[Linux-cluster] clvmd Falling back to local file-based locking

Christine Caulfield ccaulfie at redhat.com
Tue Jun 10 10:39:11 UTC 2008


Peter Haufschild wrote:
> Hallo,
> after a reboot I start cman and clvmd and the start success.
> 
> 3 Blades show's me
> 
> [root at blade1 ~]# pvdisplay
>     Logging initialised at Tue Jun 10 12:18:56 2008
>     Set umask to 0077
>     Scanning for physical volume names
>   --- Physical volume ---
>   PV Name               /dev/sdb1
>   VG Name               SATAStorage
>   PV Size               4096,00 EB / not usable 4096,00 EB
>   Allocatable           yes 
>   PE Size (KByte)       4096
>   Total PE              476906
>   Free PE               4842
>   Allocated PE          472064
>   PV UUID               I2xV9i-8SN7-ifza-CmJA-W6ez-b2cT-OL4gxw
>    
>   --- Physical volume ---
>   PV Name               /dev/sda1
>   VG Name               SCSIStorage
>   PV Size               410,19 GB / not usable 3,40 MB
>   Allocatable           yes 
>   PE Size (KByte)       4096
>   Total PE              105008
>   Free PE               19248
>   Allocated PE          85760
>   PV UUID               695f63-a0Tv-lEJ9-Xubw-f4We-ymak-TZ0ah4
>    
>     Wiping internal VG cache
> This are OK.
> 
> But 2 Blades show me 
> 
> 
> [root at blade9 ~]# lvdisplay
> lvm2    Logging initialised at Tue Jun 10 12:14:41 2008
> lvm2    Set umask to 0077
> lvm2  connect() failed on local socket: Verbindungsaufbau abgelehnt
> lvm2  WARNING: Falling back to local file-based locking.
> lvm2  Volume Groups with the clustered attribute will be inaccessible.
> lvm2    Finding all logical volumes
> lvm2  Skipping clustered volume group SATAStorage
> lvm2  Skipping clustered volume group SCSIStorage
> lvm2    Wiping internal VG cache
> [root at blade9 ~]# vgdisplay
> lvm2    Logging initialised at Tue Jun 10 12:20:13 2008
> lvm2    Set umask to 0077
> lvm2  connect() failed on local socket: Verbindungsaufbau abgelehnt
> lvm2  WARNING: Falling back to local file-based locking.
> lvm2  Volume Groups with the clustered attribute will be inaccessible.
> lvm2    Finding all volume groups
> lvm2    Finding volume group "SATAStorage"
> lvm2  Skipping clustered volume group SATAStorage
> lvm2    Finding volume group "SCSIStorage"
> lvm2  Skipping clustered volume group SCSIStorage
> lvm2    Wiping internal VG cache
> 
> cluster.conf and lvm.conf identical.
> 
> 
> This difference I could see already at boot time, when activating lvm's. 

Assuming that Google's translation to "Connection Rejected" is correct 
then it sounds like either clvmd isn't running or possibly SELinux 
getting in the way of connecting to clvmd.

You say clvmd has started, but have you check it is actually running ?

Chrissie




More information about the Linux-cluster mailing list