[linux-lvm] Can't get rid of INACTIVE read/write snapshot

[linux-lvm] Can't get rid of INACTIVE read/write snapshot

[Marcel Gsteiger]

Hi all

I'm using lvm2-2.01.08-2.1 on fc4, kernel 2.6.16-1 with linux-vserver patches.

A while ago, I mistakenly created a 39GB read-write snapshot of my /vservers partition (which, btw, seemed to work). But after some time this partition got to an overflow and since then I can't get rid of it anymore. See the lvdisplay output below.

lvchange -an or --refresh gives me a message "Can't change snapshot logical volume "vservers_snapshot"".

lvremove lets me wait, then "top" shows 100% IO-Wait but does not come to an end (at least after an hour). I don't know how much patience I would need, this box is rather fast otherwise (dual xeon 2.8GHz w/ SATA drives).

Rebooting does not help anything.

Is there a safe way to simply get rid of this snapshot? Any help would much be appreciated.

Regards
--Marcel

 --- Logical volume ---
  LV Name                /dev/vg00/lv00_vservers
  VG Name                vg00
  LV UUID                9oELP3-2l7d-gTML-2xg2-unTu-znP0-qVd3Mc
  LV Write Access        read/write
  LV snapshot status     source of
                         /dev/vg00/vservers_snapshot [INACTIVE]
  LV Status              available
  # open                 1
  LV Size                39,06 GB
  Current LE             625
  Segments               1
  Allocation             contiguous
  Read ahead sectors     0
  Block device           253:9
   
  --- Logical volume ---
  LV Name                /dev/vg00/vservers_snapshot
  VG Name                vg00
  LV UUID                cQEj3z-V2jD-QOCl-3Yea-ETQZ-Egj9-2B7nd2
  LV Write Access        read/write
  LV snapshot status     INACTIVE destination for /dev/vg00/lv00_vservers
  LV Status              available
  # open                 0
  LV Size                39,06 GB
  Current LE             625
  Segments               1
  Snapshot chunk size    4,00 KB
  Allocated to snapshot  100,00% 
  Allocation             inherit
  Read ahead sectors     0
  Block device           253:8

Re: [linux-lvm] Can't get rid of INACTIVE read/write snapshot

[Mikulas Patocka]

> Hi all
>
> I'm using lvm2-2.01.08-2.1 on fc4, kernel 2.6.16-1 with linux-vserver 
> patches.
>
> A while ago, I mistakenly created a 39GB read-write snapshot of my 
> /vservers partition (which, btw, seemed to work). But after some time 
> this partition got to an overflow and since then I can't get rid of it 
> anymore. See the lvdisplay output below.
>
> lvchange -an or --refresh gives me a message "Can't change snapshot 
> logical volume "vservers_snapshot"".
>
> lvremove lets me wait, then "top" shows 100% IO-Wait but does not come 
> to an end (at least after an hour). I don't know how much patience I 
> would need, this box is rather fast otherwise (dual xeon 2.8GHz w/ SATA 
> drives).

Hi

It looks like an unknown bug in lvremove. Send a strace of lvremove while 
it waits for that long amount of time.

Mikulas

> Rebooting does not help anything.
>
> Is there a safe way to simply get rid of this snapshot? Any help would 
> much be appreciated.
>
> Regards
> --Marcel
>
> --- Logical volume ---
>  LV Name                /dev/vg00/lv00_vservers
>  VG Name                vg00
>  LV UUID                9oELP3-2l7d-gTML-2xg2-unTu-znP0-qVd3Mc
>  LV Write Access        read/write
>  LV snapshot status     source of
>                         /dev/vg00/vservers_snapshot [INACTIVE]
>  LV Status              available
>  # open                 1
>  LV Size                39,06 GB
>  Current LE             625
>  Segments               1
>  Allocation             contiguous
>  Read ahead sectors     0
>  Block device           253:9
>
>  --- Logical volume ---
>  LV Name                /dev/vg00/vservers_snapshot
>  VG Name                vg00
>  LV UUID                cQEj3z-V2jD-QOCl-3Yea-ETQZ-Egj9-2B7nd2
>  LV Write Access        read/write
>  LV snapshot status     INACTIVE destination for /dev/vg00/lv00_vservers
>  LV Status              available
>  # open                 0
>  LV Size                39,06 GB
>  Current LE             625
>  Segments               1
>  Snapshot chunk size    4,00 KB
>  Allocated to snapshot  100,00%
>  Allocation             inherit
>  Read ahead sectors     0
>  Block device           253:8
>
> _______________________________________________
> 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/
>

Antw: Re: [linux-lvm] Can't get rid of INACTIVE read/write snapshot

[Marcel Gsteiger]

[-- Attachment #1: Type: text/plain, Size: 3234 bytes --]

.sorry for the delay, I tried again to lvremove, find the strace enclosed.

After the output "flock(3, LOCK_EX

at the end, lvremove seems to go into an endless loop (100% CPU usage).

Then I tried the following:

kill -9 <pid of lvremove>
rm /var/lock/lvm/V_vg00

to remove /var/lock/lvm/V_vg00 "by hand" and tried again. Now the system does not go to 100% CPU usage anymore, but lvremove does not terminate. See my second lvremove2.strace.

As mentioned before, I simply wanted to get rid of a overflowed snapshot and clean up my volume group, as I should extend the LVM in question. How can I safely accomplish this in my situation?

Thanks for your help and best regards
--Marcel








>>> Mikulas Patocka <mpatocka@redhat.com> 03.06.2008 21:15 >>>
> Hi all
>
> I'm using lvm2-2.01.08-2.1 on fc4, kernel 2.6.16-1 with linux-vserver 
> patches.
>
> A while ago, I mistakenly created a 39GB read-write snapshot of my 
> /vservers partition (which, btw, seemed to work). But after some time 
> this partition got to an overflow and since then I can't get rid of it 
> anymore. See the lvdisplay output below.
>
> lvchange -an or --refresh gives me a message "Can't change snapshot 
> logical volume "vservers_snapshot"".
>
> lvremove lets me wait, then "top" shows 100% IO-Wait but does not come 
> to an end (at least after an hour). I don't know how much patience I 
> would need, this box is rather fast otherwise (dual xeon 2.8GHz w/ SATA 
> drives).

Hi

It looks like an unknown bug in lvremove. Send a strace of lvremove while 
it waits for that long amount of time.

Mikulas

> Rebooting does not help anything.
>
> Is there a safe way to simply get rid of this snapshot? Any help would 
> much be appreciated.
>
> Regards
> --Marcel
>
> --- Logical volume ---
>  LV Name                /dev/vg00/lv00_vservers
>  VG Name                vg00
>  LV UUID                9oELP3-2l7d-gTML-2xg2-unTu-znP0-qVd3Mc
>  LV Write Access        read/write
>  LV snapshot status     source of
>                         /dev/vg00/vservers_snapshot [INACTIVE]
>  LV Status              available
>  # open                 1
>  LV Size                39,06 GB
>  Current LE             625
>  Segments               1
>  Allocation             contiguous
>  Read ahead sectors     0
>  Block device           253:9
>
>  --- Logical volume ---
>  LV Name                /dev/vg00/vservers_snapshot
>  VG Name                vg00
>  LV UUID                cQEj3z-V2jD-QOCl-3Yea-ETQZ-Egj9-2B7nd2
>  LV Write Access        read/write
>  LV snapshot status     INACTIVE destination for /dev/vg00/lv00_vservers
>  LV Status              available
>  # open                 0
>  LV Size                39,06 GB
>  Current LE             625
>  Segments               1
>  Snapshot chunk size    4,00 KB
>  Allocated to snapshot  100,00%
>  Allocation             inherit
>  Read ahead sectors     0
>  Block device           253:8
>
> _______________________________________________
> 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/ 
>


[-- Attachment #2: strace.zip --]
[-- Type: application/x-zip-compressed, Size: 12561 bytes --]