[linux-lvm] New disk; "Couldn't find uuid" after successful pvcreate&vgextend

[linux-lvm] New disk; "Couldn't find uuid" after successful pvcreate&vgextend

[Barnaby Claydon]

Hi all,

I've been successfully running LVM2 (2.01.04) on Debian Sarge 
(2.6.8-2-686) for months.

As I've done many times before, I added a new disk yesterday. I 
successfully ran the following, both without error:

pvcreate -v /dev/hdc
and
vgextend -v mono_media /dev/hdc

As standard practice to make sure everything was ok, I then ran:

vgscan -v

At this point I got several references to "Couldn't find uuid..." 
repeated with the same uuid several times. At this point it seemed 
logical to back-out my changes, so I tried to run the following:

vgreduce -v mono_media /dev/hdc

Unfortunately this gave the same "Couldn't find uuid..." errors as 
'vgscan'. As it seemed somewhat logical, I tried running the same 
'pvcreate' command, again without error:

pvcreate -v /dev/hdc

After 'pvcreate' had finished, I was THEN able to run 'vgreduce' and 
restore my 'mono_media' VG to working order.

At this point I was a bit baffled, running 'lvmdiskscan' after the 
initial 'pvcreate' shows the disk as present, but if I reboot after just 
that first 'pvcreate' step and run 'lvmdisksacn' after the reboot, 
/dev/hdc isn't shown.

To rule-out a hardware problem I successfully ran 'mkfs.ext3 /dev/hdc' 
and was able to mount and navigate the volume without errors. Last night 
I also left 'badblocks -svw /dev/hdc' running which is almost complete 
and hasn't found any errors yet.

I will reply with full command output once 'badblocks' has finished, but 
I hope this brief synopsis makes sense.

Thanks everyone!
-Barnaby

Re: [linux-lvm] New disk; "Couldn't find uuid" after successful pvcreate&vgextend

[Barnaby Claydon]

# pvcreate -v /dev/hdc
    Set up physical volume for "/dev/hdc" with 390721584 available sectors
    Zeroing start of device /dev/hdc
  Physical volume "/dev/hdc" successfully created


# pvdisplay -v /dev/hdc
    Using physical volume(s) on command line
  --- NEW Physical volume ---
  PV Name               /dev/hdc
  VG Name
  PV Size               186.31 GB
  Allocatable           NO
  PE Size (KByte)       0
  Total PE              0
  Free PE               0
  Allocated PE          0
  PV UUID               xacykK-MXPT-2yO0-qf2R-TBTU-VJb5-vVYw4Q


I haven't tried vgextend again since running badblocks - it looks like 
the drive is fine though:

# badblocks -svw /dev/hdc
Checking for bad blocks in read-write mode
 From block 0 to 195360984
Testing with pattern 0xaa: done                        984
Reading and comparing: done                        984
Testing with pattern 0x55: done                        984
Reading and comparing: done                        984
Testing with pattern 0xff: done                        984
Reading and comparing: done                        984
Testing with pattern 0x00: done                        984
Reading and comparing: done                        984
Pass completed, 0 bad blocks found.






James Hammett wrote:
> What do you get when you try pvdisplay?
>
> James
>
>> Hi all,
>>
>> I've been successfully running LVM2 (2.01.04) on Debian Sarge 
>> (2.6.8-2-686) for months.
>>
>> As I've done many times before, I added a new disk yesterday. I 
>> successfully ran the following, both without error:
>>
>> pvcreate -v /dev/hdc
>> and
>> vgextend -v mono_media /dev/hdc
>>
>> As standard practice to make sure everything was ok, I then ran:
>>
>> vgscan -v
>>
>> At this point I got several references to "Couldn't find uuid..." 
>> repeated with the same uuid several times. At this point it seemed 
>> logical to back-out my changes, so I tried to run the following:
>>
>> vgreduce -v mono_media /dev/hdc
>>
>> Unfortunately this gave the same "Couldn't find uuid..." errors as 
>> 'vgscan'. As it seemed somewhat logical, I tried running the same 
>> 'pvcreate' command, again without error:
>>
>> pvcreate -v /dev/hdc
>>
>> After 'pvcreate' had finished, I was THEN able to run 'vgreduce' and 
>> restore my 'mono_media' VG to working order.
>>
>> At this point I was a bit baffled, running 'lvmdiskscan' after the 
>> initial 'pvcreate' shows the disk as present, but if I reboot after 
>> just that first 'pvcreate' step and run 'lvmdisksacn' after the 
>> reboot, /dev/hdc isn't shown.
>>
>> To rule-out a hardware problem I successfully ran 'mkfs.ext3 
>> /dev/hdc' and was able to mount and navigate the volume without 
>> errors. Last night I also left 'badblocks -svw /dev/hdc' running 
>> which is almost complete and hasn't found any errors yet.
>>
>> I will reply with full command output once 'badblocks' has finished, 
>> but I hope this brief synopsis makes sense.
>>
>> Thanks everyone!
>> -Barnaby
>>
>> _______________________________________________
>> linux-lvm mailing list
>> linux-lvm@redhat.com
>> https://www.redhat.com/mailman/listinfo/linux-lvm
>> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>
>

Re: [linux-lvm] New disk; "Couldn't find uuid" after successful pvcreate&vgextend

[Barnaby Claydon]

After the below error-free commands, I went ahead and ran 'vgextend' and 
then 'lvextend' without any 'vgscan' tests in between, and the whole 
system is fine now, 'xfs_grow' went off problem free as well.

Very odd. Are there caveats about when to run 'vgscan' maybe?

Thanks for the response James,
-Barnaby



Barnaby Claydon wrote:
> # pvcreate -v /dev/hdc
>    Set up physical volume for "/dev/hdc" with 390721584 available sectors
>    Zeroing start of device /dev/hdc
>  Physical volume "/dev/hdc" successfully created
>
>
> # pvdisplay -v /dev/hdc
>    Using physical volume(s) on command line
>  --- NEW Physical volume ---
>  PV Name               /dev/hdc
>  VG Name
>  PV Size               186.31 GB
>  Allocatable           NO
>  PE Size (KByte)       0
>  Total PE              0
>  Free PE               0
>  Allocated PE          0
>  PV UUID               xacykK-MXPT-2yO0-qf2R-TBTU-VJb5-vVYw4Q
>
>
> I haven't tried vgextend again since running badblocks - it looks like 
> the drive is fine though:
>
> # badblocks -svw /dev/hdc
> Checking for bad blocks in read-write mode
> From block 0 to 195360984
> Testing with pattern 0xaa: done                        984
> Reading and comparing: done                        984
> Testing with pattern 0x55: done                        984
> Reading and comparing: done                        984
> Testing with pattern 0xff: done                        984
> Reading and comparing: done                        984
> Testing with pattern 0x00: done                        984
> Reading and comparing: done                        984
> Pass completed, 0 bad blocks found.
>
>
>
>
>
>
> James Hammett wrote:
>> What do you get when you try pvdisplay?
>>
>> James
>>
>>> Hi all,
>>>
>>> I've been successfully running LVM2 (2.01.04) on Debian Sarge 
>>> (2.6.8-2-686) for months.
>>>
>>> As I've done many times before, I added a new disk yesterday. I 
>>> successfully ran the following, both without error:
>>>
>>> pvcreate -v /dev/hdc
>>> and
>>> vgextend -v mono_media /dev/hdc
>>>
>>> As standard practice to make sure everything was ok, I then ran:
>>>
>>> vgscan -v
>>>
>>> At this point I got several references to "Couldn't find uuid..." 
>>> repeated with the same uuid several times. At this point it seemed 
>>> logical to back-out my changes, so I tried to run the following:
>>>
>>> vgreduce -v mono_media /dev/hdc
>>>
>>> Unfortunately this gave the same "Couldn't find uuid..." errors as 
>>> 'vgscan'. As it seemed somewhat logical, I tried running the same 
>>> 'pvcreate' command, again without error:
>>>
>>> pvcreate -v /dev/hdc
>>>
>>> After 'pvcreate' had finished, I was THEN able to run 'vgreduce' and 
>>> restore my 'mono_media' VG to working order.
>>>
>>> At this point I was a bit baffled, running 'lvmdiskscan' after the 
>>> initial 'pvcreate' shows the disk as present, but if I reboot after 
>>> just that first 'pvcreate' step and run 'lvmdisksacn' after the 
>>> reboot, /dev/hdc isn't shown.
>>>
>>> To rule-out a hardware problem I successfully ran 'mkfs.ext3 
>>> /dev/hdc' and was able to mount and navigate the volume without 
>>> errors. Last night I also left 'badblocks -svw /dev/hdc' running 
>>> which is almost complete and hasn't found any errors yet.
>>>
>>> I will reply with full command output once 'badblocks' has finished, 
>>> but I hope this brief synopsis makes sense.
>>>
>>> Thanks everyone!
>>> -Barnaby
>>>
>>> _______________________________________________
>>> linux-lvm mailing list
>>> linux-lvm@redhat.com
>>> https://www.redhat.com/mailman/listinfo/linux-lvm
>>> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>>
>>
>
>