[linux-lvm] hard drive crash info please

[linux-lvm] hard drive crash info please

[Randy Perkins]

hello,
	I have a lvm setup of 5 PV (disks), 1 VG,  and 1 LV.  one of the drives
just started clicking, and is not recognizable by the bios. Now when i
boot up my volume group is not recognized.
some info:

redhat 8.0
lvm 1.0.3
readable copy of /etc/lvmconf/my_vg.conf ( about 5 months old )


	How can i retrieve some of the data that is contained soley
on the other 4 drives. ?? Any help is appreciated.


	i have been looking over the list,
and it appears this thread talks about extracting data from LV
that didnt span the bad disk, but i only have one LV.
http://lists.sistina.com/pipermail/linux-lvm/2003-April/013911.html
it looks like he is using lvm2
.
	does lvm2 have additional tools that can be used to bring back some of
my data. ??

i was just keeping mp3 on this LV, but when i upgraded redhat i moved
all my data to it for awhile, and never moved it back.  my wife is
pissed cause i lost all my digital pictures, and i lost all my personal
files.  i knew better but didnt listen to myself.  the hard drive gave
no errors in the logs, which i monitor, it just up and started beating
its head against the wall.  thanks for any help

randy

Re: [linux-lvm] hard drive crash info please

[Emmanuel Varagnat]

Randy Perkins wrote:
> hello,
> 	I have a lvm setup of 5 PV (disks), 1 VG,  and 1 LV.  one of the drives
> just started clicking, and is not recognizable by the bios. Now when i
> boot up my volume group is not recognized.

I did have the same problem a long time ago. So I started to write a 
tool that would try to recover some the data from the remaining data 
available. I lost the first disk of set of 3 disk...
This is quite finished.
It uses the capability of LVM2 to deal with 'partial' VG.
It only works with ext2/ext3 filesystems and require another (temporary) 
disk where recovered data will be written.

The URL is at : http://coredump.free.fr/linux/e2retrieve.php

I also did a floppy disk with the essential (kernel 2.4.19 with LVM2 
patch, LVM2 tools, Busybox and e2retrieve).

-=( manu )=-

Re: [linux-lvm] hard drive crash info please

[Kai Iskratsch]

hi

you need lvm2 to  activate a partial volume group. if you have a backup 
of your vgcfg  then you may restore your config to a new harddisk with 
the lvm1 tools.
i had no actual vgcfgbackup so i needed to upgrade to lvm2 to activate 
the partial vg and make a backup of all partitions that were not 
affected by the crashed harddisk. then i used the vgcfgbackup on the 
partial vg and edited the meadata from the backup (lvm2 saves the backup 
data in human readable form) that all missing segments are put on the 
replacement disc. after i restored the backup everything worked again 
and i lost only the data which was on the bad disk.

kai



.

Randy Perkins wrote:

>hello,
>	I have a lvm setup of 5 PV (disks), 1 VG,  and 1 LV.  one of the drives
>just started clicking, and is not recognizable by the bios. Now when i
>boot up my volume group is not recognized.
>some info:
>
>redhat 8.0
>lvm 1.0.3
>readable copy of /etc/lvmconf/my_vg.conf ( about 5 months old )
>
>
>	How can i retrieve some of the data that is contained soley
>on the other 4 drives. ?? Any help is appreciated.
>
>
>	i have been looking over the list,
>and it appears this thread talks about extracting data from LV
>that didnt span the bad disk, but i only have one LV.
>http://lists.sistina.com/pipermail/linux-lvm/2003-April/013911.html
>it looks like he is using lvm2
>.
>	does lvm2 have additional tools that can be used to bring back some of
>my data. ??
>
>i was just keeping mp3 on this LV, but when i upgraded redhat i moved
>all my data to it for awhile, and never moved it back.  my wife is
>pissed cause i lost all my digital pictures, and i lost all my personal
>files.  i knew better but didnt listen to myself.  the hard drive gave
>no errors in the logs, which i monitor, it just up and started beating
>its head against the wall.  thanks for any help
>
>randy
>
>
>_______________________________________________
>linux-lvm mailing list
>linux-lvm@sistina.com
>http://lists.sistina.com/mailman/listinfo/linux-lvm
>read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>
>  
>

Re: [linux-lvm] hard drive crash info please

[Randy Perkins]

> Randy Perkins wrote:
>> hello,
>> 	I have a lvm setup of 5 PV (disks), 1 VG,  and 1 LV.  one of the
>> drives
>> just started clicking, and is not recognizable by the bios. Now when i
>> boot up my volume group is not recognized.
>
> I did have the same problem a long time ago. So I started to write a
> tool that would try to recover some the data from the remaining data
> available. I lost the first disk of set of 3 disk...
> This is quite finished.
> It uses the capability of LVM2 to deal with 'partial' VG.
> It only works with ext2/ext3 filesystems and require another (temporary)
>  disk where recovered data will be written.
>
> The URL is at : http://coredump.free.fr/linux/e2retrieve.php
>
> I also did a floppy disk with the essential (kernel 2.4.19 with LVM2
> patch, LVM2 tools, Busybox and e2retrieve).
>
> -=( manu )=-
>

hello,

thanks for the reply.
I went to the link you provided but there is no download available.
did you mean to say earlier "this is not quite finished".
is the download available for e2retrieve ??

I am going to have to study the man pages concerning lvm.
I basically studied it enough to get my volume working,
and since then I havent had to mess with it.

thanks again

randy

Re: [linux-lvm] hard drive crash info please

[Randy Perkins]

> hi
>
> you need lvm2 to  activate a partial volume group. if you have a
> backup  of your vgcfg  then you may restore your config to a new
> harddisk with  the lvm1 tools.
> i had no actual vgcfgbackup so i needed to upgrade to lvm2 to activate
>  the partial vg and make a backup of all partitions that were not
> affected by the crashed harddisk. then i used the vgcfgbackup on the
> partial vg and edited the meadata from the backup (lvm2 saves the
> backup
>  data in human readable form) that all missing segments are put on the
>
> replacement disc. after i restored the backup everything worked again
> and i lost only the data which was on the bad disk.

hello
thanks for your reply
I am going to reread the how-to and man pages so i can
get refreshed on what i am doing.

Is the vgcfg the file that is located in /etc/lvmconf/my_vg.conf ??
If this file is the vgcfg, how current does it need to be ??

at this point you guys have given me hope that i can restore
at least some of my data, and the good news is that the files
I really need are the smaller files.

thanks again
randy

Re: [linux-lvm] hard drive crash info please

[Emmanuel Varagnat]

Randy Perkins wrote:
> 
> hello,

Hi,

> thanks for the reply.
> I went to the link you provided but there is no download available.
> did you mean to say earlier "this is not quite finished".
> is the download available for e2retrieve ??

Not for the moment. It needs some improvements.
I was able to recover a lot of my data (more than 36Go), but without 
their names because directory data were available (or easily available). 
So I'm working on that to see if I can recover some filenames.
I hope to release it the next week.

Good luke.

-=( manu )=-

Re: [linux-lvm] hard drive crash info please

[Randy Perkins]

On Thu, 2003-05-08 at 13:29, Kai Iskratsch wrote:
> hi
> 
> if you have a backup 
> of your vgcfg  then you may restore your config to a new harddisk with 
> the lvm1 tools.

hello,
	I have made some progress on attempting to get some
of my data back out of LVM after a hard drive failure.

all responses are appreciated

I am using lvm-1.0.3 setup with 5 PV, 1 VG, and 1 LV. 

1.	I have a successful vgcfg backup as reported by vgcfgrestore.

2.	pvscan successfully reports my remaining 4 PV.

3.	The defective drive was /dev/hdk.  I have replaced the defective
	drive with a replacement of the same model and size (WD 40GB)

4.	I have ran 'pvcreate /dev/hdk'

5.	By runningy pvdisplay /dev/hd[begik] , I remembered that the 	path of
the physical volumes has changed between the backup and 	restore time. 
I was so impressed with lvm when this was taken 	care of automatically
when they originally changed.

	so my defective disk is /dev/hdk but in the file
	/etc/lvmconf/my_vg.conf it is /dev/hdg.  I was able to deduce 	this
from the UUID listed in vgcfgrestore and compare it to the UUID 	listed
in pvdisplay /dev/hd[begik].

6.	I plan to use the -o option of vgcfgrestore to place the config 	data
from /dev/hdg on /dev/hdk.

questions

When i run the command 'vgcfgrestore -n my_vg -o /dev/hdg /dev/hdk' i
get the following message:

[root@fileserver root]# vgcfgrestore -n my_vg -o /dev/hdg /dev/hdk
vgcfgrestore -- INFO: using backup file "/etc/lvmconf/my_vg.conf"
vgcfgrestore -- "/dev/hdk" may not be an actual physical volume of
volume group "my_vg"

I dont know why this didnt work.

i have succussfully restored onto /dev/hdg with 'vgcfgrestore -i -n
my_vg /dev/hdk',but that was the wrong disk info. i have since ran
'pvcreate -ff /dev/hdk' to clear the info.

I just cant seem to get the -o option to work.
I could always shut down the machine and reposition the disks according
to the backup.
how should i procede ?

After i get the vg restored where do i go then ??
if i was to guesss---
vgchange -a y
fsck /dev/my_vg/log_vol # do i really want to do this? ,which options?
mount /dev/my_vg/log_vol /mnt/mypoint

how will it handle files that are missing ?

how will it handle files that are corrupt ?

will the file names be intact ?

I am assuming that since my PE size is 8MB, that any file
under 8MB will either be there and complete, or completely missing.
I am not really concerned with extracting any files over 8MB anyway.

thanks for reading this and any help is appreciated

thanks again

randy perkins



outputs of above referenced commands follow

--- output of 'vgcfgrestore -ll -n my_vg -t'--------- 

vgcfgrestore -- INFO: using backup file "/etc/lvmconf/my_vg.conf"
vgcfgrestore -- backup of volume group "my_vg"  is consistent
--- Volume group ---
VG Name               my_vg
VG Access             read/write
VG Status             NOT available/resizable
VG #                  0
MAX LV                256
Cur LV                1
Open LV               0
MAX LV Size           511.98 GB
Max PV                256
Cur PV                5
Act PV                5
VG Size               260.80 GB
PE Size               8 MB
Total PE              33383
Alloc PE / Size       33383 / 260.80 GB
Free  PE / Size       0 / 0
VG UUID               bEvsph-VuUD-ETOW-ntYw-I90F-5JDB-CIU4mh

--- Logical volume ---
LV Name                /dev/my_vg/log_vol
VG Name                my_vg
LV Write Access        read/write
LV Status              NOT available
LV #                   1
# open                 0
LV Size                260.80 GB
Current LE             33383
Allocated LE           33383
Allocation             next free
Read ahead sectors     10000
Block device           58:0


--- Physical volume ---
PV Name               /dev/hdb
VG Name               my_vg
PV Size               37.27 GB [78165360 secs] / NOT usable 8.19 MB
[LVM: 146 KB]
PV#                   1
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              4769
Free PE               0
Allocated PE          4769
PV UUID               VzIeJK-1RXR-CY1X-HL7z-tA0S-Jleb-I29HHV

--- Physical volume ---
PV Name               /dev/hde
VG Name               my_vg
PV Size               37.27 GB [78165360 secs] / NOT usable 8.19 MB
[LVM: 146 KB]
PV#                   2
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              4769
Free PE               0
Allocated PE          4769
PV UUID               3v1hNv-VnYq-hFhX-bBKI-6gbx-21f2-jYbgs7

--- Physical volume ---
PV Name               /dev/hdg
VG Name               my_vg
PV Size               37.27 GB [78165360 secs] / NOT usable 8.19 MB
[LVM: 146 KB]
PV#                   3
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              4769
Free PE               0
Allocated PE          4769
PV UUID               pb1svL-UdjB-cvOP-xREw-udxz-z4Px-AnaXVX

--- Physical volume ---
PV Name               /dev/hdi
VG Name               my_vg
PV Size               74.53 GB [156301488 secs] / NOT usable 8.19 MB
[LVM: 165 KB]
PV#                   4
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              9538
Free PE               0
Allocated PE          9538
PV UUID               HOAlS3-7Bck-JzVD-Ds4j-FpEG-lC6L-Gu8peg

--- Physical volume ---
PV Name               /dev/hdk
VG Name               my_vg
PV Size               74.53 GB [156301488 secs] / NOT usable 8.19 MB
[LVM: 165 KB]
PV#                   5
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              9538
Free PE               0
Allocated PE          9538
PV UUID               Mu4FuR-qpqZ-glBu-QB6h-mW3s-7tVf-3iq1s3

--------------------------- end of output -----------------------------

------output of 'pvscan' -----------------------------

pvscan -- reading all physical volumes (this may take a while...)
pvscan -- inactive PV "/dev/hdi"  is associated to unknown VG "my_vg"
(run vgscan)
pvscan -- inactive PV "/dev/hdg"  is associated to unknown VG "my_vg"
(run vgscan)
pvscan -- inactive PV "/dev/hde"  is associated to unknown VG "my_vg"
(run vgscan)
pvscan -- inactive PV "/dev/hdb"  is associated to unknown VG "my_vg"
(run vgscan)
pvscan -- total: 4 [223.61 GB] / in use: 4 [223.61 GB] / in no VG: 0 [0]

-------------------------- end of output---------------------


--------- output of 'pvdisplay /dev/hd[begik]'-----------------

--- Physical volume ---
PV Name               /dev/hdb
VG Name               my_vg
PV Size               37.27 GB [78165360 secs] / NOT usable 8.19 MB
[LVM: 146 KB]
PV#                   1
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              4769
Free PE               0
Allocated PE          4769
PV UUID               VzIeJK-1RXR-CY1X-HL7z-tA0S-Jleb-I29HHV

--- Physical volume ---
PV Name               /dev/hde
VG Name               my_vg
PV Size               74.53 GB [156301488 secs] / NOT usable 8.19 MB
[LVM: 165 KB]
PV#                   4
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              9538
Free PE               0
Allocated PE          9538
PV UUID               HOAlS3-7Bck-JzVD-Ds4j-FpEG-lC6L-Gu8peg

--- Physical volume ---
PV Name               /dev/hdg
VG Name               my_vg
PV Size               74.53 GB [156301488 secs] / NOT usable 8.19 MB
[LVM: 165 KB]
PV#                   5
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              9538
Free PE               0
Allocated PE          9538
PV UUID               Mu4FuR-qpqZ-glBu-QB6h-mW3s-7tVf-3iq1s3

--- Physical volume ---
PV Name               /dev/hdi
VG Name               my_vg
PV Size               37.27 GB [78165360 secs] / NOT usable 8.19 MB
[LVM: 146 KB]
PV#                   2
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       8192
Total PE              4769
Free PE               0
Allocated PE          4769
PV UUID               3v1hNv-VnYq-hFhX-bBKI-6gbx-21f2-jYbgs7

pvdisplay -- no physical volume identifier on "/dev/hdk"

----------------------------- end of output -------------------------