22 |
>>>> [ 2794.535826] Buffer I/O error on dev sdc1, logical block 2134897, |
| 23 |
>>>> lost async page write |
| 24 |
>>>> [ 2794.535828] Buffer I/O error on dev sdc1, logical block 2134898, |
| 25 |
>>>> lost async page write |
| 26 |
>>>> [ 2794.535830] Buffer I/O error on dev sdc1, logical block 2134899, |
| 27 |
>>>> lost async page write |
| 28 |
>>>> [ 2794.535832] Buffer I/O error on dev sdc1, logical block 2134900, |
| 29 |
>>>> lost async page write |
| 30 |
>>>> [ 2794.535835] Buffer I/O error on dev sdc1, logical block 2134901, |
| 31 |
>>>> lost async page write |
| 32 |
>>>> [ 2794.535837] Buffer I/O error on dev sdc1, logical block 2134902, |
| 33 |
>>>> lost async page write |
| 34 |
>>>> [ 2842.568843] sd 9:0:0:0: [sdc] tag#0 FAILED Result: |
| 35 |
>>>> hostbyte=DID_ERROR driverbyte=DRIVER_SENSE |
| 36 |
>>>> [ 2842.568849] sd 9:0:0:0: [sdc] tag#0 Sense Key : Hardware Error |
| 37 |
>>[current] |
| 38 |
>>>> [ 2842.568852] sd 9:0:0:0: [sdc] tag#0 Add. Sense: No additional |
| 39 |
>>sense |
| 40 |
>>>> information |
| 41 |
>>>> [ 2842.568857] sd 9:0:0:0: [sdc] tag#0 CDB: Write(10) 2a 00 01 04 a4 |
| 42 |
>>>> 58 00 00 f0 00 |
| 43 |
>>>> [ 2842.568859] blk_update_request: I/O error, dev sdc, sector |
| 44 |
>>17081432 |
| 45 |
>>>> [ 2842.568862] buffer_io_error: 20 callbacks suppressed |
| 46 |
>>>> |
| 47 |
>>>> nmon says sdc is 100% busy but doesn't show any reading or writing. |
| 48 |
>>I |
| 49 |
>>>> once pulled the USB stick in a situation like this and I ended up |
| 50 |
>>>> having to reformat it which I need to avoid this time since the file |
| 51 |
>>>> is crucial. What should I do? |
| 52 |
>>>> |
| 53 |
>>>> - Grant |
| 54 |
>>> |
| 55 |
>>> Whatever you do, do NOT try to unplug the USB you were writing to |
| 56 |
>>unless you |
| 57 |
>>> first manage to successfully unmount it. If you do pull the USB |
| 58 |
>>stick |
| 59 |
>>> regardless of its current I/O state you will likely corrupt whatever |
| 60 |
>>file you |
| 61 |
>>> were writing onto, or potentially more. |
| 62 |
>>> |
| 63 |
>>> You could well have a hardware failure here, which manifested itself |
| 64 |
>>during |
| 65 |
>>> your copying operation. Things you could try: |
| 66 |
>>> |
| 67 |
>>> Run lsof to find out which process is trying to access the USB fs and |
| 68 |
>>kill it. |
| 69 |
>>> Then see if you can remount it read-only. |
| 70 |
>>> |
| 71 |
>>> Run dd, dcfldd, ddrescue to make an image of the complete USB stick |
| 72 |
>>on your |
| 73 |
>>> hard drive and then try to recover any lost files with testdisk. |
| 74 |
>>> |
| 75 |
>>> Use any low level recovery tools the manufacturer may offer - they |
| 76 |
>>will likely |
| 77 |
>>> require MSWindows. |
| 78 |
>>> |
| 79 |
>>> Shut down your OS and disconnect the USB stick, then reboot and try |
| 80 |
>>again to |
| 81 |
>>> access it, although I would first create an image of the device using |
| 82 |
>>any of |
| 83 |
>>> the above tools. |
| 84 |
>> |
| 85 |
>> |
| 86 |
>>dd errored and lsof returned no output at all so I tried to reboot but |
| 87 |
>>even that hung after the first 2 lines of output so I did a hard reset |
| 88 |
>>after waiting an hour or so. Once it came back up I tried to do 'dd |
| 89 |
>>if=/dev/sdb1 of=usbstick' but I got this after awhile: |
| 90 |
>> |
| 91 |
>>dd: error reading ‘/dev/sdb1’: Input/output error |
| 92 |
>>16937216+0 records in |
| 93 |
>>16937216+0 records out |
| 94 |
>>8671854592 bytes (8.7 GB) copied, 230.107 s, 37.7 MB/s |
| 95 |
>> |
| 96 |
>>It's a 64GB USB stick. dmesg says: |
| 97 |
>> |
| 98 |
>>[ 744.729873] sd 8:0:0:0: [sdb] tag#0 FAILED Result: hostbyte=DID_OK |
| 99 |
>>driverbyte=DRIVER_SENSE |
| 100 |
>>[ 744.729879] sd 8:0:0:0: [sdb] tag#0 Sense Key : Medium Error |
| 101 |
>>[current] |
| 102 |
>>[ 744.729883] sd 8:0:0:0: [sdb] tag#0 Add. Sense: Unrecovered read |
| 103 |
>>error |
| 104 |
>>[ 744.729886] sd 8:0:0:0: [sdb] tag#0 CDB: Read(10) 28 00 01 02 78 e0 |
| 105 |
>>00 00 f0 00 |
| 106 |
>>[ 744.729889] blk_update_request: critical medium error, dev sdb, |
| 107 |
>>sector 16939232 |
| 108 |
>>[ 744.763468] sd 8:0:0:0: [sdb] tag#0 FAILED Result: hostbyte=DID_OK |
| 109 |
>>driverbyte=DRIVER_SENSE |
| 110 |
>>[ 744.763472] sd 8:0:0:0: [sdb] tag#0 Sense Key : Medium Error |
| 111 |
>>[current] |
| 112 |
>>[ 744.763474] sd 8:0:0:0: [sdb] tag#0 Add. Sense: Unrecovered read |
| 113 |
>>error |
| 114 |
>>[ 744.763478] sd 8:0:0:0: [sdb] tag#0 CDB: Read(10) 28 00 01 02 79 00 |
| 115 |
>>00 00 04 00 |
| 116 |
>>[ 744.763480] blk_update_request: critical medium error, dev sdb, |
| 117 |
>>sector 16939264 |
| 118 |
>>[ 744.763482] Buffer I/O error on dev sdb1, logical block 4234304, |
| 119 |
>>async page read |
| 120 |
>>[ 744.786743] sd 8:0:0:0: [sdb] tag#0 FAILED Result: hostbyte=DID_OK |
| 121 |
>>driverbyte=DRIVER_SENSE |
| 122 |
>>[ 744.786747] sd 8:0:0:0: [sdb] tag#0 Sense Key : Medium Error |
| 123 |
>>[current] |
| 124 |
>>[ 744.786750] sd 8:0:0:0: [sdb] tag#0 Add. Sense: Unrecovered read |
| 125 |
>>error |
| 126 |
>>[ 744.786753] sd 8:0:0:0: [sdb] tag#0 CDB: Read(10) 28 00 01 02 79 04 |
| 127 |
>>00 00 04 00 |
| 128 |
>>[ 744.786755] blk_update_request: critical medium error, dev sdb, |
| 129 |
>>sector 16939268 |
| 130 |
>>[ 744.786758] Buffer I/O error on dev sdb1, logical block 4234305, |
| 131 |
>>async page read |
| 132 |
>> |
| 133 |
>>I haven't tried to mount it yet. Any suggestions? |
| 134 |
>> |
| 135 |
>>- Grant |
| 136 |
> |
| 137 |
> Try ddrescue. |
| 138 |
> It tries to skip the really bad parts. |
| 139 |
> |
| 140 |
> Then try data recovery on copies of the resulting image file. |
| 141 |
|
| 142 |
|
| 143 |
I did: |
| 144 |
|
| 145 |
# ddrescue -d -r3 /dev/sdb usb.img usb.log |
| 146 |
GNU ddrescue 1.20 |
| 147 |
Press Ctrl-C to interrupt |
| 148 |
rescued: 62742 MB, errsize: 32768 B, current rate: 0 B/s |
| 149 |
ipos: 8672 MB, errors: 1, average rate: 14878 kB/s |
| 150 |
opos: 8672 MB, run time: 1h 10m 17s, remaining time: n/a |
| 151 |
time since last successful read: 5s |
| 152 |
|
| 153 |
# mount -o loop,ro usb.img /mnt/usbstick |
| 154 |
mount: wrong fs type, bad option, bad superblock on /dev/loop0, |
| 155 |
missing codepage or helper program, or other error |
| 156 |
|
| 157 |
In some cases useful info is found in syslog - try |
| 158 |
dmesg | tail or so. |
| 159 |
|
| 160 |
# mount -o loop,ro -t ntfs usb.img /mnt/usbstick |
| 161 |
NTFS signature is missing. |
| 162 |
Failed to mount '/dev/loop0': Invalid argument |
| 163 |
The device '/dev/loop0' doesn't seem to have a valid NTFS. |
| 164 |
Maybe the wrong device is used? Or the whole disk instead of a |
| 165 |
partition (e.g. /dev/sda, not /dev/sda1)? Or the other way around? |
| 166 |
|
| 167 |
How else can I get my file from the ddrescue image of the USB stick? |
| 168 |
|
| 169 |
- Grant |
Archives