[Linux-cluster] Cluster - process entered in D-state

[Marcos Ferreira da Silva]

I have two nodes in a cluster.
When I try to relocate the vm web2 to node2 then the second node freeze.

[root at vserv4 ~]# clustat
Member Status: Quorate

  Member Name                        ID   Status
  ------ ----                        ---- ------
  vserv3.uniube.br                      1 Online, rgmanager
  vserv4.uniube.br                      2 Online, Local, rgmanager

  Service Name         Owner (Last)                   State
  ------- ----         ----- ------                   -----
  vm:web1              vserv3.uniube.br               started
  vm:web2              vserv3.uniube.br               started

[root at vserv4 ~]# clusvcadm -r vm:web2 -m vserv4.uniube.br
Trying to relocate vm:web2 to vserv4.uniube.br...Success
vm:web2 is now running on vserv4.uniube.br

[root at vserv4 ~]# clustat
Member Status: Quorate

  Member Name                        ID   Status
  ------ ----                        ---- ------
  vserv3.teste.br                      1 Online, rgmanager
  vserv4.teste.br                      2 Online, Local, rgmanager

  Service Name         Owner (Last)                   State
  ------- ----         ----- ------                   -----
  vm:web1              vserv3.teste.br               started
  vm:web2              vserv4.teste.br               starting

when I use ps ax:

17349 ?        S<     0:00 /bin/bash /usr/share/cluster/vm.sh start
17356 ?        S<     0:00 python /usr/sbin/xm create web2 restart=never
--path=/etc/xen
17357 ?        D<     0:00 /usr/bin/python /usr/bin/pygrub -q
--output=/var/lib/xen/xenbl.12384 /storage/web-vms/web2.img


/var/log/messages

Feb 28 15:30:58 vserv4 clurgmgrd[10259]: <notice> Starting stopped
service vm:web2
Feb 28 15:31:00 vserv4 openais[5349]: [CKPT ] checkpoint_find returned 0
Calling error_exit.
Feb 28 15:31:40 vserv4 last message repeated 4 times

After this I have to reboot the vserv4.

The filesystem is gfs2 in RHEL5.

/dev/mapper/VGWEB-LVWebVMs on /storage/web-vms type gfs2
(rw,hostdata=jid=1:id=655361:first=0)
/dev/mapper/VGWEB-LVWeb on /storage/web type gfs2
(rw,hostdata=jid=1:id=786433:first=0)

[root at vserv4 ~]# uname -a
Linux vserv4.teste.br 2.6.18-53.1.13.el5xen #1 SMP Mon Feb 11 13:41:50
EST 2008 x86_64 x86_64 x86_64 GNU/Linux



-  
_____________________________
Marcos Ferreira da Silva
DiGital Tecnologia
Uberlândia - MG
(34) 9154-0150 / 3226-2534